Brainstorming alternatives - The Fast Feedback Cycle - Scenario-Focused Engineering (2014)

Scenario-Focused Engineering (2014)

Part II: The Fast Feedback Cycle

Chapter 7. Brainstorming alternatives



Now it’s time to change gears and shift from the first half of the Fast Feedback Cycle, which is fundamentally about understanding the problem to be solved, into the second half of the cycle, which focuses on creating solutions. The first step in solving the problem you’ve identified is to explore a wide range of alternatives; then you narrow in on a few specific approaches to pursue.

The Brainstorm stage is one of the easiest ways to get a big bang for the buck with as little as a few hours of effort, and it should be an integral part of almost any problem-solving exercise, large or small, technical or not. However, we’ve found that this stage is where our intuition as analytical engineers is most likely to lead us astray, and some of our biggest mistakes and oversights can be traced back to this stage.

And note that while the last stage was about converging—making decisions about which problems you will focus on and at which scope—now you are going to switch gears and again diverge as you strive to generate lots of alternative ideas for a solution. This converge/diverge pattern happens repeatedly during the Fast Feedback Cycle, and we’ll discuss it in more detail in Chapter 10, “The importance of iteration.”

Where does innovation come from?

Deep down inside, every one of us dreams about having the next big, bold idea, an innovation so terrific that it will become a household name, make all our competitors jealous, and maybe even redefine the industry or disrupt a market. Ever the optimists, we see possibilities around every corner, and the ambitious among us gravitate toward projects that hold the potential for this kind of game-changing innovation.

Patterns of successful innovation

It’s worthwhile to take a minute and think about where innovative, commercially successful ideas have actually come from. Is there a common lineage, a pattern of some sort, or even a secret formula that suggests a path for discovering the next big thing? Let’s consider a few examples of industry-changing innovations in the history of computing.

Evolution over long time periods

Not too long ago, the first general-purpose computers with word processors became available to college students. Suddenly, students no longer needed to retype an entire chapter of their thesis when they had to insert a new paragraph in the middle of a page. Over the years, modern word processing has not just enabled greater ease in the act of writing; it has also fundamentally changed how we go about the writing process. It’s no longer painful to reorganize your thoughts or to develop your ideas more deeply, with basic tools like the insertion point and cutting and pasting literally at your fingertips. Making repeated revisions on screen has become the norm—you no longer need to desperately hope that you’ll get the job done with one rough draft and a single edit pass, and then hire a typist. Word processing has pervaded almost every aspect of modern life, from writing a Facebook post, to sending email, to jotting down a camping trip packing list. But where did the idea of word processing come from anyway? Was it a brand-new, revolutionary idea that had never been conceived of before, or did it have more evolutionary roots?

Going way, way back, of course, there was paper and ink. Then came the Gutenberg printing press in the 1400s, and a few centuries later the invention of the manual typewriter, popularized by the Remington in the 1870s, which introduced the QWERTY keyboard we still use today. In the modern age, widespread typewriting by secretaries and high standards in the business world spurred the invention of Liquid Paper correction fluid (originally called Mistake Out) in 1956. Then came a succession of electric typewriters, such as the IBM Selectric (1961), that helped people type even faster. In marketing for the Selectric, the term “word processing,” or “W/P,” was coined, which morphed in meaning over the following decades. And then came a very popular innovation, when a special-purpose whiteout tape was added to typewriters that allowed typists to back up a few characters to fix mistakes without having to pull out the paper and dab on whiteout. A slightly bigger transition to an electronic typewriter included a built-in, one-line LCD screen, where you could see and edit an entire line of text before it was actually committed to the page. This grew to a two-line screen and then a four-line screen. In parallel, the personal computer entered the scene, and it naturally evolved to take on the functionality already available on a typewriter and make it available on an even larger screen. Here, the first dedicated word-processing computers were born, such as the Digital DECmate, and this functionality soon became available on general-purpose personal computers as well. From there, steady improvement in command-line interfaces gradually evolved into better and better WYSIWYG (What You See Is What You Get) interfaces as the personal computer came of age, with floppy disks and printers growing up alongside. In the past couple of decades, rich-edit text boxes that embed word-processing functionality have become commonplace in nearly every modern computer application.

When you think about this progression, what ended up being a revolutionary innovation in modern computing was actually the result of steady evolutionary improvements from multiple, interconnected sources over a long period of time. It’s hard to look at that historical progression and point to any single moment or transition and identify the single golden moment that changed the world, or exactly which step invented word processing. This is a great example of the first pattern of successful innovation: how long-term, evolutionary improvements lead to important innovations.

Combining ideas

Let’s look at another example. Where did the idea for free web-based email come from? Hotmail was one of the first of these services. As with word processing, web-based email has deeply affected modern life, from how we do business, to how we communicate with friends, to how often we bother getting out an envelope and a stamp to send snail mail to pay a bill. But what was really innovative about Hotmail?

At the time of Hotmail’s founding in 1996, the idea of email itself was hardly new. Universities and corporations had long had email systems, using proprietary clients such as Microsoft Mail, Lotus Notes, cc:Mail, or various UNIX or VAX systems. Home users could get email service as well through ISPs such as America Online or CompuServe. At that time, neither the Internet nor websites were brand-new, as the Internet revolution was already well underway. And certainly the idea of providing a web service for free was not new, as most of the nascent Internet industry at that time focused on attracting users and usage more than profits.

Hotmail is a great example of combining three already-existing ideas: email, a website, and free to consumers. None of the building blocks of Hotmail were in themselves brand-new, but the particular combination of these existing ideas turned out to be magical and resulted in a revolutionary innovation that has been copied by many. The same basic formula that Hotmail pioneered continues to be the norm almost two decades later. This example illustrates a second pattern: combining existing ideas in new ways is a common path to innovation.

Continuous improvement

How about a more recent success: Facebook? Where did Facebook come from? Again, you can trace an industrywide, evolutionary history, starting with UNIX bulletin boards and AOL member profiles back in the 1990s to the 1995 launch of for discovering old high school friends. All of these had a notion of a customizable webpage dedicated to one person’s profile. Then, in 1997, popped up, showing how closely you were connected to the well-known actor Kevin Bacon. The degree of separation within a hierarchy of friends has been a widely used concept ever since, although SixDegrees itself didn’t last long (it shut down in 2001). Then came the popular but irreverent “Am I Hot or Not?,” an attractiveness voting site created by UC Berkeley students James Hong and Jim Young, later mimicked by Mark Zuckerberg at Harvard with a site called Facemash. Friendster was born in 2002 and was also wildly popular for a time, followed closely by LinkedIn and MySpace in 2003. Facebook was actually rather late on the scene, starting as an Ivy League-only social network in 2004, initially aimed at hooking you up with a friend of a friend for a date, before opening its doors to the general public in 2006.

When you examine the history, you see an evolutionary story behind the beginning of Facebook. Furthermore, even though many notable features have driven adoption and usage, none of them were actually universally loved ideas when they shipped, nor were they available in the very early versions, nor was Facebook always the first to ship them. For instance, Facebook’s news feed was released to significant controversy about compromising personal privacy, and only in 2006, two full years after Facebook’s launch.1 The famous Facebook Like button was conceived by the development team in 2007 as the Awesome button, but shipped two years later, in 2009, as the Like button, after competitor FriendFeed independently released a similar “Like” feature in 2007 to minimal fanfare. (FriendFeed was eventually acquired by Facebook.)2

It’s easy to think that Facebook sprang fully formed out of Mark Zuckerberg’s head and has always been the way we see it today, but when you look at the timeline, its journey really was much slower and more evolutionary. The first version of Facebook was a great idea, and compelling for its time, but it lacked much of what we now recognize as the hallmarks of the Facebook experience. This, of course, in no way discounts Facebook’s remarkable popularity, but it makes the point that brilliant and successful innovation does not need to be birthed from radically new ideas that are unlike anything that has come before. Nor does successful innovation need to happen all at once in a single big bang at the very first release. Nor is it one single killer feature that paves the way. In fact, innovation rarely happens that way.

Alongside an evolving industry and some deft combination of existing ideas, which follow the patterns of innovation we’ve already discussed, another factor behind Facebook’s success was a steady stream of constant, small improvements that grew Facebook from a Harvard dorm project (2004) to a company that almost a billion people use every day (2014)—and it’s worth pointing out that this journey took an entire decade. Tiny improvements like the color and placement of a link or the wording on a sign-up screen all affected usage and growth by some small amount. For instance, the news feed underwent a series of tweaking and privacy-control improvements to get the controls just right. When you add together each of the things that have had a positive effect, those little things tend to add up. And over not just a few years, but over a decade, they add up a lot. Facebook is very much a data-driven company. It has a culture of analyzing data, trying things out, measuring effectiveness, and iterating continually, and the company’s management team tolerates risk as a necessary part of making these continuous improvements. Facebook is a great example of the third common pattern of successful innovation—innovation is often achieved through persistent, continuous improvement over successive releases.

Continuous improvement isn’t a new idea, by the way, and it is very popular with online services, and increasingly so with the advent of big-data analysis and A/B testing that can power constant fine-tuning. Historically, Toyota has also relied heavily on continuous improvement in optimizing its manufacturing processes—a concept they call kaizen, which also led to the genesis of the Lean movement.

Innovation does not happen all at once

With those patterns in mind, take a moment to look around and notice all the technology success stories in the world around us today. On the surface, these innovations may seem to have dropped out of the sky, fully formed as breakthrough ideas that were conceived by a brilliant, creative genius, team, or company. But think about the history and the timeline behind those innovations—never mind the execution required to make those innovations a reality.

The fact is that all innovative ideas have a history, even those that end up revolutionizing a category or an industry. They all have a much more pedestrian, evolutionary path than appears on the surface—building on previous ideas, combining and recombining ideas, trying different approaches until a winning combination is found, and then building, optimizing, and continually improving. It’s also important to remember that along the path to success were plenty of false starts, missteps, and failures, as well as plateaus, waiting for someone to take the idea one step further to find the magic combination. These ideas took a lot longer to go from concept to success story—years, not months—than we often think about.

Three important implications stem from the realization that successful innovation is a lot more evolutionary in nature.

Innovation may be closer than you think

The first implication is that people often reach too high when they’re looking for an innovative, game-changing idea. We strive for a brand-new approach that no one has ever thought of before, that doesn’t look anything like the current available technology—a radical, revolutionary invention. Yet there are plenty of examples of products that were too far out there, too far ahead of their time, which simply did not catch on.

The first touchscreen tablet was shipped by Microsoft 15 years ago, but it didn’t sell very well at the time, which is quite surprising to contemplate today in the age of the iPad. Pitney Bowes executives anticipated the idea of the cashless society in the 1960s,3 suggesting things like electronic fund transfers, barcode scanning, ATMs, and credit/debit cards, which are now commonplace, of course, but at the time seemed inconceivable and took decades to actually happen. The Segway was an amazing technological feat, but it still largely remains a curiosity, hardly the personal transportation revolution that its inventor, Dean Kamen, had predicted. These ideas were too different, too far ahead of their time to win mass-market approval and adoption when they first appeared.

On the other hand, many products and services that have been wildly successful commercially were initially criticized for not being particularly novel ideas or new technology, but they were the right combination at the right time. Practically speaking, this means that a more evolutionary style of innovation offers a lot more opportunity than you might realize. Just the right improvements, tweaks, and new combinations of existing ideas may be all you need to have a surprisingly large impact on utility and desirability for your customers.

Successful ideas satisfy deep human needs (that aren’t obvious)

The second implication is that we mistakenly believe that we will instantly recognize a great solution when we see it. However, even the launch of the iPad came with its share of people saying, “Who needs a giant iPhone? It’s a toy that will be used once, and will be a flash in the pan.” The Facebook news feed, arguably the company’s most important element for sustained, regular usage, was initially released to significant controversy. Even the release of a runaway success has a mixed chorus of prognosticators, some of whom love your solution and some of whom don’t.

However, if you look more deeply, you see that successful innovations found a way to meet a deep customer need really well. And because unarticulated needs are tough to recognize, the power of these solutions was not universally predicted in advance or fully appreciated right away. Sometimes innovators are thoughtful about what deep needs they are solving, and sometimes it seems more like they just got lucky. But either way, satisfying the deep need was the root of their success and carried them through some mistakes, detours, and imperfect solutions.

For example, word processing served a huge population of students, professionals, and consumers who needed to write things down and who didn’t always know ahead of time exactly what they wanted to say. At the time, it would have been easy to identify the surface needs of people using typewriters—they have poor spelling skills, they don’t take time to plan their writing, they need to be able to rewrite quickly, and so on. However, in hindsight we know that the deep need was embedded in the cognitive complexity of the writing process itself, not in the gap in typing skills between professional secretaries and the general public, which was not at all obvious at the time.

Hotmail and other free web email services allowed people to communicate almost instantly with friends across the street or across the globe—a behavior borne of the deep-seated human need to communicate with others. But at the time, some argued that nothing would ever replace the intimacy of a handwritten letter, and email would be good only for doing business and never for personal communication. But it turned out that being able to instantly send pictures and greetings to a family member across the world was a deeply satisfying way to communicate, not to mention cheap and convenient.

Facebook and other social networks capitalize on numerous big-T human truths: from staying in touch with loved ones and getting advice from people you trust, to baser instincts like wanting to be visibly popular by collecting a large list of friends, comparing details of others’ lives with your own, and even a touch of curiosity-driven voyeurism. Despite Facebook’s popularity, people still argue over whether its news feed intrudes on privacy, or they denounce the addictive habit of constantly checking your feed for updates. But Facebook is addictive exactly because it satisfies a deep human desire to be part of a community and be in the know, and that turns out to be a pretty powerful driver of human behavior. Even though Facebook’s culture of continuous improvement certainly helped spur things along, it would have gone nowhere if a core human need wasn’t being filled in the first place.

The bottom line is that even unquestionably successful innovations have their doubters and detractors, and the extent of their success may have been far from obvious ahead of time. But whether premeditated or not, successful innovations are fueled by serving deep human needs extremely well.

SFE in action: Adding emoticons to MSN Messenger

Kelvin Chan, Business Evangelist, Microsoft Developer Partner Experience

When I was a program manager working on the MSN Messenger team in 2001, the engineering team was rushing to catch up to the feature set of its competitor at the time, AOL Instant Messenger (AIM), as well as having to make continual infrastructure improvements to support our exponentially growing user base. The list of work was long, and the team was small, so every feature proposed for addition was heavily scrutinized and debated.

We had an existing feature in the MSN Chat web app called emoticons, which allowed users to include graphical smiley faces in their messages alongside their text. I argued to port this feature to MSN Messenger. People had been using smiley faces of all sorts in their messages for years, with text characters such as these:

:-) ;) :) :(

The idea was to automatically convert these text-based smiley faces into tiny images that represented a graphical smiley (or frowny) face and to provide a drop-down menu that displayed a list of smileys you could include in your message. You could also select a number of other emoticons, from animals to musical notes to a little red devil.

Porting the code from MSN Chat to reuse it in MSN Messenger was pretty straightforward, so the cost was small. The work item was approved, and the feature shipped in the next release of MSN Messenger. Almost immediately, it became the headline feature for the release, eclipsing work items that the team had spent orders of magnitude more time and energy on but were nowhere near as popular or valued by users. How was it that this minor tweak of a feature resonated with users so deeply?

I wish I could say that we predicted that graphical emoticons would be so successful in MSN Messenger. In truth, we just got lucky. But as it turned out, the idea of emoticons spread rapidly across many different form factors over the years, from instant messaging, to email, to word processors, to mobile phones. And a full panoply of smileys was merged into the 2010 Unicode standards. Future versions of MSN Messenger added ever larger libraries of emoticons, including animated ones, hidden “secret” emoticons, and the ability to upload your own custom images. However, it was really the simple, facial-expression smileys that received the lion’s share of use. Arguably, the others were merely a distraction, proving that we still didn’t fully understand the true unarticulated needs that this feature was addressing.

But why were emoticons so successful? Because they helped people express emotions in the context of their textual messages, and emotions are core to the human communication experience. Both email and instant messaging can suffer from misinterpretation because the context isn’t clear—a remark that was intended to be sarcastic is taken literally, a friendly suggestion is heard as criticism, a loving remark just doesn’t seem all that intimate. Emoticons can alleviate some of the misinterpretation.

Simple text emoticons had already gotten the idea halfway there, but making emoticons graphical brought the idea mainstream, making them easily understandable and accessible to all users. Emoticons solved a deeper problem for users—avoiding the embarrassment of being misunderstood.

There is a science behind innovation

The third implication of innovation being evolutionary is perhaps the most important one: there is actually a science behind innovation, and there is a discipline around idea generation that can be learned and cultivated over time. Successful innovations are much less about a lone genius who suddenly has an inspiration and much more about a sustained effort, gradual improvement, and paying attention to and learning from feedback, which is the essence of the Fast Feedback Cycle. Even when a clever new idea is the seed, a lot of effort needs to go into tending and nurturing and pruning for the idea to be successful in the marketplace.

This chapter will help you learn the practical science and specific techniques behind how to unleash your gray matter to come up with better ideas to help fuel the Fast Feedback Cycle. Sometimes you will be hoping to find a magic new combination that will put you on a path to disrupting the competition, whereas more often you simply need to come up with a great solution to a reasonably tactical problem. Whatever your goal, there are techniques that will dramatically improve your odds of finding an optimal solution with high efficiency. The important thing, you’ll find, is not to focus exclusively on generating a single killer idea, but to iterate several ideas in parallel and use feedback to combine and refine the most promising ideas into a truly innovative, winning solution.

Image Mindshift

Whose job is it to be creative? One myth about innovation is that the really good ideas all come from a very small set of creatively gifted souls, and your best bet is to make sure you have one of those people on your team and listen to what she or he says. While it’s true that some people seem better than others at finding out-of-the-box ideas, it’s not at all clear that this skill is innate or even that a “creative” person’s ideas will always be golden.

Even for Apple, the biggest ideas did not actually start with Steve Jobs. At Apple, an elite team of industrial designers (which Jobs presumably helped recruit and hire) played that role. Jobs’s main role was as gatekeeper, decreeing when an idea was “good enough.” Steve Jobs consistently maintained incredibly high standards, blocking many products from releasing until they met his exacting scrutiny. He created and insisted on a strong culture around design, creativity, and putting the customer first. He identified the problems to be solved, but he did not generate many of the ideas for the solutions themselves. Those innovations came from a very strong design team, which, incidentally, made use of many of the techniques that we discuss in this book. The magic at Apple was a talented team that had a shared priority for customers, an iterative design process, and a gatekeeper with very high standards who was willing to wait for the right solution to emerge.

So whose job is it to be creative? It’s the job of the team, and the job of leaders is to support an innovative culture. Prepare for iteration, let people try things, celebrate failure, measure results, and do it again. Creativity is not the job of any one person.

You have to get over the notion that you, or someone on your team, or someone you want to hire is going to suddenly receive a lightning bolt of supreme insight, the single, brilliantly creative, innovative, market-disruptive, billion-dollar idea. And even if this happened, you are not likely to recognize it as such, at least not right away. If you take the time to study the history of innovative products, you see that innovation just doesn’t happen that way.

Explore lots of alternatives

Linus Pauling, the only person ever to win two undivided Nobel Prizes,4 famously said,

If you want to have good ideas you must have many ideas. Most of them will be wrong, and what you have to learn is which ones to throw away.

We wholeheartedly agree. To get started creating solutions, first you generate lots of ideas and explore as many alternatives as possible in an efficient, lightweight way, giving yourself the greatest chance at finding a new approach, combination, or improvement that may turn out to be wildly successful in the marketplace. Think of it as a numbers game. Your goal is to give yourself the best odds of discovering a promising new idea. Generating alternatives is the essence of this stage of the Fast Feedback Cycle.

Think about an architect who is helping you remodel your kitchen. Would you be satisfied if he drew only one floor plan for you to consider before construction began? Or would you expect to consider a few options, to see what happens if you move the door or put the stove in the corner instead of on the island or add a second sink? Architects have learned through experience that it’s worth the effort to explore multiple ideas on paper, iterating a few times to ensure that both they and their clients have thought through as much of the plan as possible and thoughtfully considered several viable alternatives. Once the contractor starts building, the cost to change your mind skyrockets. The value of considering lots of alternatives up front when remodeling is obvious.

Is writing software so different? Yes, software is infinitely updateable, and you can make design changes, try different approaches, fix bugs, and often push updates almost instantly. However, those updates aren’t free, especially when you consider all the time spent not just by development and testing but also in design and deployment, plus the impact on downstream operations, product support, and legacy support. Also, adding functionality is one thing, but taking a feature away can upset some of your customers, and even if they are only a minority, they may be quite vocal. Adding hardware devices into the mix complicates things dramatically, with physical constraints that dictate longer manufacturing timelines and hard limits to what you can change later in firmware or software.

One nightmare scenario is that you discover very late in the cycle that your architecture is fundamentally flawed. Or, after having invested in a hardware design, you find that customers think the device is simply too big. Or perhaps you discover that customers just don’t seem to like your approach enough to switch from their current provider to your new service. You definitely want to avoid those nightmare scenarios. How best to do it?

Software development, too, can benefit from considering alternatives up front in a cheap and efficient way. Exploring and comparing alternatives can help you better predict which approaches are most likely to work and help eliminate the weaker options before you invest too much time. When you have only one solution under consideration, by definition it’s your best idea. But once you have a few different ideas, the benefits and deficits of each approach come into sharper focus. Furthermore, getting feedback from customers not just on one idea at a time but on several that customers can compare and contrast helps make customer feedback more effective as well.

The power of blends

Remember the photographs of the three-dimensional mouse prototypes in Chapter 3, “Take an experimental approach”? One important phenomenon the mouse example highlights is that at every stage, the team didn’t choose specific designs to move forward with in their design process. Rather, at each stage the team combined and recombined the best ideas from multiple prototypes based on feedback from users, while also considering hardware manufacturing constraints, and it used that feedback to inform its next generation of prototypes. This combinatorial mixing of ideas allowed the most promising aspects of a proposed solution to be retained, while swapping out other aspects that might not be working as well.

Many teams we’ve worked with have also commented on how the best ideas emerged from blending different approaches, and that they never would have considered that blend had they not taken the time to generate lots of alternatives. Looking for blends is a central approach behind implementing the innovation pattern we called “combining ideas” earlier.

Image Vocab

A blend is a combination of different aspects of multiple solution approaches that together create a new approach that is markedly better than any of the original alternatives that were considered.

Successful blends are all over the place. Consider Skype: it created the hybrid technical solution of a peer-to-peer infrastructure that keeps video chat bandwidth off its servers but still uses a web service for centralized services such as authentication. Or consider Snapchat, which combined existing technologies for sharing photos, video chat, and instant messaging in a unique way, making photos as ephemeral as a video chat stream. The best ideas don’t often leap fully formed out of one person’s head. Rather, they are more likely to be partly your idea and partly the ideas of others, plus other bits that no one on the team originally thought of but are obvious once you connect ideas together this way. By generating lots of alternatives, and spending some time wrestling with them, you give yourself a better opportunity to find ways to mix and match ideas to discover unique solutions like these.

If you remember back to your math classes in college, it turns out that setting yourself up to create blends of ideas is particularly important when you think about innovation in terms of finding mathematically optimal solutions. Read on.

The math behind innovation

Our intuition as software engineers often leads us to believe that if we find a good working solution to a problem, all we need to do is continually iterate and improve on that working solution to eventually reach the best, most-optimal solution. This is one area where the Agile philosophy of choosing the simplest possible solution for any problem can run you aground unintentionally—it is a reasonable heuristic that the simplest solution is often the best place to start, and we certainly don’t want to overengineer for hypothetical future flexibility or architectural elegance. Sometimes, however, the simplest technical solution creates a suboptimal user experience or carries other hidden tradeoffs, and you may not realize those downsides until much later if you aren’t looking out for them.

Your instincts might lead you to imagine the problem space to look something like what’s shown in Figure 7-1.


FIGURE 7-1 A graph of a single, optimal solution in a complex, three-dimensional plane.

You imagine that all possible solutions are on the path to the one optimal solution. You figure that once you get going on building something, you’re somewhere on the foothills of that mountain peak, and as long as you keep iterating and improving your solution, eventually you’ll reach the top—that is, you’ll land on a globally optimal solution. Engineers are natural iterators who constantly seek to improve their solutions by fixing bugs, adding functionality, integrating more partners, improving error handling, and so on. So, by and large, engineers do climb whatever hill they’re on and make improvements little by little. However, most engineers typically iterate only one solution idea at a time, and that’s where the trouble starts. The reason that iterating only one solution at a time is troublesome is that most problems are complex and multifaceted, and you’re trying to solve for many variables at once. The landscape of possible solutions is much more complex than you often realize at first. There are actually many good solutions to any given problem, perhaps even more than one great solution, and those solutions may be quite different from one another both in concept and in implementation. Each potential solution may trade off key aspects of implementation elegance, performance, scalability, usefulness, usability, and desirability in dramatically different ways.

Consider a much more complex, three-dimensional mathematical plane, Figure 7-2, where the good solutions to a given scenario are represented by local maximums—small hills on the contour diagram.5 There are also a handful of truly great solutions that have much higher peaks, and one outstanding approach that is the clear global maximum. If you start your project with a reasonably good, workable approach in mind, you can gradually improve your solution as you iterate and will likely climb to reach one of those small peaks, or local maxima. If you are very lucky, you might happen to start off on the foothills of that highest peak and reach the global maximum—because you picked a good starting point, or, in mathematical terms, a good seed.


FIGURE 7-2 A graph of multiple possible solutions, some more optimal than others, in a complex three-dimensional plane.

But what if you are not so lucky? How do you give yourself the best odds at iterating toward the most globally optimal solution and not getting stuck at the top of a much smaller hill? Rather than enthusiastically rushing to code up the first good approach you come up with, what if you start off by exploring and iterating several diverse approaches, trying out ideas that represent several different regions of that plane? What if you combine and recombine the best aspects of those ideas to help you identify which neighborhood of solutions is most globally promising and provides the best balance of tradeoffs for your situation? As in mathematics, the more diverse a set of seeds you start your iteration with, the better the odds you have for discovering a more globally optimal solution in a complex universe.6

Why it’s so hard to find a global maximum

Dr. Craig Wiegert, Associate Professor of Physics, University of Georgia

The task of optimizing a function (finding its maxima or minima) is ubiquitous in the sciences. For example, nature strongly favors minimum energy configurations of a system. So, chemists use this principle to describe molecular bonding in terms of the arrangement of atoms that minimizes the total energy of the molecule. Physicists model the physical properties of crystalline solids and amorphous glasses by seeking the lowest energy structures. In biology, protein chains spontaneously fold into their functional shape as the molecule lowers its overall energy in a complex aqueous environment.

Finding a local optimum is relatively easy, at least conceptually speaking. From a starting point, just go downhill (or uphill) until you can’t go any farther! (This method is known, not surprisingly, as “hill climbing.”) Algorithms such as gradient descent, the conjugate gradient method, and others generally improve the computational efficiency by carefully choosing directions in the search space. Many of these algorithms assume that, near the optimum, the function behaves quadratically (like a multidimensional parabola).

In most situations, though, it’s important to determine the global energy minimum, the one that nature finds seemingly without effort. A local minimum might have almost the same value as the global minimum, but that’s often not good enough to be useful. There’s also no guarantee that this local minimum will be anywhere near the global minimum in the search space. That’s especially problematic because even a small difference in the predicted atomic positions in a crystal can dramatically change material properties such as tensile strength, conductivity, or ferromagnetism.

The bad news for us all is that it’s mathematically impossible to find the global optimum of any arbitrary function. In other words, any optimization algorithm, no matter how general, will fail on some classes of functions. The slightly better news is that, under certain assumptions about the function—for example, that it’s smooth, bounded, and/or in a finite-sized search space—finding its global optimum becomes merely fiendishly difficult.

Much of the difficulty relates to the size of the search space. Each new coordinate or degree of freedom in the system adds another dimension to the search space. Consider a simple molecule such as water. To fully determine its molecular structure and properties from first principles, you need to minimize the energy function not just over the three atoms’ positions, but also all 10 of the electrons. Naïvely, that’s a 39-parameter search space (13 objects x 3 spatial dimensions)! Accounting for symmetries, and ignoring some of the nonbinding electrons, allows a considerable reduction of the dimensionality; that still leaves two atomic degrees of freedom and eight electrons that need to be optimized. Now imagine the computational complexity of putting thousands of molecules of water together and investigating their molecular interactions. This is why physicists still don’t have a complete theoretical understanding of liquid water.

Without a universally applicable method of finding global optima, practical optimization procedures in huge search spaces rely on casting a wide-enough net. Some algorithms, like the Nelder-Mead simplex method, are deterministic but relatively inefficient. Many others are stochastic; they increase efficiency at the expense of introducing randomness into the search process. These include algorithms like simulated annealing (where a large initial search is gradually “frozen” into smaller and smaller regions) and evolutionary algorithms (where the best candidates from many small regions of the search space are remixed into new candidates). Fundamentally, all of these approaches are based on the same principle: evaluate a large variety of starting points across a complex search landscape, and narrow in on the optimum based on those results. Starting the optimization with only a few points, or with points that are too similar to each other, practically guarantees a nonoptimal solution.

Even for mathematicians and physicists, no matter which method they use, finding a true global optimum is a genuinely difficult endeavor. Thankfully, even though the software problems we work on are complex, finding the single, absolute maximum may not be as critical for us. In fact, there may be several comparable maximums, and practically speaking, we’d be pretty happy if we found any one of those truly great solutions and successfully avoided getting stuck on a smaller hill. To more reliably find a great, optimal (or nearly optimal) solution, you can take a lesson from the experts. To find optimal solutions in complex spaces, start out by casting a wide net and try multiple seeds as you narrow in on an optimal solution.

It’s so easy to barrel down the wrong path, quickly picking an approach that looks reasonable, coding it up, and releasing it, only to find yourself fixing bugs and responding to customer feedback and pushing updates constantly, working hard but never realizing the success that seemed imminent with such an auspicious start. Later, when a competitor comes out with a variant that ends up gaining the popularity you hoped for, you may realize that the root problem was something more fundamental in your approach—the underlying architecture that defined the core performance characteristics, the user experience that needed to have been simpler and more streamlined, a key aspect of the customer’s end-to-end experience that you didn’t notice, the developer whose apps ended up driving customer adoption and purchase, or that you did a great job solving the wrong problem.

Looking back, the salient postmortem question often is “Did you ever consider alternative approaches that might have led to something similar to your competitor’s winning solution?” Sometimes the answer is yes—that an intentional decision was made not to follow that path, and that will happen from time to time. But all too often the answer is no—that approach never occurred to you—you went down the first viable path that seemed to meet the customers’ needs and didn’t notice that there was an alternative to consider. Without a broad-enough set of inputs, no amount of iteration will lead you there. That sort of failure lies not in the iterative process, but in not considering a broad - and diverse-enough set of alternatives from the very start.

It’s no surprise why people used to say wait for v3 before buying a Microsoft product. By the third version (that is, after the third cycle of iteration) Microsoft had finally explored enough of a complex solution space to find a more globally optimal solution—or, at the very least, it had reached the actual peak of a local maximum. These days, however, you want to get the iteration done well before the customer experiences a final solution, whether it’s version 1 or version 15. Furthermore, in today’s marketplace, customers demand not just good solutions but great ones, and they don’t want to wait years to get there. So the bar is set higher than ever before.

The exploration stage is about giving yourself choices. You will never look as broadly at solution possibilities as you will at the start of a project, so take advantage of that time to generate lots of diverse ideas and explore the space of possibilities. Later in the project, you will naturally narrow your thinking and your choices. But at this early stage, you should explore as wide a variety of alternatives as possible and give yourself the best possible odds at combining them into a unique approach that leads to a globally optimal solution that will bring the success you’re hoping for.

The problem with tunnel vision

The first step in exploration is to generate a lot of ideas. If only it were that easy. You see, there is a problem. The human brain is hardwired to lean in the opposite direction.

Let’s illustrate with a story. Can you remember a time when you had a plan for something, you had written a spec perhaps, or worked out an architecture for a new component, or sketched a design for a new backyard deck? While you were pretty happy with your plan, you had a niggling thought in the back of your head that maybe that plan wasn’t perfect, maybe you were forgetting something, maybe there was a better way. Perhaps you even sat down with a blank piece of paper and told yourself “Let’s come up with a better idea,” but all you could do was stare at the blank piece of paper?

We’ve all been in that situation at one time or another. You just can’t seem to break out of your current mode of thinking to produce something new. The problem is that once you have a solution in mind, even if it isn’t a particularly great solution, it’s really, really hard for your brain to generate other alternatives. The longer you’ve thought about that solution, and the more you’ve explored it in detail, the harder it is to even imagine the possibility of viable alternatives.

What’s going on? This isn’t a personal character flaw, and additional time or brute-force effort will have little effect. It’s actually a well-understood side effect of how our brain works, specifically how it learns. We call this phenomenon tunnel vision. Sometimes we refer to a person, or even a team, as being “stuck in the tunnel.”

Image Vocab

Tunnel vision is a neurological phenomenon where once you believe you have solved a problem and have a particular approach in mind, your brain becomes blind to the possibility that alternative solutions may exist, and your ability to generate new ideas becomes greatly compromised.

Neuroscience break by Dr. Indrė Viskontas

In the 1960s and 1970s, neuroscientists named Terje Lømo and Tim Bliss discovered that when two neurons connected by a synapse are stimulated at the same time, they become associated. This means that when one fires in the future, the other is more likely to fire, too. We say that the synapse has been “strengthened” because the connection between these two neurons is stronger.

You might have heard the adage “neurons that fire together, wire together,” which is short-hand for this type of synaptic plasticity. Let’s say that there’s a set of neurons in your brain that turns on, or fires, when you see the refrigerator door. And there’s another group nearby that fires when you see any type of food. Now, whenever you open the fridge, both sets of neurons fire because you see the door and you see the food at the same time.

Over time, these two groups of neurons fire together often enough that the connection between them, or the synapse, gets stronger. Just seeing the fridge door is enough to fire your food-detector neurons. The activity of these neurons is now connected, and the fridge door is associated with food.

Connecting cells through their wiring, and making these connections more efficient, is the very essence of learning: it’s how the brain changes with experience. The capacity of neurons to connect to each other gives us the ability to learn, but it also has a downside. Much of what we learn over time gets ingrained in the wiring of our brains.

A common analogy for this side effect of our brain’s wiring is to think of the brain as being a beautiful, pristine sledding hill, full of fresh, fluffy white snow. You are sitting in a sled at the top. The first couple of times you head down the hill, it’s quite easy to lean this way or that way and cut new paths in the snow. However, once you have taken the same path in your sled a few times, a rut forms, and it’s quite tough to get your sled to go anywhere else but in the rut. This is much like what happens at a neurological level. When the brain encounters a novel challenge, it can easily consider multiple potential solutions, recruiting different neural networks to the task, and it can quite flexibly consider different approaches. However, the longer you think about a particular solution, the stronger that particular neural pathway becomes, and the less likely it is that you can jump out of that neurological “rut” to see a different alternative.

Image Vocab

Fresh snow is the mental state most conducive to creativity, when you haven’t yet formed a clear opinion or decided on a solution to a given challenge. The term stems from the analogy of the brain being full of fresh, fluffy snow before a rut, or tunnel vision, forms.

As you can see, tunnel vision happens for a predictable neurological reason. When you start thinking about a particular idea a lot, you “learn” it—in neuroscience terms, the neural pathway gets reinforced every time you use it. Your brain can easily be convinced that there are no other options because the one you hold in your mind fits the pattern so well, and the more you think about it, the more ingrained your opinion becomes. Once that neurological rut is carved sufficiently deep, it becomes increasingly hard for your brain to generate a different idea, a different approach, a different solution. For some people, tunnel vision sets in faster than for others. Other people seem to tolerate ambiguity a bit longer. But no one is immune to the physiology of the brain.

Image Mindshift

Tunnel vision feels so good. One reason we are likely to fall so quickly into tunnel vision is that having a firm plan in mind is a powerful and satisfying feeling, especially because it replaces the much-less-pleasant feelings of ambiguity and anxiety about an uncertain future, which are inherent in the early stages of a project. Once we feel like we’ve hit upon the “perfect solution”—or even what looks to be a workable approach to a challenging problem—we can exhale, telling ourselves that it’s all downhill from there. We quickly get caught up in the details of thinking through exactly how that approach might be implemented, naturally gravitating to our strengths in technical design and development and working through details. Tunnel vision can feel downright intoxicating, and it’s easy to bring a team along with you, because everyone is eager for the clarity that a definite path forward brings.

In fact, the siren song of tunnel vision is so incredibly tempting that it can easily cause a team to entirely skip the first half of the Fast Feedback Cycle—clearly identifying their target customers, understanding their customer’s unarticulated needs—and go directly to brainstorming product ideas without any regard to what problem the team is trying to solve, what is going to be desirable for its customers, or its business strategy. Skipping these steps is even more dangerous than not generating enough alternatives, because you will likely end up building a solution for a problem that your customers may not care about or for a fictional customer who doesn’t exist in the real world.

Whatever the cause, remember that if you have zeroed-in on a solution approach too soon and left the complex plane unexplored, odds are that barreling down that well-worn rut in the snow will land you at a local maximum, not at the game-changing optimal solution you might have aspirations for.

Tunnel vision happens. It just does, it is part of the human condition. People and teams who understand what tunnel vision is, how it manifests, and how to mitigate it are able to generate more creative, innovative ideas. Those who can’t, don’t. It’s that simple. Become self-aware, and learn to plan for and mitigate tunnel vision.

Mitigating tunnel vision

Your job is to predict that tunnel vision will happen and not let it surprise you. If you recognize that tunnel vision is inevitable because it’s a fundamental property of our brain wiring, your best course of action is to delay tunnel vision as long as possible and take advantage of “fresh snow” by using the time when your brain is most able to think creatively to generate and explore a wide variety of alternatives. Remember that your instincts may be telling you to dig deep into the first promising idea you discover. You need to develop substantial discipline to combat those instinctual urges and delay tunnel vision from setting in before you are ready.

Neuroscience break by Dr. Indrė Viskontas

Have you ever noticed how young children, those under the age of five, can find very creative uses for everyday objects? A yogurt container becomes a drum. Mom’s eyeglasses can double as a toothpick. A wall signifies a blank canvas, waiting to be drawn on. Then, by about age seven or so, as kids learn what the traditional purposes are for different objects, they lose the ability to dream up new functions. This is the point in their development when they begin to show evidence of a common cognitive bias: functional fixedness. Functional fixedness prevents us from finding novel uses for everyday objects—once we conceive of a tool’s purpose, it’s very hard to develop alternative ways to engage with it.

This bias likely helps us learn about our world, but it also can hamper creativity. The classic experiment during which the term was coined involved asking participants to solve a peculiar problem: they were given a candle, a book of matches, and a box of thumbtacks and asked to affix the candle to the wall such that no wax would drip on the table below it. The solution requires literally thinking outside the box: the thumbtacks have to be removed from the box so that the candle can be placed inside it, and a thumbtack then is used to pin the box to the wall.

Most participants failed to find the solution, but when they were presented with an empty tack box, the solution came easily. They needed a nudge to consider an alternative use for the objects. By the same token, when we generate only one potential solution to a problem, we tend to fixate on it—at the expense of other creative options. Generating a series of solutions is a great way to overcome fixedness and the functional equivalent of writer’s block.

You can’t delay tunnel vision forever, and indeed you benefit in other stages of the Fast Feedback Cycle from being in the tunnel, which helps you focus on the task at hand without constantly reopening decisions or questioning yourself. However, you don’t want to go there until you have fully explored the space of alternatives. Here are some tactics you can use to delay or mitigate tunnel vision until you are ready for it.

Brainstorm early

The first tactic to mitigate tunnel vision is simply to recognize that there really is a time when you are at your most creative. Your brain will be dramatically more capable of generating a broad variety of alternative solutions in the very beginning of a project, before the tunnel begins to form.

Beware that some teams want to wait until the “right time” to brainstorm solution ideas—until all the right people can attend a key meeting or until a particular technical constraint is decided. There is truly a right time to brainstorm, but waiting is most likely the wrong call. Once you have identified the next set of customer problems to solve, whether in your first iteration or your fifteenth, your brain will start spinning on possible solution ideas and that will bring on tunnel vision, whether you intended it or not. It is essential to take the time up front to generate lots of diverse alternatives and go as broad as possible when you and your team are physically and psychologically most able to be creative.

Quantity, not quality

As Linus Pauling said, the goal at this point is to generate a large quantity of ideas. However, one of the quickest ways to turn off the flow of creativity is to start deciding whether the ideas you’re generating are any good. As soon as someone hears that an idea he suggested “will never work” or “the customer wouldn’t like that,” it makes that person less likely to contribute again. It is so easy to unintentionally dampen or completely extinguish the fire behind an idea-generation exercise, simply with poor body language and offhand comments.

The problem starts when someone offers an out-of-the-box idea, which may understandably sound silly or impractical at first. (Remember that every new innovation probably sounded pretty crazy at first.) If the team accepts that unusual idea and builds from it, sometimes magical things can happen. But unfortunately, it’s quite likely that a team member might roll his or her eyes or use other body language to unconsciously or subconsciously say, “That will never work” or “That’s a stupid idea.” Without really intending it, a team member might say something out loud that unwittingly carries the same message, such as “Anyway . . .” or “Moving right along . . .” Despite the usually neutral intent, the negative subtext is understood loud and clear by the recipient as well as by the other participants in the room, and can significantly reduce people’s willingness to take a risk and contribute to the conversation, especially to share that wacky out-of-the-box idea rolling around in their head.

Image Tip

“Yes, and . . .” is a popular technique borrowed from improvisational comedy that can be very useful during idea generation to ensure that all ideas are valued and encouraged without judgment, either explicit or implicit. The idea is for each person to say “Yes, and . . .” at the beginning of every comment they make or any new idea they contribute. When consistently used in the context of a brainstorming activity, these two small words can make a world of difference in instituting a welcoming and nonjudgmental environment that encourages every team member to feel comfortable and participate fully.

The antidote is to make sure that every idea is welcomed and celebrated, no matter who says it, what it is, or where it comes from. There will be plenty of opportunities to sort through the ideas later. For now the sole focus is on generating lots and lots of ideas, regardless of how crazy, unfeasible, or terrific they might turn out to be. At this moment, focus exclusively on quantity, not quality.

Breadth before depth

Another key tactic is to mindfully seek to generate a wide variety of ideas, exploring as much of the width of the solution space as you possibly can before talking about any ideas in depth. Think of it as doing a breadth-first search through the space, knowing that you can go back to think through an idea’s depth later. Keep asking yourself, “How else could we solve this problem?” Your focus right now is purely to maximize the variety of ideas generated.

Imagine that as you generate ideas, you’re creating a long, long list of pointers, with just bare minimum information attached, to help you remember what you were thinking as a jumping-off point. Your list may contain words and phrases, or it may also include visualizations like sketches and rough storyboards. You will come back to this long list several times in future iterations of the Fast Feedback Cycle and have many chances to continue brainstorming at a deeper level to explore the details of the most promising of those ideas.

Go fast

The trick to doing a breadth-first exploration is that you can’t pause very long on any one idea, especially the good ones. You have to develop the habit to write down all of your ideas (practical or not, promising or not—don’t try to decide now) and then force yourself to move on. Maintaining a fast pace helps to keep ideas fluid in your head, and you have a better chance of sparking new connections between different ideas when they are separated by seconds and not minutes.

Think of it like staying on your toes in soccer. Sure, you can stand on the field flat-footed, but you’ll turn faster and react more nimbly to the game if you stay on your toes, ready to move at any moment. For this kind of exercise, intentionally go fast. Some teams even set explicit goals to generate a certain number of ideas per hour to help keep the pace up. Between 60 and 100 unique ideas per hour is considered a good rate for a group brainstorming session.

Keeping a fast pace is particularly difficult when someone comes up with what appears to be a brilliant idea . . . which leads us to the next tactic for mitigating tunnel vision.

Don’t fall in love with your first good idea

You’ve just encountered a difficult problem that needs a creative solution. You understand that tunnel vision is going to set in soon, so you’ve kicked off a quick brainstorming session. After 10 minutes of slogging through a quagmire of irrelevant, impractical, and uninspiring ideas—BAM!, lightning strikes, and someone comes up with an approach that appears to be unique, exciting, pragmatic, and certain to delight your customers. What should you do?

Write it down and move on, asking yourself, “How else could we solve this problem?”

That promising idea won’t go away. There is no reason (other than to invite the onset of tunnel vision) to linger on that idea or to start thinking about its details, implications, viability, or constraints. Just write it down with all the other ideas and take comfort in the fact that you have captured it—it won’t be forgotten, and you can come back to dig into it sometime later on.

But this is much easier said than done. This is exactly the situation in which your instincts will mislead you. Every bone in your body will want to chase this beautiful idea and discuss the possibilities—the way it might work, how it might look, what implications it might have for the architecture. The problem is that if you do that, tunnel vision will rapidly kick in. And once that happens, you will be much less able to generate any more alternatives, especially not meaningfully different approaches. In addition, it’s very likely that an entire roomful of people will be brought into the tunnel with you.

Consider this. What if you discover that the first awesome idea you fell in love with doesn’t work out for some reason? (You know, it’s not as good as you thought it was at first . . . which is almost always the case with first ideas.) Because you didn’t generate other viable alternatives yet, you have no options to fall back on, and you literally have to start from scratch. And starting from scratch will be extremely difficult this time, at least with this group of people, because you are all already in the tunnel. Likely, you will move forward with that idea and won’t realize until quite a bit later that it has some fatal tradeoffs or that there was a fundamentally better approach that you never considered, but by then it will be too late (or at least very expensive) to change.

Avoid this problem by developing the discipline to write down the good idea and move on. Over time, you and your team will form the habit of doing this and it will become easier and easier. We’ve seen teams develop a culture and vocabulary around this approach. When they come across this situation, we hear people say things like:

“Great idea. Write it down. We’ll think about it more later.”

“Let’s take advantage of the fresh snow. How else could we solve this problem?”

“Awesome. Love it. What else?”

“Yikes—tunnel vision is setting in . . . let’s move!”

Instead of derailing the brainstorming to go deeper on that first great idea, you have to mindfully postpone that thought process until after you have done a complete breadth-first pass. As you take notes, don’t write down a lot of details and don’t have a lengthy conversation about every idea. Just capture enough of the idea that you can trigger your memory later. Aim for a sentence or a phrase or a very rough sketch; a paragraph is probably too much.

SFE in action: The algebra of ideas, or how two dumb ideas make a great idea

Bob Graf, Senior User Experience Researcher, Microsoft Engineering Excellence

Sandy, a program manager on my team, walked into my office to get input on the design of his feature from my UX perspective. Instead of just signing off on it, I asked him numerous questions about the users’ roles, responsibilities, goals, and needs. I also asked him to describe our business needs and strategic needs and why the feature was important.

After our conversation, I was able to say with confidence that his design addressed everything we talked about. However, it was only one possible solution. I told him we should generate six ideas to make sure we explored all options.

I sketched his design on the whiteboard and labeled it “#1.” Only five more to go. The next two ideas came fast, but then we got stuck. I looked at him; he looked at me. We shrugged. Stuck! I said we couldn’t leave the office until we generated three more. To reach that goal, I told him I had a dumb idea and sketched it on the board. Two more to go. That dumb idea led to another, even dumber one. I personally created a safe environment for us to think creatively and without censoring our thought processes with premature evaluations.

Suddenly, magic happened. As we both looked at the two dumb ideas, the best solution simultaneously leaped into our minds. I quickly sketched it on the board. We felt we had the best solution, but we followed up with a pros-and-cons evaluation for each of the six possible solutions. When we were done, we knew idea #6 was the superior solution we needed to build. We would have never arrived at it had we not made a commitment to generate six ideas and allowed each idea to spark new ones.

I can summarize our experience into three principles for successful design thinking:

Image Establish a safe, open environment, where free thinking is encouraged.

Image The quantity of ideas is essential; there are no bad ideas.

Image Keep interactions lightweight in both process and tone.

Employing these principles allows magic to happen.

It takes a village

It is unreasonable to expect that any one person can fully explore any solution space on his or her own. That is why many techniques for generating and exploring ideas involve multiple people in one way or another. It turns out that harvesting thoughts and energy from multiple brains—and perhaps more important, multiple points of view—is more than just a way to get out of a rut, it is a valuable and necessary ingredient to the creative process.

Ideally, you involve people with as much diversity as possible—consider different skill sets, different personal experiences, different career trajectories, different personality types, different backgrounds, different ages and life stages, different problem-solving styles, and even different motivations for solving the problem in the first place. Having people bounce ideas off one another, stimulating each other in unpredictable ways, is a key part of the formula for keeping ideas flowing and avoiding becoming stuck in a rut prematurely.

Team members also have a role to keep each other honest, to keep the discussion moving and not linger too long on any one idea. It’s much easier to see when a teammate is drilling too deeply into an idea (and risking getting stuck in the tunnel) than it is to see that in yourself. Find a nonjudgmental way to communicate that it’s time to move on. Anything from tossing a red bandanna to the offender, tagging them with a Nerf toy, or empowering anyone to cut off a discussion with a scripted reminder such as “too much detail” or “How else could we solve this?” can work. Define the plan ahead of time so that everyone knows the rules, and keep it positive, playful, and not confrontational.

Help! I’m in the tunnel

Now that you know about tunnel vision, you need to be vigilant and self-aware. Once you find yourself experiencing that intoxicating feeling of having found the perfect solution, you need to recognize that you may genuinely and firmly believe that no other alternative approach exists—even if one is plainly visible to others. If you have already generated plenty of alternatives, feel confident that you have explored the solution space, and have captured those ideas in a form that will trigger your thinking later, this is not necessarily a problem. But if you suspect that you have entered the tunnel too early, without having yet explored the space fully, or have simply run dry of ideas before you feel that you’ve generated enough, you need to find a way out of the tunnel. Here are some techniques that are pretty broadly useful when you need to get out of the tunnel.

One of the best steps to take is to solicit different ideas from a new set of people. Recruit people who have not yet been infected by your idea and use their “fresh brains” to see things that you may not. They may help you see alternatives, blends, and improvements that you likely would not find on your own. Even a few seemingly random new ideas can help unclog your thinking so that you notice alternatives you were not aware of before. Usually you can productively riff on an idea once it is suggested by someone else, even if you are too far in the tunnel to notice that alternative approach on your own.

Image Vocab

Fresh brains? Sorry, this isn’t a reference to the latest zombie craze. We use the term fresh brains to refer to people whose heads are still metaphorically full of fresh snow. They have not been exposed to your current thinking about the problem space and are better able to see ideas that you may not.

The first step to getting out of the tunnel is to acknowledge that you are in fact stuck. Something like this: “Hey gang, we’ve been spinning around this same idea for a while now. I think we’re in the tunnel. Let’s go find some fresh brains and see if we can get unstuck.” Teammates who have not been deeply involved in your project are a good place to look for fresh brains. They are usually easy to find simply by walking down the hall or picking up the phone. A five-minute conversation by the water cooler may be all it takes to find a new perspective that gets you out of the tunnel.

Beware of limiting yourself to tossing around ideas only with others on your immediate team. You may readily convince each other that no other viable alternatives exist because you are all stuck in the same tunnel. It’s very easy to bring teammates into the tunnel with you.

Even when you find some new people to brainstorm with, be sure you follow the rules and encourage new ideas. Don’t try to sell or validate the ideas that took you into the tunnel in the first place. You are still looking for quantity, not quality. You are looking for a new approach or blend of approaches that will get you out of the tunnel and able to explore new paths of thinking again.

An even better source of fresh brains is your customer base. Customers will almost always approach problems from a perspective different from yours. It is a good general practice to generate ideas with customers purely because of the different perspective they bring, which can be hard to replicate even with a diverse team.

Good ideas come from the strangest places

Regardless of the techniques you use to generate ideas, predictable patterns occur. In our workshop, to illustrate how these patterns emerge and what they look like, we conduct a little contest. It turns out that everyone loves a little competition, and the brainstorming contest really gets the post-lunch energy flowing again.

The room has already been set up with tables for four or five people, ideally with representatives from different disciplines at each table. We ask each table to nominate a scribe, and for a little added motivation we mention that a prize will be awarded to the winning team. The table that comes up with the greatest number of unique ideas wins. We set a timer for one minute and then announce:

Ready, set . . . the focus of the brainstorm is:

“Name as many ways as you can to send a message to another person.”


Immediately the energy in the room skyrockets. People shout, laugh, think, and write frantically. This is brainstorming. For most people, it’s fun, engaging, exhilarating. But to use it as a productive and efficient ideation tool, it takes practice—intentional practice by the entire team, and not everyone is good at it at first.

After one minute we call time and ask people to count how many ideas their group came up with. We don’t care whether the ideas are any good, it’s just the raw number that matters right now. The average number of messaging ideas each table generates is about 20. The most ideas we’ve ever seen produced in a minute is 42—more than double the average. We’ve also seen plenty of tables barely get to a dozen in a minute’s time; these teams get stuck on an idea and can’t seem to get past it. We figure out which table came up with the most ideas and then announce the prize: the winners get the honor of reading their list of ideas to the group.

Here’s an example of a typical winning list of ideas for how to send a message to another person:

Email, letter, text message, skywriting, smoke signals, airmail, carrier pigeon, Morse code, radio, television, webpage, discussion board, facial expression, frown, smile, spitting, walking away, handshake, kiss, hug, song, poem, ballad, haiku, limerick, gift, present, gift card, phone call, Skype, flaming arrow

Diverse people create more diverse ideas

While you read through the list, did one of these entries trigger a new idea for you? Chances are that it did, and you have a couple of new contributions you could add to the end. Your ideas, in turn, might trigger additional ideas for other examples that hadn’t occurred to you. That’s the magic of brainstorming with multiple people, each coming from a different perspective and bringing multiple skills, life experiences, and backgrounds to help maximize the number and variety of alternatives generated.

In the classroom, people are always astounded at how many different ideas their own table came up with that are not on the list that was read out loud. It is an experiential reminder that different people will often generate very different ideas, even given the same problem statement.

Embrace the cousins

Do you notice any patterns in that list of ideas, and the order in which ideas are mentioned? The first three ideas—email, letter, text message—are all centered on typed or written text. It’s very common for ideas to come out in spurts like this during a brainstorming session, producing a set of related ideas that are “cousins” of each other.

Image Vocab

Cousins are a succession of similar ideas mentioned during an idea-generation session that all are closely related to one another.

Some people have trouble with cousins. Inwardly, they feel that a “cousin” idea might not be different enough from other ideas on the list to be worth saying out loud. For instance, in our example, a person might hear “email, letter, text message” and think about contributing “SMS,” but she might second-guess herself because, well, isn’t SMS just the same thing as a text message? This doubt is your inner perfectionist peeking through, worrying that someone will call you out for contributing an uninspired brainstorming idea, one that doesn’t seem new or different enough from what is already on the list. You feel like you are cheating somehow.

Truthfully, you are better off saying your idea out loud no matter how similar or dissimilar it is. Otherwise, you’ll spend a lot of mental energy censoring yourself. This makes it hard to stay in the fast-paced flow of the brainstorm so that you can keep riffing on ideas and contributing your unique perspective. Furthermore, saying “SMS” might have triggered someone to think of MMS messages, which might have led to sending videos, photos, Snapchat, photobombing, etc. (None of which are mentioned on the list, did you notice?)

You can make forward progress in your brainstorming with even the closest of cousins. And worst case, even if it goes nowhere, so what? You used only a few seconds of time, big deal. Bottom line: if the idea pops into your head, say it; don’t self-censor. It’s better for you, and it’s better for the team’s results overall.

Encourage lateral jumps

When we get to the end of a run of cousins, there’s usually a short pause. Then, out of the blue, a new idea—skywriting—is suggested. Where did it come from? Note that skywriting is followed by another set of cousins: smoke signals, airmail, carrier pigeon. And when this run of cousins ends, there’s another pause, and then a jump to another completely different line of ideas: Morse code, radio, television.

This pattern is typical of a healthy brainstorm. One idea starts a stream of thinking along a particular vein. When that vein runs dry, a different thought is suggested and it’s followed for a while. There may be a pause between streams of thought, and then there’s a jump to a new topic. Those jumps are called lateral jumps. They don’t take a line of thought forward; they jump sideways, or laterally, to a different perspective on the problem or solution.

Image Vocab

A lateral jump is a shift to a new topic while brainstorming that is seemingly unrelated to the previously generated ideas.

It turns out that lateral jumps are very powerful. The more lateral jumps that occur during idea generation, the more diverse a set of options you’ll end up generating. The more that you can cultivate an environment that encourages and stimulates lateral jumps, the more ideas you will generate and the more likely that you will hit upon the raw materials for the winning combination you are searching for.

Suspend disbelief about wild ideas

One of our favorite techniques for encouraging lateral jumps is called challenge assumptions. The idea is to challenge a basic assumption about your problem and see whether by removing that constraint you can find an alternative through a side door that you might not have noticed while looking at your problem head-on.

For example, what if we suggested that you design a new kind of coffee cup, but we challenged the assumption that it would have a bottom. What would a coffee cup look like without a bottom? It’s a basic assumption that all cups have some sort of a bottom. How would it stand on its own if it had no bottom to rest on? Your first instinct is to reject the whole notion and say, “That’s impossible!”

But stay with us for a second. Suspend disbelief and think about it. What might a coffee cup look like if it had no bottom? What if it was a different shape? A rounded bottom, like a child’s tip-free sippy cup? Perhaps a sphere? Or maybe a cone shape? A spinning cylinder that leveraged centrifugal force? An antigravity vortex? Maybe some sort of device that acts on vacuum pressure? What if it looked like a bag of intravenous (IV) fluids from a hospital—imagine, an IV drip of your morning coffee! Or what if the cup was like an IV bag but with a straw attached to the bottom? Hmm, I guess there might be some alternatives to consider after all.

But can you make the leap from there and imagine inventing the CamelBak, a water-containing backpack with a long straw meant for keeping you hydrated on hiking outings? Or perhaps you might leap in a different direction and imagine the beer hat, which comfortably perches two cans of beer on your head, with a long straw that lets you sip at will?

Both of these are actually pretty useful, intriguing concepts and have enjoyed commercial success in their time. How likely is it that you would have come up with these kinds of ideas just thinking about alternatives for regular old coffee cups? Forcing yourself to challenge an assumption allows you to think about the problem differently and open yourself to a different class of solutions you might not have noticed otherwise.

This example illustrates the importance of being playful in generating ideas, of not taking anything too seriously or at face value, and being willing to suspend disbelief and assume that, for just this one moment, anything is possible. Being willing to consider wild ideas, playing around with them, and letting cousins and lateral jumps happen are central behaviors to cultivate to get the most out of brainstorming.

Of course, not every wild idea will be practical or even possible. Many may feel downright crazy, ridiculous, corny, or foolish. However, it’s only after following a string of wild ideas and seeing where that thread goes that you might stumble on an approach that proves to be a real winner, that is actually practical and implementable and perhaps not so crazy after all. And that idea probably would not have occurred to you otherwise. At this stage, ideas are cheap, so don’t be pound wise and penny foolish. If you’re willing to spend a few minutes of suspended disbelief to see where that wild idea might lead, you dramatically increase your chances of hitting on an out-of-the-box approach that no one has noticed before, not even your competition. And worst case? You spent only a few minutes on a dead end.

The point to remember is that you are more likely to stumble onto a unique approach through a side door rather than going at a problem head-on. This idea is so powerful that specific techniques, called lateral thinking techniques, have been developed that help spur lateral jumps in your idea generation. These encourage you to knock on those side doors and give you more chances to find new, unusual solutions. We’ll discuss several lateral thinking techniques in more detail in the tools and techniques section later in this chapter.

Image Mindshift

Who gets credit for that good idea? If you witnessed hundreds of brainstorming sessions and idea-generation activities, you would notice that you can’t predict where the best or craziest or most interesting ideas will come from. You can’t predict who will initiate the lateral jumps in the group’s thinking, and you can’t give credit to any single person who comes up with what proves to be a brilliant idea—once you realize that this brilliant idea came at the end of a long string of related and unrelated ideas, all of which created the environment that stimulated the brilliant idea to surface at all.

The notion of who gets the credit becomes even more meaningless the more that a team becomes truly collaborative and operates as a unit rather than as a collection of individuals. A strong, collaborative team shouldn’t feel like they are competing with their colleagues for who gets the bigger bonus this year. Instead, everyone has skin in the game and is pulling together for a shared goal. When a team gets to that high level of interdependence and trust, its work becomes more about getting to the right decision for the customer (and not jockeying for whose idea is the best). In turn, this attitude makes it easier to admit and recover from failures and more fluidly take advantage of the unique skills of each team member at the right time. Developing team-wide trust and deep collaboration is essential for many of the practices we discuss in this book to take root, and brainstorming is a crucial place where it becomes very clear whether your team has that trust or not.


When searching for new ideas, sometimes the best thing to do is walk away, or at least give yourself some space for a while. Marinating is a weird “anti-technique,” but it is an extremely effective concept to keep in mind when you’re generating ideas. You’ve worked on a problem for a while and perhaps gotten a bit stuck, overwhelmed, bored, or tired. Then you intentionally stop thinking about it and go do something else—work on a different problem, take a shower, go for a walk, play soccer, take a nap—and by some sort of magic, a new insight or connection or idea hits you hours or days later, when you least expect it, when you aren’t even thinking about the problem. Or when you do come back to a problem hours or days later, you find that you have a much better understanding or think of ideas that you didn’t have before.

Image Vocab

To marinate is to stop actively thinking about an idea to give your brain downtime for processing the idea in the background, making it more likely for your brain to notice unusual connections or new insights.

Most of us have experienced this at some point in our lives and have heard stories about it happening to famous people. Whether it’s Watson and Crick dreaming about the DNA double helix or Archimedes’s supposed eureka moment in the bathtub, it’s a fascinating phenomenon of the human brain that it appears to keep working on a problem subconsciously long after the conscious mind has stopped.

Neuroscience break by Dr. Indrė Viskontas

Psychologists see the creative process as having four distinct stages: preparation, incubation, illumination, and verification. Because incubation, by definition, occurs outside our conscious awareness, it’s hard to study. But there have been some interesting insights into its brain basis discovered recently by neuroscientists. In particular, sleep seems to be an important context for incubation: studies have shown that a solution to a problem, or a new idea, can often present itself after a period of sleep.

During sleep, our brains replay what happened during the day, and important information is consolidated, while at the same time, the irrelevant things that we thought about or experienced get erased. There’s even new evidence that the amount of fluid in your brain increases during sleep, washing away the metabolic byproducts of all the activity that your neurons were engaging in during the day. Cleaning up this waste prepares your brain for the next day’s activities. If we skip or truncate our sleep, not only do we have trouble functioning the next day, but we also are more likely to forget what we were trying to learn the previous day.

If you don’t have time for a full rest, however, just doing a relatively simple activity can also boost your incubation productivity. Studies have shown that a task that doesn’t require your full attention, like a walk or an errand, can boost your chances of experiencing a eureka moment. And the longer the incubation period, within reason, of course, the more likely you are to find a solution unconsciously.

While marinating is not fully understood, it’s well established that it does happen. So use it to your advantage. You can encourage better ideas by intentionally allowing ideas to marinate—giving time for ideas to ripen, like a fine red wine or an aged cheese.

There appear to be several ways to encourage marination—that is, to encourage brain states in which you are more likely to have a eureka idea through background processing. Some people say that activities that allow your brain and body to relax or get into a repetitive rhythm are more conducive to producing a flash of insight. Many people report experiencing an aha moment while showering, bathing, taking a walk, going for a run, or during other exercise. Similarly, meditation experts report that they achieve an altered, more creative state while meditating.

Image Tip

Try using those precious few minutes of semiwakefulness in the morning as a time to gently call a particular problem or challenge to mind and encourage your brain to start chewing on it while your brain is feeling open and unhindered by full wakefulness. You may be surprised by the results.

Next time you need to schedule a work session, take care to consider whether breaking up the time over several days might be better than one long session. The marinating that naturally happens between each day might give you better results if you spread out the focused work sessions a bit. Also, the next time you feel blocked and tunnel vision is looming, or the team seems to be rehashing the same issue over and over, don’t keep beating on the front door; stop and come back later. Sometimes, a little time or a trip to the coffee shop is all it takes to get back on track. If nothing else, taking some time might help you get out of the weeds, and you’ll come back later with more of a balcony view.

SFE in action: Getting unstuck

Norman Furlong, Principal, Greenbook Inc.

When I worked for Boss Logic, a NeXTStep startup, I managed a group of developers in Silicon Valley. We worked out of a hillside house overlooking San Mateo. In those days there weren’t many resources available for Objective-C developers, so we were pretty much on our own. Almost every day, sometimes more than once in a day, a cry of anguish would emanate from the other room. This would warn me that one of the devs had gotten blocked and was in need of a distraction. I would spring into action and drive said developer down the hill for a cuppa Joe. On the way, we’d talk about the coding problem, and sometimes I could offer some cogent insights. But most often, we’d change the subject and talk about music, women, cars—anything but code.

Usually within 10, but rarely more than 20, minutes after we’d get back to the house, I’d be rewarded with cries of “Eureka!” coming from the now-unstuck developer. This routine, with minor variations, repeated itself all summer. When I sensed the whole team needed a break, we’d all head down the hill. The summer we worked together in that house was extremely productive, due in part to our devs getting time to walk away from the problem and marinate, releasing their conscious mind from the burden of solving the problem and allowing their subconscious to step in and do some of the heavy lifting.

Explore stage: Key tools and techniques

In this section we go into more detail about different techniques you can use, individually or as a team, to generate solution alternatives. First, we’ll talk about visualization techniques, such as sketching, storyboarding, and drawing flow charts and block diagrams. These are very powerful but lightweight ways to generate and explore alternatives. Then we’ll discuss brainstorming approaches that can be used equally well in concert with sketching or with more traditional verbal techniques. We’ll also introduce the broad topic of lateral thinking techniques and related approaches that help you supercharge your ideation by shifting perspectives to help generate more out-of-the-box ideas.

Image Mindshift

Use the whole brain. Remember when we talked about the power of including a diverse set of people when your team generates ideas? We suggested that a necessary step in the creative process is to harvest the brains of a diverse population, with different sets of experiences, backgrounds, perspectives, and so on. Well, the same approach applies to generating ideas with your own brain. To access the power of your whole brain, use idea-generation techniques that involve different types of thought processes and different parts of your brain.

You might write words, act out situations, draw pictures, or even use physical objects to explore different shapes and forms in physical space. You might brainstorm on the spur of the moment, or you might ask people to mull over the topic for a while before getting together to share their ideas. While not all of these modalities apply to any given problem or any given team, many of them will. By using different sets of techniques, you can bring more and different types of brainpower to bear on the problem at hand. The more modalities you use when you explore ideas, the larger the variety of ideas you are likely to uncover.

Visualization techniques

Among the many ways to express ideas as you generate and explore different approaches to solving your problem, the most obvious is to use words, creating long lists of ideas expressed as short phrases, sentences, or paragraphs. This is a valuable way to capture ideas, but it relies predominantly on the language functions of your brain, which are linked with logical, analytical thinking.

Visual and spatial approaches, such as building physical objects, sketching, and storyboarding, use other parts of your brain and have distinct advantages over verbal methods for many types of problems. Furthermore, some of the problems engineers encounter are highly visual in nature; they involve user interfaces, architectures, or physical devices that specifically lend themselves to a visual idea-generation approach.

Neuroscience break by Dr. Indrė Viskontas

The left brain/right brain distinction is a lot murkier than most people realize. There are many connections between the two hemispheres of the brain, and the idea that each hemisphere acts alone is not supported by neuroscience. What’s more, creativity engages the left side of the brain just as much as the right, with a recent meta-analysis showing no clear evidence for a greater role played by the right side in creative thinking. But for most people, the conscious mind is dominated by language—we think in words more often than in pictures or other symbols. So the language centers in the left brain are sometimes thought of as “dominating” much of our thinking.

There is evidence that our frontal cortex can control neural activity in other parts of the brain, like our medial temporal lobes, where our long-term memories are stored. By switching approaches and engaging parts of the brain that are not involved in language processing, we might be able to release other parts of the brain that are inhibited by the dominant language areas and let them “speak” for themselves. In fact, in studies of dementia patients who lose the ability to communicate verbally because of a progressive neurodegenerative disease, we sometimes see an emergence of visual creativity: that is, as their language regions degenerate, other parts of their brains, like the parietal and visual cortices, can have a greater influence on behavior.

Let’s explore a few ways to capture and explore ideas visually.


A particularly versatile, but frequently overlooked, mode for generating ideas is to use some form of sketching. You might sketch different ways to depict an icon, draw several alternative flow charts to describe a workflow, whiteboard different architectural block diagrams, sketch alternative layouts for a user interface, or perhaps link several sketches in comic-book fashion to rough out an end-to-end experience in a storyboard.

The visual nature of drawings can help you notice connections and ideas that you would be very unlikely to see if you just described the same ideas in words. Many teams find that when they take the time to do some rough sketching as part of exploring alternatives, they discover ideas that they would not have found another way.

It’s helpful to think of sketching as just another idea-generation tool; it’s brainstorming with pictures instead of words. As such, tunnel vision applies to drawing just as much as to any other brainstorming technique. So keep in mind all the principles about why you should go broad first, not fall in love with your first good idea (or good picture), and not judge ideas until later. The ultimate goal of a visual idea-generation exercise is to sketch as many different ideas as you can in very rough form, mindfully postponing the details to generate the greatest diversity of ideas possible before tunnel vision sets in.

Image Tip

We strongly encourage people to sketch with a marker, which naturally writes with a broad line, helping you to keep detail to a minimum and your drawings rough and simple. This keeps you in the zone for idea generation, rather than straying into drawing with distracting details. Also, you can’t erase a marker, which helps stop you from striving for a beautiful drawing and to just go with what you first put down, mess and all. The bold look of a drawing made with a marker is also ideal for sharing with others across a table or pinned on the wall. These drawings are much easier to see at a distance than a pen or pencil drawing.

It’s important to understand that sketching is not about drawing beautiful pictures, it’s about exploring different ideas. We don’t care if our drawings aren’t beautiful. In fact, they may be only semi-legible. But if a sketch means something to the person who drew it and helps him explore that idea, it’s good enough. Simple sketches are the most effective at this point to focus on the essence of the idea being proposed. Sketches do not need to be pretty, or accurate, or complex. They just need to capture and communicate enough of the idea that you can return to it later to develop the idea more fully.

Image Mindshift

Beware of beautiful drawings (and great artists). Just because one person on your team happens to be an amazingly talented artist, and her sketches look light years better than everyone else’s, does not mean that her idea is the best or that everyone else should sit back and let the team’s artist do all the drawing. Remember that the goal of this stage of the Fast Feedback Cycle is to generate lots of alternative ideas and to leverage the diversity of perspectives of everyone on the team. If you’re going to use everyone’s visual brainpower, everyone needs to be holding a pen and drawing. A rough, messy, imperfect sketch is just as likely to carry a promising idea as a visually beautiful one.

Even a very rough sketch can help you better communicate your thinking to a teammate and create a shared understanding of an idea. With words, ideas can easily be misinterpreted or misunderstood, because different people unconsciously imagine very different implementations of the same basic idea yet still believe they are talking about the same thing. With ideas expressed as pictures, the gap between possible implementations begins to close. Drawings have much higher information density because they inherently include aspects like relationship, size, and proximity in addition to the specific content being drawn. As the old adage goes, a picture is worth a thousand words.

Unfortunately, many engineers believe that they can’t draw, so they sometimes avoid sketching as an idea-generation technique and miss out on all its creative benefits. Some may happily sketch an architectural block diagram on a whiteboard but shy away from trying to draw anything more complicated than that. We see it in our workshop all the time; when we start talking about sketching, the room gets tense. If we don’t start with a warm-up exercise, it’s common for some people in the room to refuse to pick up a pen when we ask them to try sketching out ideas for their scenario.

In the workshop, after introducing the idea of sketching and how simple it can (and should) be, we have another timed contest. We ask everyone in the room to fill a page-size grid of 25 circles with a quick sketch, one drawing per circle. It can be a happy face, a sunset, a boat . . . it doesn’t matter, as long as people don’t use numbers or letters. The goal is to fill as many of the circles as possible in one minute. An example is shown in Figure 7-3.


FIGURE 7-3 An example of the circle exercise.

When the minute is over, we ask everyone to hold up their drawings. Most are able to fill about half the sheet, but only a few fill in all of the circles. We ask the people who filled the page to share their secret to success. They say things like “keep it simple,” “no details,” and “follow a thread of cousins.” At this point, most folks in the room are feeling a bit more confident in their ability, and many are itching for another chance. We give them a fresh sheet and set the timer for a minute. This second time, we observe that almost all the participants are able to fill the entire sheet with sketches.

After watching literally tens of thousands of engineers sketch their solution ideas in our workshops (not just this warm-up exercise), we can say with confidence that every engineer is capable of drawing the simple sketches needed to explore ideas. In fact, we are continually amazed at the quality of the sketches we see, and the quality of the ideas generated through sketching, especially from people who emphatically declare that they are terrible at drawing.

SFE in action: But I can’t draw!

Lisa Mueller, Senior User Experience Lead, Microsoft Corporation

On our team, after the scenarios were written and signed off, diverse teams (including UX, PM, Developers and Test) were put together to start on paper-and-pencil prototyping. In the beginning, these teams were very apprehensive about starting unless they knew that a UX team representative would attend. They felt this was the “drawing” phase, and since they didn’t have any experience with drawing, their confidence was low and drawing was something these small teams of engineers just didn’t feel comfortable doing. In the end, if a UX team member couldn’t attend, many of these prototyping sessions were canceled and rescheduled.

We anticipated that our approach to paper prototyping was going to be to get into groups of two to three people to sketch out possible walk-throughs of each scenario. However, because so many people were so hesitant to draw, we changed our procedure. Instead of asking the small teams to sketch together, we directed the team members to work individually for one hour and to create their own walk-through sketches. Afterward, each person would present his or her sketch or drawing of the scenario to the rest of the team.

As each person presented his or her sketch, the team would highlight ideas that seemed particularly good with a sticky note and a star. By the time we got to the third presentation, the energy in the room was high and becoming very collaborative. Individuals began to notice and appreciate the different types of ideas that were coming from different people and that each team member brought a new strength to the table. The PM brought subject matter expertise. UX brought the UI framework. Developers brought the step-by-step plan, and testers brought detailed content. The combination of these strengths was very powerful. After this sketching-brainstorming meeting, everyone became much more engaged because they had the realization that this exercise wasn’t about how well you could draw but about leveraging the strengths and perspectives from across the team.

SFE in action: How (and why) to draw anything

Dan Roam, international best-selling author of The Back of the Napkin and Show and Tell. (All drawings © Dan Roam, 2014. Provided with the author’s permission.)

Here are two interesting data points:

Image More of our brain is dedicated to processing vision than to any other mental task.

Image More of our body’s energy is consumed by our brain than by any other organ.

Adding those two facts together tells us something important: as humans, we are essentially walking, talking vision-processing machines.

For scenario-focused engineers, this offers a critical but overlooked insight: if we want to maximize our innate problem-solving skills, we should structure our problems visually. By intentionally enabling our visual mind to actively engage in defining, structuring, and formulating problems, we will discover that problem solving can be faster and more creative than ever.

The problem is that we rarely intentionally engage our visual mind. Why? Because we are afraid to draw. So step number one in becoming better visual thinkers is to relearn the simple art of drawing.

Here are six quick exercises to get you started:

1. All drawing starts with five simple shapes.


2. By combining these shapes, you can draw almost anything your mind can conceive. Can you draw these? What else can you draw?


3. Since we usually find people at the center of most problems, it will be helpful to be able to draw them. Simple figures are good enough. Can you draw these?


4. Stick figures are ideal for showing individuals and emotion. Can you draw these?


5. Block figures are good for showing action. Can you draw more like these?


6. Blob figures are ideal for showing groups and relationships. Can you draw more like these?


I hope that by showing these simple examples, you can see how easy drawing can be. Now think about how you might apply this to clarifying your challenges.

Drawing and sketching is a very powerful technique, both for generating creative ideas and for communicating thoughts, ideas, and concepts to others. You do not need to be an artist, and your sketches do not need to be sophisticated or beautiful. By drawing the kinds of simple figures that Dan Roam demonstrates, you can describe just about anything visually.


The specific sketching technique that we find most useful for generating ideas about an end-to-end experience is storyboarding. Instead of sketching a single picture, you string together several pictures into a sequence of actions or steps to form a storyboard.

Storyboarding originated in the film industry as a comic-book-style approach to drawing out sequences of movie scenes. Storyboards help filmmakers think through all the different camera angles, shots, and actions they need to capture to create the final film. It’s very expensive to reshoot a scene if there is a mistake or a missing camera angle, so it behooves film directors to carefully consider all the pieces they need to thread together to form the final experience before they are on location with actors, gear, and staff.

Storyboarding is helpful for mapping out software experiences as well. It is particularly useful for the kinds of scenarios that engineers are typically solving because it encourages you to think through sequences of interactions, not just individual moments. If you’re serious about building end-to-end experiences for your customers, storyboards are your first, best tool to start exploring different end-to-end sequences that might solve a given customer scenario. Just like in scenario writing, the storytelling nature of storyboarding helps you get to a deeper level of understanding and empathy with your customers, which may suggest new avenues for inspiration.

It’s critical to approach storyboarding in your first couple iterations of the Fast Feedback Cycle as still being in idea-generation mode, trying to generate multiple diverse solution approaches before tunnel vision sets in. In later iterations, more-refined storyboards play a role in helping you fine-tune ideas and detail improvements to the experience, but the focus in the beginning should be on very quick, rough storyboards that explore as many different sequences of interaction as possible. Having alternatives will help you develop a feel for which ones flow most smoothly and best create the customer experience you are aiming for. Just like with other idea-generation techniques, only after you generate many divergent ideas do the relative merits of each approach become clear and the decision points become more obvious.

Figure 7-4 shows several different storyboards of a mobile app concept that would make it easy for friends who are physically near one another to find each other and meet up. Each storyboard shows a different sequence of how the user interface might be designed, in very rough form. These storyboards explore different ways to upsell a new feature and also manage privacy and consent for sharing your physical location with friends through a mobile app.


FIGURE 7-4 Four different storyboards exploring alternative user interface flows for a mobile app.

Here are some tips to keep in mind when storyboarding:

Image Storyboards show sequences A storyboard should have a minimum of three frames and can have many more. Frames can be a mix of comic-book drawings of stick figure customers in a situation, a sketch of a possible user interface flow, or a sequence of steps that the customer might interact with.

Image Keep the details to a minimum Notice that storyboard drawings are very simple and straightforward. They leave out a lot of details. We’re talking line drawings, no artistic flourishes, no detailed pictures; just the core idea with a minimum of ink.

Image Brief captions are enough Writing a brief caption under each frame can be helpful to give context or communicate the story if it’s not obvious. However, there’s no need to write every single word in the user interface you are imagining. Some wavy lines to indicate where text might go and a keyword or two is generally enough to get the idea across. Remember that you’re going for speed so that you can produce multiple ideas.

Image Show people and emotion While some storyboards focus directly on what is happening on the screen, many storyboards include frames that show the user before, during, or after the experience. Don’t worry about having to draw people—simple stick figures are just fine, but adding a quick facial expression (smile, frown, etc.) or thought bubble can carry a lot of meaning about the customer’s context, emotion, and state of mind.

Image Don’t get attached The more effort you put into making a storyboard beautiful, the more you will like it. Indeed, you might fall in love with it. And that will lead to tunnel vision, making it hard for you to imagine that other ideas exist.

Flow charts and state diagrams

Some problems lend themselves to a flow chart or state diagram, showing the logic of how each state flows into the next. This approach is particularly useful for deeply technical problems, but it can also be used for drawing user interface flows to work out more detailed interactions or to fine-tune a proposed interaction in later iterations of the Fast Feedback Cycle. Some teams will develop their more-refined storyboards into user interface (UI) flow charts. These show how the UI might fork at various decision points in the interface: if the user chooses A versus B, what happens next for each of these choices.

This technique is not often the best choice in the first couple of iterations of the Fast Feedback Cycle. However, after you generate some ideas and get feedback, pulling your most promising ideas into a flow chart can help you refine those ideas and formalize the details and the flow at the next level of detail. Here are a few things to consider as you refine the flow:

Image Switch the order If your proposed flow asks the user to log in first, flip the logic so that authentication happens later on or at the very end.

Image Challenge each branch Is this decision fork absolutely necessary? What if you just don’t ask? What happens if you pick a reasonable default instead?

Image Take different customer points of view If your flow chart describes the experience of the end user, also draw the logic for how the experience might be for an administrator, purchaser, partner, or another customer type in your ecosystem.

Block diagrams, interfaces, and architectural drawings

The diagrams most often seen on the whiteboards of engineers are architectural block diagrams of one sort or another, often with lines, arrows, or lollipops connecting them. Block diagrams are excellent for describing modular components of software, abstraction layers, interfaces, APIs, and relationships between larger components such as servers, services, and databases. You already know how to draw these diagrams, and the finer points of formal architecture drawings are beyond the scope of this book.

However, many of the problems we’ve discussed in this chapter lurk beneath the surface for architecture diagrams as well. In particular, watch out for tunnel vision in your architecture diagrams and be open to true iteration, not just minor tweaks and fine-tuning. We hear too many stories of developers who create one reasonable architectural drawing and then, as they get feedback, start decorating their diagram with the parts they missed—a link to the authentication service here, a missing component there, a link between that component and this other one that will need to share data . . . and pretty soon you wind up with a spaghetti diagram. By that point, however, the proposed architecture is so firmly established in everyone’s brains that it can be hard to see other alternatives. Indeed, it’s easy to convince yourself that this is just a complex problem, so you resign yourself to building a complicated architecture as well. But is that really true, or is that tunnel vision talking?

We do an exercise in class in which we ask the groups at each table, after they have iterated on their storyboard concepts, to sketch out what a technical infrastructure to implement their concept might look like. We set a specific goal to draw three different architecture diagrams of whatever sort is most appropriate for the problem the participants’ have identified. It’s striking how few groups are able to come up with even a second architectural approach, never mind three. Yet as an instructor (with a fresh brain full of fluffy white snow), it’s easy to walk over to a table, hear the basic idea, and suggest a few quick questions, such as, “Have you considered a peer-to-peer approach? Or a client-server approach? A cloud-service approach? A rich-client approach?” These are all familiar, basic architectural paradigms, but somehow the tendency is to zoom in to a particular approach right away and go so deeply into that idea that other options are forgotten.

What we observed was that teams that are able to come up with a few meaningfully different architectural diagrams also report that they didn’t pick just one of the alternatives to move forward with. Rather, their favored implementation ended up being a blend of different ideas that were suggested. They often remark that they would not have likely considered that blend had they not forced themselves to come up with multiple alternatives in the first place.

Just as with user experiences, a good starting point for designing stronger architectures is to consider multiple options at the beginning. That means that when you sit down to draw some block diagrams of how your architecture might look in your next release, or some lollipops to work out the interface between this component and that one, the behavior to cultivate is to not stop after you work out one reasonably good approach. Keep drawing, and force yourself to come up with another credible alternative, and ideally more. Bring colleagues into that process to bring more perspectives and other approaches into consideration. The idea is to draw many possible architectures so that later you can compare and contrast their pros and cons and allow the simplest and most elegant blended solutions to emerge.

Image Mindshift

Think like an architect. One thing we’ve noticed working with teams is that expert software development architects naturally use a lot of the techniques we discuss in this book. However, they do it largely by instinct, not because they’ve been taught to. Perhaps it was precisely these instincts that led them to be architects in the first place.

For instance, expert architects believe that an architecture or an interface cannot be written in a vacuum—it’s best to be paired with another team that will actually use your architecture or interface to be sure you’re building something that works in real life. We’ve also observed that many expert-level architects intuitively think about customer and developer desires and needs as well as the end-to-end usage scenarios in which their architectures will be used. They intuitively generate multiple ideas, and before they pick an approach, they investigate the relative merits of those approaches (by researching documentation and discussion groups, networking with others who have experience, playing around with the code, or building functional or semifunctional prototypes). They are willing to iterate, sometimes many times, before settling on a final design.

The bummer is that many of these steps happen in their head, or when working solo on a whiteboard, making it hard for other, more-junior developers to observe and learn this craft and approach. This also means there are fewer opportunities for feedback and for incorporating diverse perspectives that might have enriched the plan. Over the years we’ve also heard occasional complaints from architects who did all the right things, yet had trouble convincing the team of their final recommended approach because they didn’t bring the team along on the journey of their thought process.

We encourage architects who have this predisposition and skill set to practice their craft in a more public way, to teach others, better influence their teams, and ultimately build better architectures. And if you are a software developer who hopes to be an architect someday, pay attention: these are core skills to develop that will help you achieve your career goals.

Hold a sketchfest or charrette

Of course, all of these sketching techniques can be done solo or in groups, but it is a rare engineering team that has developed a strong habit of sketching to generate ideas. One productive technique to encourage more sketching and start building that muscle is to schedule dedicated time for a team a team “sketchfest,” sometimes also called a “charrette.” Get a broad group of team members together in a room, remind everyone of the current scenario or technical challenge, and have folks pick up markers and start sketching. After everyone has had time to sketch some ideas, have people present their more interesting concepts to each other, in pairs or in small groups, to allow ideas to mingle and encourage blends to emerge. Plan time for a second or third round of sketching, mix up the sharing groups, and be sure to collect the ideas that are generated as you select the most promising ideas to move into the next stage of the Fast Feedback Cycle.

Physical models

If you are aiming to build a physical product of some sort—a hardware device, a handheld gizmo, or a peripheral—it naturally makes sense to explore ideas physically in three-dimensional space. Whether you use modeling clay, paper, cardboard, foam core, wood, epoxy, plastic, or other materials, the goal here again is to explore many alternatives very quickly so that you can compare and contrast the best aspects of each possibility, consider blends of different approaches, and get a more visceral, physical sense of how these form factors might look and feel. (For an example of exploring ideas in a physical medium, using mostly rough-shaped clay and fiberglass models, review the Microsoft mouse example in Chapter 3.)

Using 3-D models is essential for building hardware or any kind of design for a physical object, but it is not a broad-purpose technique. An expert in designing 3-D objects (and all of the attendant complexities of structure, materials, function, and human factors) is called an industrial designer. So many technical details are involved in building a physical device of any sort that we strongly recommend that you employ a professional industrial designer.

Brainstorming techniques

Originally developed by marketing executive Alex Osborn in 1942,7 brainstorming is a term that refers to a family of techniques that allow you to quickly generate a large number of ideas, usually taking advantage of the inherent diversity in a group of people. Unfortunately, the word “brainstorming” is often overloaded in its use and can refer to anything from sending email with some ideas to a couple of people, to chatting out by the water cooler, to a formalized ideation session with a facilitator.

When we begin the process of adapting our workshop for a new team, we first try to assess what iteration, research, and creativity techniques the team already embraces. It’s been interesting to us to observe that virtually every team reports that they already engage in brainstorming activities. But, when we dig in, we soon discover that few teams have the knowledge (never mind the disciplined practice) of engaging in actual brainstorming techniques, other than in the loosest sense of the term.

Image Vocab

Brainstorming is a family of structured, idea-generation techniques that can be used to generate a large set of divergent ideas for a given stimulus.

Group brainstorming

The classic brainstorming technique is to gather a group of people in a room and spend anywhere from 20 minutes to a couple of hours focused on generating ideas for a given topic. Meet in a comfortable place with lots of whiteboard space, have a facilitator to keep the conversation moving forward, and have a scribe write down all the ideas. Use your chosen scenario as the topic, and strive to fill the whiteboard with ideas. Sounds simple enough, but it’s actually quite challenging if you’ve never done it before in a structured way. However, when it is done well, group brainstorming can be a tremendously valuable source of creative input to the Fast Feedback Cycle.

We go into detail about how to facilitate and lead a group brainstorming session in the “Deep dive” section later in this chapter.


One piece of feedback we get about the “send a message” brainstorming exercise in our workshop is that it’s not always fun for introverts. Verbalizing ideas doesn’t get the creative juices flowing for some people, and we are sometimes asked questions like:

“I’m an introvert and I’m not comfortable shouting out my ideas that way. I like to think alone and mull things over. Does that mean I can’t participate well in group ideation?”

Given that many talented engineers lean toward being introverted, this is a great question to ponder. And while a traditional group brainstorming session is a valuable technique, it does tend to be more comfortable for extroverts and for people who are comfortable speaking off the top of their head. With people and teams who may be more introverted, you might generate more ideas if participants write down their ideas instead of speaking them out loud. It’s also true that many of these people appreciate the opportunity to bounce ideas around with others, but only after they’ve had some time to gather their thoughts first. There are alternative ways to running a brainstorming session that rely less on extemporaneous talking but still take advantage of a diverse team’s fresh brains. We call these variations quietstorming.

Here are a few quietstorming techniques you can try:

Image Entry tickets Have each participant show up with three to seven ideas as their ticket to enter the brainstorming session. Go around the room and have each person share his or her ideas. Allow others to jump in, and encourage piggybacking and new ideas to emerge. However, be sure to postpone judgment and discussion of the ideas until later; this is still about generating lots of ideas, not about investigating them.

Image Start with quiet time Set aside a specific amount of time at the beginning of the session for each participant to quietly write down his or her initial ideas, each on its own sticky note. Usually 5-10 minutes is plenty. Then, go around the room and ask participants to share those ideas one by one, and riff away. As each idea is presented, place the sticky note on a large piece of butcher paper on the wall to create a dynamic map of ideas, grouping similar ideas near each other. It’s best to have each person share one idea at a time so that the latter part of the session doesn’t degenerate into a “readout” of each person’s private (and usually long) brainstorming list, which does not encourage piggyback ideas to emerge. When someone mentions an idea that another person also wrote down (or mentions a cousin idea of some sort), that person should speak up and go next.

Image Pass the paper Have the participants sit around a table, and give each person a piece of paper. Set a timer for a couple of minutes. Everyone then writes down a few ideas, and when the timer sounds, passes their paper to the left. Next, each person reads the ideas on the sheet they received, adds their thoughts to that list (stimulated by what is already there), and again passes to the left. Continue this for several rounds.

Image Computer mediated Use an online chat room, an instant messaging group, or other electronic tool in which many individuals can type their ideas in a visible, sequential way, each at his or her own computer. This technique is particularly good for large groups. It’s impractical for more than about 20 people to brainstorm in a room without talking over one another or having to wait for a turn to talk, but they can all type at the same time. It’s also a good approach for a distributed team that isn’t located together. A computer-mediated approach can also mitigate some of the cultural issues that can impede group brainstorming, with its perceived anonymity and the emotional safety of typing rather than speaking. However, this technique does not work well for ideas that are better expressed with a quick sketch, an out-loud verbal description, or pantomime.

Image Solicit ideas over time Set up ways for customers, partners, or employees to submit a cool idea whenever it occurs to them, and leverage those ideas as inspiration when exploring alternatives. In this way you can create a pipeline of ideas that can incorporate the thinking of a large organization or community of users. However, expect that many of the ideas you capture in this way will not align with your business strategy, target customer, or chosen scenarios, especially if you do not limit ideas to a particular topic or challenge, so be careful that you don’t get blown off track chasing a shiny object. Even so, some of the ideas will align with your goals, and occasionally one of these ideas might help uncover a potential customer need you hadn’t noticed yet that inspires your future plans.

Individual brainstorming

A few techniques worth mentioning work best for individuals to generate ideas on their own. Of course, these can be used together with the more group-oriented ideas, but they tend to be better suited to individuals or a very small group.

Image Mindmapping A mindmap is a popular way to brainstorm ideas, where you create a visual record of which ideas are connected to others as you brainstorm. Figure 7-5 shows an example.


FIGURE 7-5 A simple mindmap. Large mindmaps can easily become complex enough to cover a large sheet of butcher paper.

Image Stream of consciousness Write down your ideas as quickly as possible in a narrative form and don’t stop—write down whatever pops into your head. If you prefer, you can do this orally, record the audio, and transcribe it later.

Image Keep a notebook Keep a notebook with you at all times, and jot down ideas whenever they occur to you. If you see an interesting idea in the world, note it or draw a quick sketch of what you found interesting. If you wake up at 3:00 a.m. with a brilliant idea, note it and go back to sleep. You get the idea. Later, flipping through your notebook during a brainstorming activity can help trigger ideas for you.

Image Don’t ignore serendipity A chance meeting with an expert, an unusual flyer in your mailbox, a coincidental situation that helps you see from a new perspective can all provide unexpected inspiration. Be ready and open for what ideas are looking for you, and when a great opportunity happens to come by, don’t be shy—speak up, ask questions, get involved—it might just be the missing piece you’ve been looking for.

Supercharging your idea generation

A few powerful techniques can help supercharge your idea generation. These techniques encourage diverse ideas and ensure that you don’t miss anything obvious. They all work on the principle of shifting your perspective to look at problems from a different direction.

Practically speaking, these techniques are most useful when you start feeling dry, when you reach a point where you feel like you can’t think of any more ideas. Whether you are in the middle of a raucous brainstorming session, quietstorming on paper, molding with clay, or sketching out storyboards, try one of these techniques when you get stuck to look at the problem from a different perspective. Especially if you can suspend disbelief for a while and encourage wild ideas, these techniques can often open up a whole new line of thinking.

Lateral thinking techniques

First introduced by Edward de Bono in the 1960s, lateral thinking techniques are tools specifically designed to encourage more lateral jumps in your idea generation.8 They force you to look at a problem from an unusual perspective, in a sense tricking your brain into looking at the same problem in a radically different way. By turning the problem upside down or inside out, you increase the chances that you will see a connection or notice a new way of approaching it that you may not have found by looking at the problem straight ahead.

Here are our favorite lateral thinking techniques that we call on again and again in our work with teams:

Image Challenge assumptions Our example from earlier in the chapter about designing a coffee cup without a bottom is based on challenging a basic assumption about your problem space. What other kinds of assumptions can you challenge? Let’s say you’re building a user interface. Could you do it without asking the user to type anything? Could you build a website that never required the customer to log in? Could you build a compute-intensive game without using any local computer CPU? Remember, the point of challenging assumptions isn’t to impose that restriction on your final solution. But by attempting to solve the problem in a more constrained space, you might notice a new approach that wouldn’t have otherwise occurred to you.

Image Random input and association Open the dictionary and pick a random word. Or use one of the random word generators you’ll find on the Internet, built for this purpose (yes, they exist!). Force yourself to find an association between the word that comes up and your chosen problem. Sometimes (but not always) that association will open up a new way of thinking about your problem. For example, how is a company’s performance review process like a canoe? Well, it involves two people, one who does the paddling, and one who does the steering, and if either person rocks the boat, the whole thing capsizes.

Image Reversal This is one of our favorite techniques because it is so easy to do. Instead of brainstorming against the chosen problem, brainstorm against the opposite problem. For example, ideas could run dry pretty quickly if the problem is “design a great emergency room.” However, if you turn the problem around and brainstorm all the ways you can make an emergency room terrible, the ideas flow like a geyser. Make the patients wait forever. Make the chairs uncomfortable. Give the patients nothing to do or nothing to look at. Insist on complete quiet. Keep it dirty. And so on. Now reverse those ideas and make sure that you are solving all of those problems.

There are many, many lateral thinking techniques besides these, as well as other resources available to develop lateral flexibility, from puzzles and games to workshops and worksheets. Just search for “lateral thinking,” and you’ll find tons of interesting options to try out.


SCAMPER is an acronym that stands for Substitute, Combine, Adapt, Modify, Put to another use, Eliminate, and Reverse. It was invented by Bob Eberle, though it includes many ideas that were part of Alex Osborn’s original conception of brainstorming. Very similar to lateral thinking techniques, each of the keywords within SCAMPER encourages you to modify your thinking in various ways to help you see other solution approaches. For instance, “Substitute” might trigger you to wonder whether there is a way to substitute an existing product for the one you’re proposing to build, or if you might substitute a different type of material for the one you are considering. You can find many listings of SCAMPER-inspired questions on the Internet.

Six thinking hats

Six thinking hats is another idea developed by Edward de Bono. Each hat encourages you to look at a problem from one of six different perspectives. It is a helpful technique to ensure that you don’t miss anything obvious and are complete in your exploration. Like the lateral thinking techniques, it can help get you moving again when you feel like your idea generation is slowing down.

The idea is to “wear” each of the six different color hats when exploring a problem: for example, the white hat focuses on gathering information, the red hat focuses on hunches or intuition, and the black hat plays devil’s advocate. One of the benefits of devoting a specific block of time to each hat is that it limits a group from falling into habitual thought patterns, such as always playing devil’s advocate, and forces the group to consider each of the perspectives in turn. A good description of the six thinking hats and training is available from the de Bono group at

Turning the corner: Deciding which ideas to move forward

Once you’ve spent some time diverging and creating a pile of ideas, it’s easy to get lost or even overwhelmed. Which ideas are the best ones? Do you start with the sketches or the brainstorming list? How do you pick the ideas to move forward with? Are any of the ideas any good? Are there patterns? Can you combine some to create something new and even more interesting?

Just as you did at the end of the Observe stage (see Chapter 5, “Observing customers: Building empathy”), when you had mountains of data (in that case, from research about customers), it’s again time to converge and make some decisions about which ideas are the most promising. Several techniques can help you turn the corner and start sorting through and making sense of all the ideas you’ve generated, and this will help you forge a path forward.

The basic idea is to cull your list of ideas, identify a few that seem most promising for supporting your business, technical, and experience goals, and use those as input for the next stage of the Fast Feedback Cycle. In the next stage, you will rapidly prototype to flesh out those ideas just enough that you can get some customer feedback. Your goal here is not to pick a single idea to move forward with, but to winnow your list down to several good alternatives that you can rapidly prototype and then see how customers respond. Here are some approaches to help you figure out which ideas to move forward with.

Write it down

You can make the decision process a lot easier by being sure to write down your ideas all along, and archive them in a reasonable way. Use something as lightweight as a manila file folder full of sticky notes, or a folder in the cloud where you plunk all your notes, sketches, and photographs of whiteboards. Generally speaking, doing a lot of data entry and organization is not terribly useful at this stage, so keep it effective but lightweight. The important goal is to be able to scan through your ideas in one place. You don’t want to have to dig around to locate them or, worse, try to recall your ideas from memory. You want these lists, sketches, and notes easily available because you might use them later for inspiration or to reconsider some alternatives if you find that your initial approaches aren’t working as well as you hoped.

One particularly efficient way to capture ideas is to write them on sticky notes, one idea per note; for larger sketches, keep each sketch on a single sheet of paper. Whether you are sketching storyboards or flow charts, brainstorming ideas verbally, quietstorming, drawing mindmaps, or doing anything in between, putting each idea on its own piece of paper makes a lot of the processing we’re about to talk about a ton easier. Sure, you can always transcribe each idea onto a sticky note later for sorting, but if you keep your notes on individual sticky notes in the first place, you’ll maximize your overall efficiency.

Consider the feasibility

You are narrowing down a pile of ideas. It’s time to get real. You need to take a hard look at the feasibility of those ideas, both for the technology involved (Can it be built and at a reasonable cost?) and for your business (Is it competitive; will anyone pay for it?). The first thing to do is sort out your ideas a bit so that you can see which ideas align to your business strategy, which support a great customer experience within your chosen scenario, and which are most feasible to build with available technology. The most promising ideas are those that support all three. Having a great customer experience supported by reasonable technology will get you nowhere if no one pays for it. Having a great business model and a terrific customer experience won’t work either if the technology isn’t reliable. You really need all three. Deciding which ideas to move forward with is the trigger to start bringing business and technology thinking into the solution and design process.

But this is not the time to do in-depth competitive analyses of each possible alternative, or to do serious development costing either—there are still too many ideas to consider. However, this is a good time to include a few senior people who can offer a rough judgment of whether a particular idea aligns with the business strategy or would support a compelling marketing message. In particular, you want people who can quickly identify ideas that are in direct conflict with your business plan. You should also include some senior technologists who can say whether an idea appears to be reasonably feasible. Ideas that are patently impossible or are in conflict with your strategy should be culled from active consideration.

One way to get this sort of input on the technical side is to do T-shirt-size estimations. For example, estimate whether a particular idea is small, medium, or large, where each size is an order of magnitude larger than the previous one. Small might mean that an idea could be implemented in an amount of time measured in days. Medium means it would take weeks, and large would be an investment of months. You’ll be tempted to use extra-small for teeny-tiny ideas that are smaller than small, and extra-large for ideas that seem so impossibly huge that they appear to be unbounded. You can add these sizes if you like, but the intent is to keep it simple because these estimates should not be considered binding. Rather, they are very quick shorthand to help engineers convey to their leaders and the broader team the order of magnitude of work that is likely to be needed to deliver an idea.

Image Mindshift

T-shirt-size costing without guilt. Engineers often get quite nervous about T-shirt estimates because they don’t have nearly enough data to make a true estimate at this stage in the process and do not want to have to guess. They rightly argue that many ideas can be implemented “the easy way,” with minimal configuration and customization, or they can be implemented “the complicated way,” with specialized algorithms, custom animations, and other added complexity. At this stage in the process, no one has any idea where on that continuum delivery of the final experience will be. On top of this, engineers really don’t like being wrong, and they worry about the consequences of a bad estimate. They worry if they say that an idea is small but turns out to be much more involved than they realized that the team will hold them to that estimate, they’ll be perpetually behind schedule, and they’ll be blamed for slipping the schedule. It’s no wonder that engineers are resistant to T-shirt-size costing.

The trick to making T-shirt-size costing work is never to use T-shirt estimates for building schedules and never hold individuals accountable to these estimates. The only value of a T-shirt-size estimate is to get a rough idea about whether an approach is buildable at all—to ensure that you are focusing on ideas that balance business, experience, and technology. Sometimes the answer is to give a range of sizes. To know that an idea is somewhere between medium and large is still useful information for prioritization.

If T-shirt-size estimates continue to give your engineering team heartburn, you could switch to using a yes/no format: yes, the idea is buildable; no, the idea is not buildable. However, in practice, almost everything is buildable given enough time, so the yes/no format leaves valuable information out of the equation. An order of magnitude of time is a pretty good proxy for technical complexity as long as you don’t try to add more precision than that.

And, of course, we all know from long experience that estimates, even from the most senior engineers, tend to be wrong. It’s a good bet to assume that everything will take longer than you imagine. That’s another good reason to not use a T-shirt-size estimate for more than a super-quick gut check, which is after all, what it is.

A particularly productive way for a group of engineers to do T-shirt-size estimation is to have them vote in real time as a group. That way, no single person is responsible for the estimate, which helps reduce anxiety, and voting as a group also takes into consideration factors from different perspectives. The way to vote it is to announce an item, and then on “Go” everyone holds up one (small), three (medium), or five (large) fingers to express their judgment for that item. If there are disagreements, the people with the largest or smallest estimates explain their reasoning, which sometimes brings an important consideration to the foreground, and usually a group consensus is quickly reached.

Affinity diagramming redux

Another helpful way to make sense of a large pile of ideas is to sort them into an affinity diagram. Just as with customer data, you can group ideas into ones that are similar to each other and see which buckets emerge. Having each idea on a separate piece of paper or sticky note makes this process dramatically easier. You can build an affinity diagram with ideas that are primarily textual, primarily sketches, or a mix. There is a detailed description of affinity diagramming in Chapter 5.

But affinitizing solution ideas has some differences from working with customer needs. With customer needs, an important channel of information is the size of the affinity group, which serves as a proxy for how loudly that need was heard. When you group solution ideas, even a category with one idea may turn out to be the most valuable. So don’t pay too much attention to group size this time.

The primary value of an affinity diagram of solution ideas is to collapse into a single group all the ideas that are essentially variants of the same concept. Doing this helps show you which ideas are meaningfully different from one another. You want to avoid moving forward with several ideas that are really just cousins and are too similar. Rather, you want a diverse set so that you explore the solution space and give yourself the best mathematical odds at finding an optimal solution. Pick one approach (or a blend) from each of the most promising groupings to move into the next stage of the Fast Feedback Cycle.

Look for the blends

As you generate ideas, you should always be on the lookout for ways to blend or combine ideas to come up with a new approach or to improve an aspect of a given idea. Similarly, as you look over your ideas, you may notice blends that you didn’t before. It’s totally fair and, in fact, a really good practice, to propose that an idea you move forward into the next stage be a combination of one idea with another.

Update the museum

If you built a museum from your customer data in a room or a hallway, this is a good time to update it with the solution ideas from your brainstorming. Seeing multiple ideas in a single eyeful, walking by them every day on the way to the elevator, and exposing everyone to new ideas generated by other members of the team can all be helpful for making connections and noticing possible blends. An affinity diagram lends itself nicely to being posted in a hallway or other museum location.

Updating your museum is also a great way to collect feedback and focus attention on the latest ideas. Tack up pens and sticky notes so that passersby can add comments. You might be surprised at how often leaders (who are so pressed for time) walk by such a museum, have something catch their eye, ponder the wall, and leave comments or contact you in email. Think of your museum as a manager glue trap, except you aren’t catching mice.

Collective decision making

Now that you’ve sorted and culled, you need to make some decisions about which ideas are the most promising and which should be moved forward into the next stage of the Fast Feedback Cycle. In collaborative project teams, usually no single person is charged with making the decision about which ideas move forward. The team is expected to come to consensus on most decisions and involve their managers or other leaders only at a significant decision point or if an irreconcilable disagreement occurs. This is generally a good thing because it means that you rely on the diversity of the team to make better-quality decisions. However, consensus-based decision making can also be very slow, and if there are strong opinions on the team, it can be hard to break a tie. (We discuss some techniques below that can make this easier.)

It’s also really important at this point to be sure that you’re getting input from stakeholders about all three key aspects of your project’s success: business, technology, and customer experience. Be sure to involve people familiar with these areas in your decision making and use their background and insights to judge what will work best in the area of their expertise. Remember that a product that is terrific in two areas—say technology and experience—but weak in the third, will not be a commercial success. You are looking for ideas that resonate with all three areas: business, technology, and experience.

Here are a few tips and techniques that can be helpful in making collective decisions. Typically, these are not used for major strategy decisions but to unblock forward progress in an uncertain world, where no one on the team is entirely certain about the best course of action. But you won’t know more until you do more iterating, so you need to pick some seeds to start with.

Image First, realize that especially in the first few iterations of the Fast Feedback Cycle, you should not choose a single best idea to move forward with. Rather, you typically select three to five alternatives that seem the most promising to flesh out in more detail. Ideally, these ideas represent a range of relatively safe ideas as well as more novel or risky approaches. It’s a lot easier to narrow and winnow to a small set than to decide on exactly one approach.

Image This is not a one-time decision. Very quickly, you will get customer feedback, and if none of the first ideas pan out, you will have the opportunity to try some others in your next iteration.

Image Keep the focus on ideas, not on people. The point is about trying to find the optimal solution for your customers’ problem; it’s not about having your idea win. It’s actually best if the team takes the attitude that no individual is the owner or the creator of any of the ideas.

Image If your team is relatively small and each member was personally part of every step in the idea-generation process, making a decision may be as easy as a quick conversation that reaches consensus about the most interesting ideas that emerged.

Image If consensus does not come easily, dot voting is a very popular and easy way to use the diversity of a group to make better decisions. Allow a wide variety of people on the team to vote for their favorite alternatives, and then see which items get the most votes. You can do this with a whiteboard full of ideas or with a preculled list of the top 10 or 20 ideas. Give everyone on the team a small number of colored sticky dots (or a marker to draw their own), and ask them to put a dot next to the three or five ideas that they think are the most promising. It’s helpful if the dots are somewhat anonymous—that is, show what you voted simply by marking a dot rather than by signing your initials or your name. Even when a discussion appears contentious, a dot-voting exercise can usually help the team settle on a path forward. The dots show where the real consensus is, which may be a bit different from what the most emphatic debaters are pushing for.

Image Another way to get quick group consensus on a list of issues is to use your thumb. Announce an idea, and have people point thumbs up to indicate that they like that idea or thumbs down to indicate they don’t (point sideways to show indifference or indecision). This works best on a team with a high degree of trust and for a decision that is not particularly contentious.

Deep dive: Group brainstorming

There’s nothing quite like a good brainstorming session to generate a bunch of great ideas and elevate the energy of a team. Maybe you already engage in some level of brainstorming. Even so, this section is still worth reading. To get the most from your brainstorming, it’s important to understand and follow some rules and guidelines.

Follow the ground rules

A set of “brainstorming ground rules” is commonly credited to IDEO, a design consultancy that’s well known for using creative techniques to solve very difficult problems for their clients. The core of these guidelines dates back to the 1950s, when advertising executive Alex Osborn published the book Applied Imagination. These rules have since been widely circulated and adopted throughout the business world, and they have stayed remarkably true to Osborn’s original list.


Although these rules were conceived for brainstorming sessions, we believe they represent broad principles that apply to any sort of exercise whose main goal is idea generation, whether that’s a raucous group brainstorming session at a whiteboard, a quiet sharing of ideas around a table, a sketchfest, or a storyboarding exercise. Posting and following these rules helps create an environment that keeps the focus on generating ideas and helps postpone the onset of tunnel vision.

Image Stay focused on the topic A productive idea-generation exercise has a clear topic in mind, an open-ended question that has plenty of room for alternatives but isn’t so broad in scope as to be unwieldy. A scenario, user journey, customer outcome, or technical challenge all work well as a starting topic. Keep idea generation from drifting too far off topic to ensure that you use your time to fully explore the given problem space.

Image Go for quantity Your goal at this stage is to generate lots and lots of ideas. Success for a brainstorming session should be measured purely by the number of ideas generated in a given amount of time. IDEO recommends numbering your ideas as they are captured to reinforce this goal and also setting an explicit target number to reach in a given amount of time. Sixty to a hundred ideas per hour is considered a good rate.

Image Defer judgment Originally stated by Osborn as “No criticism of ideas,” this principle goes hand in hand with quantity. At this stage, your goal is lots of ideas, regardless of whether those ideas are any good. Wait until later to decide which ideas you will move forward with. Criticizing ideas, even in very subtle ways, can be extremely destructive to the activity, and will quickly stifle enthusiasm and keep participants from contributing for fear of critique.

Image One conversation at a time This practice is needed both to capture an accurate record of all the ideas generated and so that participants hear every idea and can be stimulated by them.

Image Build on the ideas of others Piggybacking, or building on another participant’s idea, is expressly encouraged. In this way ideas bounce around the room and are often greatly shaped by this exchange. Ideas developed this way have no clear ownership, which is very desirable both for cultivating team buy-in and for discouraging personal ownership, which can impede collaboration. Be careful, however, that building ideas on one another aims at generating new variants or combinations of ideas, not just digging into the details of an idea, which can lead to tunnel vision.

Image Encourage wild ideas Creating a playful environment and team culture that explicitly encourages wild, impractical, or exaggerated ideas is ironically one of the most powerful ways to supercharge idea generation, as these unusual ideas can sometimes lead to a breakthrough that no one had ever considered before. This is critical behavior to develop, yet it is also one of the hardest for practicality-focused engineers to become comfortable with.

Image Be visual (or physical) Don’t be afraid to sketch a picture, act out an idea, or grab a few items in the room to mock up a quick physical prototype. Sometimes words aren’t the best medium with which to capture an idea. A rich, flexible environment, with toys, models, materials, and lots of drawing surfaces, helps stimulate ideas and the demonstration of ideas in other modes.

Facilitating a group brainstorming session

The secret to a successful brainstorming session is a great facilitator. With a skilled facilitator, a brainstorming session can be hugely productive, even with a fairly novice group of participants. However, with a poor or inexperienced facilitator, a session can easily go south for any number of factors. Here are the most important things for a facilitator to keep in mind:

Image Pick an appropriate space The environment can do a lot to both shape the mood and encourage divergent ideas. An informal room with couches and wall-to-wall whiteboards sets a mood quite different from a conference room dominated by a long table. Both environments can work, but we prefer using a more informal space when possible. Bring in physical artifacts that can serve as jumping-off points, inspiration for ideas, or fodder to explain an idea. These might be office supplies or even toys, such as Legos, building blocks, or plastic figurines. You may also bring in items related to your specific problem space, such as photos of your target customers or posters containing key insights from your research.

Image Invite the right people Get a good mix of experience, disciplines, backgrounds, and perspectives on the problem. Consider not including the manager or leader of the effort if their presence might prevent people from contributing wild ideas, especially until a brainstorming culture is well developed within the team. Keep in mind that given the limitation of 1 person talking at a time, having more than about 20 people in a room makes it hard for everyone to participate fully. For larger groups, schedule multiple sessions or encourage team members to vote with their feet to attend brainstorming sessions on certain topics but not others, based on their interests and passions.

Image Set the mood A productive brainstorming session is playful, informal, open to new ideas, and absolutely welcoming of every person and every idea, no matter how seemingly outlandish. Getting this wrong is the single biggest reason a brainstorming session fails—participants are unwilling to contribute because they fear criticism or worry that people will think their ideas are foolish.

Image Set the ground rules and enforce them Post rules visibly (see “Follow the ground rules,” earlier). Firmly correct and redirect any comments that don’t follow the rules, paying particular attention to anything that could be interpreted as even a subtle criticism. If left unchecked, this can cause many people to clam up and dramatically impact the effectiveness of the activity.

Image Keep the group focused on the chosen topic Start the session by reminding everyone who your target customer is and the scenario or problem statement. As the conversation progresses, allow some leeway for tangents, as sometimes these can be productive sources of alternative approaches and creative ideas, but a facilitator needs to be ready to pull the conversation back if it strays too far afield for more than a few minutes.

Image Capture every idea in a visible way Some facilitators will keep a running list by writing each idea on the whiteboard as it is contributed. Others will delegate the job to someone else, who writes on the whiteboard or types on a laptop that is projected onto a screen. Sometimes participants or a scribe capture each idea on a sticky note and post that on the wall or on butcher paper. If the idea is presented in visual form, the contributor should be the one sketching it on a sticky note or the whiteboard to capture it. The important thing is to capture all ideas so that everyone can see them. As the session progresses, being able to point to a previous idea and notice a connection or a possible blend is a powerful way to generate additional ideas.

Image Keep the pace up Encourage a fast pace, and make sure that your scribe can keep up with the conversation. Speed encourages more flexible thinking and a greater likelihood of seeing connections between ideas. Don’t go longer than about an hour of focused work—you just can’t keep up the pace much longer than that.

Image Constantly offer the reminder, “How else could we solve this?” Remember that the goal of a brainstorm is to generate as many ideas as possible. The facilitator needs to be vigilant to ward off tunnel vision and to not allow the group to discuss any idea in too much detail. Keep pushing the group to look for new approaches. Assume that there is always another approach you haven’t considered. Even if you don’t see it at the moment, someone else in the room invariably will.

Image Welcome offbeat ideas If an offbeat idea comes up, the facilitator should be extra supportive to help people suspend disbelief and give it due consideration. Play with it, extend the idea, see where the thread leads. Especially with a team that is new to brainstorming, it’s important for the facilitator to encourage wild ideas—to hold open the door for novel, out-of-the-box ideas to emerge.

Image Manage a dry spell If the conversation stalls, inject a provocative question or a lateral thinking technique to get things moving again. Generally speaking, it’s best for the facilitator to not contribute ideas to the brainstorming session, especially at the beginning. If the facilitator offers ideas early in the session, it can set the wrong tone and encourage people to sit back and let the facilitator take the lead rather than have everyone jump in and contribute. However, later in the session, if the well starts running dry, the facilitator might inject an unusual, thought-provoking idea to help get things moving again.

Image Save decisions for later Don’t end a brainstorming session with a discussion or decision on which ideas will get pursued. And certainly don’t allow tradeoff conversations to arise in the middle of your idea generation. Those are just other ways to judge ideas, and we know what happens when you unintentionally introduce judgment and criticism into brainstorming. Focus this time just on idea generation. Schedule a separate time to sort through them, make sense of your work, and pick a few ideas to flesh out in the next stage of the Fast Feedback Cycle.

Image Make sure the results get used One of the biggest problems with even a productive brainstorming session is that the results get forgotten or sidelined. Perhaps only one idea was actually chosen to move forward with, and the rest were forgotten. Or worse, the brainstorming session was a token activity to appease the team when the real plan was generated behind closed doors by the senior leaders, and may have little relationship to the brainstorming activity or its results. This is a sure way to tank any future attempts at brainstorming, not to mention team morale. Be sure results get saved in an accessible location. Be sure that they are used, both in the current iteration as well as when they’re needed for inspiration in future iterations of the Fast Feedback Cycle. Communicate back to the participants how the ideas were used and what the next steps are.

Concluding a brainstorming session

You can use the end of a brainstorming session to get the opinions of the people who attended. As people leave the room, have everyone vote for three to five of the ideas that seem most promising to them. A vote could be as simple as a check mark on the whiteboard, or ideas can be flagged with a sticky note or a colored label. These votes aren’t meant to be binding, but they help alert the facilitator or the owner of the decision area to promising ideas that might merit further investigation and more detailed exploration.

SFE in action: In defense of brainstorming

Scott Berkun, best-selling author of The Myths of Innovation (

Periodically, popular articles arise decrying how flawed brainstorming is. Jonah Lehrer wrote a popular article in The New Yorker, “Groupthink: The Brainstorming Myth,” but there have been, and will be, many others. Most of these articles have poor frameworks, miscasting what brainstorming was designed to do and how ideas in workplaces are actually developed.

I have no stake in brainstorming as a formalized thing. Even in my essay “How to Run a Brainstorming Meeting,” I explain its strengths and weaknesses. It’s a method, and I’ve studied many idea-generation methods. If done properly, in the right conditions, some of them help. I’m not bothered by valid critiques of any of them. However, sweeping claims based on bad logic and careless thinking need to be addressed.

Here are four key things Lehrer doesn’t mention, which shatter his conclusion:

Image Nothing matters if the room is filled with fools or strangers (or both). If you fill a room with thoughtless people who do not know each other, no method can help you. The method you pick is not as important as the quality of people in the room. The most important step in a brainstorming session is picking who will participate (based on intelligence, group chemistry, diversity, etc.). No method can instantly make fools smart, the dull creative, or acquaintances intimate. The people in [Charlan] Nemeth’s research study, the one heavily referenced by Lehrer, had never met each other before and were chosen at random. A very different environment than any workplace.

Image Brainstorming is designed for idea volume, not depth or quality. [Alex] Osborn’s (the inventor of brainstorming) intention was to help groups create a long list of ideas in a short amount of time. The assumption was that a smaller group would review, critique, and debate, later on. He believed most work cultures are repressive, not open to ideas, and the primary thing needed was a safe zone, where the culture could be different. He believed if the session was led well, a positive and supportive attitude helped make a larger list of ideas. Osborn believed critique and criticism were critical, but there should be a (limited) period of time where critique is postponed. Other methods may generate more ideas than brainstorming, but that doesn’t mean brainstorming fails at its goals.

Image The person leading an idea-generation session matters. Using a technique is only as good as the person leading it. In Nemeth’s research study, cited in Lehrer’s article, there was no leader. Undergraduates were given a short list of instructions: that was the entirety of their training. Doing a “brainstorm” run by an fool, or a smart person who has no skill at it, will disappoint. This is not a scientific evaluation of a method. It’s like saying “brain surgery is a sham; it doesn’t work,” based not on using trained surgeons, but instead undergraduates who were placed behind the operating table for the first time. (See Scott G. Isaksen and John P. Gaulin, “A Reexamination of Brainstorming Research: Implications for Research and Practice,”

Image Generating ideas is a small part of the process. The hard part in creative work isn’t idea generation. It’s making the hundreds of decisions needed to bring an idea to fruition as a product or thing. Brainstorming is an idea-generation technique and nothing more. No project ends when a brainstorming session ends, it’s just beginning. Lehrer assumes that better idea generation guarantees better output of breakthrough ideas, but this is far from true. Many organizations have dozens of great ideas but fail to bring those ideas into active projects, or to bring those active projects successfully into the market.

Brainstorm stage: How do you know when you are done?

The Brainstorm stage is critical for generating and exploring many diverse ideas, giving you the best chance at hitting on a truly optimal solution that balances your business needs and technical abilities and delivers a great customer experience.

There are lots of techniques you can use to generate ideas, from brainstorming, to sketching storyboards, to modeling with clay. No matter which techniques you use, you know you are ready for the next stage when you have generated:

Image A collection of the alternatives you have generated, which represent a dozen (or perhaps lots more) meaningfully different approaches to solving your problem or scenario, not just cousins of the same basic idea. Your collection should include some sketches or other visualizations, not just lists of words, and be archived in a lightweight way that you can refer to in future iterations.

Image Three to five alternatives chosen from that larger collection as the most promising, which warrant deeper exploration in the next stage.

The good news is that the Brainstorm stage is the stage in the Fast Feedback Cycle that likely needs the least amount of total time investment. With as little as a few concerted hours of idea generation, you should be ready to move on to the next stage and start prototyping your most promising ideas so that you see whether they actually resonate with customers the way you thought they might.


1. Sam Biddle, “The Guy Who Invented Your Facebook News Feed Just Quit Facebook,” Gizmodo, June 15, 2012,; Farhad Manjoo, “Facebook News Feed Changed Everything,” Slate, September 12, 2013,;

2. Alexia Tsotsis, “Facebook’s ‘Like’ Button Used to Be the ‘Awesome’ Button,” TechCrunch, October 5, 2010,; Om Malik, “Why Facebook Wants FriendFeed,” August 10, 2009,

3. Pitney-Bowes judged that these would be a precursor to the paperless society, which would have deep implications to their core business of selling postage meters and paper-handling equipment for the mailroom and the post office.

4. Pauling was awarded the Nobel Prize in Chemistry in 1954 and the Peace Prize in 1962. Marie S. Curie also won two prizes, but she shared one of them (the Nobel Prize in Physics in 1903).

5. Arguably, the complex problems we solve have many more dimensions, but that’s tough to visualize, and three dimensions illustrate this concept well enough.

6. Stretching the analogy further, one could say that the complex plane is changing over time. What was an optimal solution five years ago is no longer as noteworthy, as new technologies and possibilities enter the market. Disruption becomes possible when a new global maximum appears that is not noticed by the incumbent.

7. Alex F. Osborn, Applied Imagination (New York, Scribner, 1979).

8. Edward de Bono, The Use of Lateral Thinking (London: Cape, 1967).

9. The full article is available at A related article (“In Defense of Brainstorming”) is at You can read Lehrer’s original article at