Considering the Ethics of Customer Analytics - Analytics for Product Development - Customer Analytics For Dummies (2015)

Customer Analytics For Dummies (2015)

Part IV

Analytics for Product Development

Chapter 16

Considering the Ethics of Customer Analytics

In This Chapter

arrow Getting informed consent

arrow Continuing ongoing experiments

In 1963, Yale Psychologist Stanley Milgram paid volunteers $4 to “teach” another volunteer new vocabulary words. If the learner got the words wrong, he or she received an electric shock! Or so the participants believed … no shocks were actually given. Regardless, a majority of participants believed that they were shocked, which led to a concern over the ethical treatment of study participants.

Even though no one is getting shock treatments for wrong answers anymore, the debate over ethical treatment continues. There will always be problems with the way some companies handle data, especially companies that exploit new technology that people aren’t familiar with. But you can avoid being one of those companies.

warning If you don’t treat your customers or study participants in a fair manner (or in a manner they expect), not only will you risk losing your current customers, but it also may turn loyal customers who are currently promoters into former customers who discourage others from purchasing or using a product or service.

Getting Informed Consent

In the years since the Milgram studies, it’s become common to provide participants with an informed consent form. It lets participants know the general topic of the study, and informs them that they can stop (opt out) any time, regardless of what the experiment demands. Here’s how you go about getting informed consent from your participants:

· Get signatures. Even when you’re not conducting a lab-based study, you’re collecting data from volunteers and subjecting them to questioning, analysis, and observation. If the data deals with sensitive information (financial data or other personal data), consider having them sign an informed consent document stating that they understand the purpose of the test, what their participation means, and what their remuneration (if any) is.

· Offer a verbal explanation. Whether or not you provide a formal document for participants to sign, always explain what participants will be asked to do (at least in high-level terms) and ask if participants understand or have questions.

· Be as clear as possible. A lot of user research happens well beyond the confines of a laboratory. No matter what environment you conduct your experiments in, always make it clear what users are volunteering for. New technology and methods can lead to a blurring of ethical lines.

· Include how you are using their data. Everyone always wants to know how his or her data is being used. Reassure participants that their data will remain private.

In the following sections, I discuss some interesting experiments conducted by well-known brands that ultimately went wrong because their users didn’t understand what they were consenting to. I also explain how you can avoid the same mistakes.


Early in 2014, Facebook made headlines for conducting a large-scale experiment on a small fraction of its billion users — small still being 700,000 users.

Users unknowingly had their newsfeeds manipulated for one week to show them either positive or negative postings. The results of the study showed that exposure to positive posts resulted in users producing positive posts, too. The same was true of the negative posts. In other words, emotional sentiments were contagious. Good news didn’t lead to people feeling glum; it actually led them to feel better.

The results were quickly overshadowed by outrage over Facebook using information to manipulate users.

Facebook discloses the latitudes it can take in its privacy policy and terms of use, but few people read, much less understand, the implications of terms and conditions and privacy policies. For example, on average, customers spend just six seconds looking at the terms and conditions of their software products. These are documents that, if read in full, can take several minutes to read, so those six seconds probably represent the time it takes customers to click “Accept” without considering what they are consenting to.

The big takeaway for you is to always make sure your participants know they’re agreeing to participate in a study, even if they sign a consent form. After all, those Facebook users all agreed to Facebook’s terms of uses, but not many knew that meant Facebook could filter what they saw on their timelines.


The dating website OKCupid admitted that, among other things, it paired up people who were poor matches according to their dating algorithm. The results of the experiment suggested that the act of telling someone that she was a good match was as important as actually being a good match. Users were notified that they were involved in a study and were shown the correct compatibility percentages after it was concluded.

This sort of manipulation was likely covered under the terms and conditions the users agreed to when using the website. But just like with Facebook, most users probably didn’t read or understand those terms and conditions.

Also, OKCupid told users about the study after it was over. To avoid the same mistake, be up front with customers about how information may be changed and give them a way to opt out. If you explain how the data manipulations and analysis will ultimately benefit them, you may have a high opt-in rate, and likely fewer irate customers.

Amazon and Orbitz

Amazon has been a pioneer of many things on the web. In 2000, it was revealed that Amazon was adjusting the pricing of some of its products based on past browsing behavior. So one customer would pay more or less than the next person who bought the same item.

Orbitz also came under fire for revealing that it prioritizes hotels based on data that showed Mac users were 40 percent more likely to book a 4- or 5-star hotel than PC users. Mac users would be shown more expensive alternatives when searching.

The line between using customer segments to personalize products and experiences is a fuzzy one. Be sure customers understand how recommendations will change based on their preferences or profile.

The financial planning website invited some participants to use a new beta feature that separated business and personal accounts. After a year of collecting data from users who meticulously entered financial information, turned off the feature without notice. Previously entered data and reports were no longer accessible. These actions were permissible under the terms of use.

Your takeaway is to always tell participants what you’re doing with their data, and what access they’ll have to their own data. Most people are reassured when their data ultimately remains under their control.

Deciding to Experiment

Most people understand that their actions and data are being monitored and used for commercial purposes. But you can minimize their concerns by following these tips:

· Privacy/anonymity: Where possible, ensure both privacy and anonymity. When you collect survey data or conduct a usability or findability study, avoid collecting personally identifiable information unless it’s absolutely necessary. When it is, be sure the participant is informed and take effective measure to keep that information secure. The National Security Agency (NSA) has made headlines for how it has data on millions of Americans and that data is associated with names, Social Security Numbers, and addresses. In the commercial world, real names and identities are generally less important.

· Disclosure: At the very least, you need to disclose what you’re doing — or might do — with user information. Write the language in the terms of use clearly. Avoid legalese that few understand.

· Retention and access to customer data: If you collect data from customers, especially data that requires labor (like providing content), make it available for those customers to download or use in other forms. The more sensitive the data (family photos, financial information, taxes), the more important it is. If that’s not possible, make it clear up front.

Privacy, disclosure, and anonymity are necessary but not sufficient when your goal is to treat your participants in an ethical manner. You can take the additional step to make sure users comprehend what they’re participating in. This gives users an opportunity to opt out of your study if they want to.

tip Test how well users actually understand the terms they’re agreeing to. You can set up a study to assess how well participants comprehend the terms and conditions. For example, I worked with a large credit card issuer and presented several variations on wording and images so customers understood things like interest rates and fees. A series of open- and closed-ended questions was asked to assess their comprehension after being exposed to the terms. Terms were picked that led to the highest level of comprehension and helped iterate the language to write more understandable terms.

The popular practice of A/B testing is one of the most effective ways for website owners to understand which design elements lead to higher purchases, donations, or registration rates (see Chapter 10). But at what point does A/B testing — and other techniques that manipulate customers and their data — become unethical?

Professional organizations like the User Experience Professionals Association (UxPA) available at and Direct Marketing Association (DMA) have codes of conduct that offer a good guide for researchers ( Among other things, the codes cover similar topics I discuss here, including consent and disclosure. The organizations’ guides also delve into marketing to children and using sweepstakes.