A Baby with a Royal Platinum Card? Personalization Fails and How to Stop Them

Ever think of getting your kid to charge your next vacation on his credit card? That might have been a possibility for the parents of 2-year-old Evan Hart, who were bemused when their little boy was offered a credit card with thousands of frequent-flyer miles because of his “admirable credit history” and, presumably, impeccable choice of airline.

Then they remembered they’d bought him his own individual seat on a transatlantic flight, rather than spend 10 hours holding him as a free-riding (but heavy and squirmy) “lap baby.” Little Evan couldn’t actually say “major personalization fail,” but his parents could, and they had a great laugh with all their friends, family and social media contacts.

“The worst part is that neither my wife or I got the same offer,” says Evan’s dad with a chuckle.

Good Personalization Pays, But Bad Personalization Fails are Stinky and Sticky

Targeted personalization has become an essential part of the marketer’s toolkit—and for good reason. By tailoring experiences to individual customers, you can win their loyalty, boost conversions, and increase your revenues over time. The ever-increasing power of data—and the martech solutions that use it—makes such approaches more potent than ever.

At the same time, not all martech solutions are up to the task, and many use too little data—or the wrong data—to do a complete job. And personalization that’s just a little bit off entails brand risks ranging from bemused or confused customers to those who take lifelong offense at really big fails. Imagine how you’d feel if you searched for suicide prevention hotlines and techniques to help a depressed friend, and then started getting offers for coffins and funeral homes?

It’s easy to see how if you make a misstep, you could alienate the very people you want to reach, and even end up in legal trouble. The growing availability of personal data can be a mixed blessing. Data-driven personalization can improve your interactions with customers—but it can also give you more ways to stab yourself in the foot.

So what are some common ways to fail at personalization—and strategies to help your business steer clear of these pitfalls?

1. Personalization that targets the wrong users

   Shutterfly had a marketing misfire in 2014 when it sent an email intended to congratulate new parents on their babies. Instead, the message went out to many users who had no children at all, including some who were coping with painful experiences such as miscarriage or infertility.

   So how does this kind of mistake happen? Human error is one likely cause. (Shutterfly apparently intended to send the email out to customers who had recently bought birth announcements.) But such snafus become far more likely if you’re working with incomplete, inaccurate or outdated data.

   Possible solutions:

   • Base your decisions on complete customer profiles, not unverified assumptions.
   • Continuously refine and update the data in your profiles. Break down data silos within your organization, and use all the data sources at your disposal to provide accurate insights into your individual customers. For example, many companies use enterprise customer data platforms, or CDPs, to build complete profiles and provide near-real-time updates.
   • Build up your stores of internally collected (first-party) data, which tends to be relatively reliable, rather than relying entirely on third-party sources.

2. Creeping out your customers with offers that are too personal

   At the opposite extreme, some personalization can be disturbingly accurate from the consumer’s point of view.

   If you were a senior in assisted living, how would you like to receive a Christmas basket from a nearby mortuary? You might feel like the people who sent it know all too much about you—and don’t necessarily have your best interests in mind.

   With vast amounts of customer data available, your company may be tempted to push personalization as far as possible. But as it turns out, most people don’t like businesses watching them too closely. In a recent survey by InMoment, 75 percent of consumers found most types of personalization to be at least somewhat creepy. What’s more, they said creepy behavior would influence how they acted toward a brand.

   To complicate the issue, different individuals and demographics may tolerate different levels of personal outreach from brands. What strikes you as innocuous and benign might seem intrusive or aggressive to someone else.

   Possible solutions:

   • Be selective in using data, especially data that customers are likely to consider personally sensitive, such as financial or health status.
   • Use more subtle approaches that don’t reveal intimate knowledge of customers’ lives. For example, a home supply company could tailor content differently for couples and single people, but in a way that doesn’t directly relate to their relationship status.
   • Don’t make assumptions about your audience. Instead, find out what kinds of personalization your specific customers like and dislike. Personalized discounts or shopping experiences, for instance, might be more popular than targeted ads.
   • Obtain consent, and be transparent about the information you collect. Customers are more likely to approve of personalization when a company uses data they agreed to provide.

3. Breaking privacy laws like the GDPR or the CCPA

   Governments around the world are taking bigger and bolder steps to protect online privacy. As a result, over-aggressive personalization can have serious legal costs, even for what seem like minor infractions.

   In January 2019, for instance, French regulators hit Google with a $57 million fine for violating the European Union’s General Data Protection Regulation (GDPR). Google’s infractions? Failing to explain how it collects users’ personal data, and to gain their consent before showing targeted ads.

   No doubt you have every intention of following the law. Even so, it requires constant effort to keep up with new laws and regulations at both the national and state or provincial level. And even though some laws sound regional, regulations like the California Consumer Privacy Act (CCPA) often end up behaving a little like national laws, since nobody really wants to have to operate on multiple standards.

   Faulty data collection and integration could also mean trouble, even if you’re trying to stay compliant. For example, your customer profiles might not contain an adequate record of how and where you collected data from users, or whether or not they’ve opted in to receive personalized communications.

   Possible solutions:

   • Ensure someone in your organization is responsible for understanding evolving privacy requirements, and include this person’s input in your decisions about personalized marketing.
   • Adopt a consistent set of best practices for personalization, taking into account the laws in all jurisdictions where your company operates. Update these guidelines as new laws and regulations emerge.
   • Ensure you have enterprise-level systems, such as a CDP, that can collect, consolidate and maintain all the data needed to protect users’ privacy.

4. Ignoring context, or not updating campaigns when context changes radically

   “Congrats, you survived the Boston Marathon!” That was the subject line of an email a major shoe brand sent out to runners who finished the iconic race in 2017.

   The message would have been fine if it had referred to almost any other athletic event. But for customers who remembered the deadly bombing at the Marathon just four years earlier, the wording came across as insensitive, and the company’s CEO rushed to issue an apology.

   As this example shows, it’s easy to overlook the context of your personalized campaigns, with potentially disastrous results. In some cases, the problem isn’t even lack of knowledge, but failure to connect the dots.

   Possible solutions:

   • Take the time to get personalization right. Use detailed, complete unified profiles to craft your messaging, and Test content on a small scale before deploying it more widely.
   • Gather relevant data about the context. This could involve steps as simple as running a Google search for relevant keywords — for example, “survive” and “Boston Marathon.”
   • Train your personnel in brand safety and build consideration of such issues into your planning and creative process.

5. Letting automation run amok

   A troupe of animated tomatoes and carrots are dancing the can-can, celebrating a popular brand of vegetable juice. What could be more innocuous?

   The only problem: this cheery popup ad is playing over a tragic article about a woman trapped in a “persistent vegetative state.” She’s “a vegetable,” according to someone quoted in the piece.

   This all-too-true story points to the risks of automated targeting. When machines make decisions with inadequate guidance, the unintended consequences can be brutal.

   In this case, the ad server was just following orders by responding to the keyword “vegetable.” Unlike a human being, it couldn’t judge what the word meant in context—or understand why it should avoid displaying that particular advertisement in the middle of a profoundly moving story about a family’s agonizing dilemma.

   Possible solutions:

   • Consider the risks of each campaign, and try to build in failsafes. The creator of the dancing veggies ad, for example, could possibly have avoided the problem by including “vegetative” or “coma” as negative keywords.
   • Focus on types of personalization that rely on genuine understanding of individual customers, rather than forcing machines to make risky guesses about their identity or interests.

Is Targeted Personalization Worth the Hassle?

If you ignore personalization, you may be ceding ground to your competitors. For example, 80 percent of consumers in a 2018 study from Epsilon said they were more likely to buy from companies that provide personalized experiences.

The right technology can help you create such experiences while limiting your risks. An enterprise CDP, in particular, unifies all your data and provides actionable insights for personalization across all channels. By using Treasure Data as their CDP, companies such as Kirin, Shiseido and Wish have served customers with personalized shopping, product recommendations and more.

And as mentioned before, a CDP can help you avoid personalization fails by creating an accurate, complete and up-to-date view of each individual customer. But technical fixes are only part of the solution. Your entire marketing organization needs to recognize the potential downsides of personalization, as well as its benefits—so that you can do it better than ever.