The First Rule of Personalization: How to Not Annoy the Customer

Personalization is more than an advertising industry buzzword—it is one of the most common goals of modern marketers. But the term’s meaning isn’t quite so clear. To many, it is even elusive. So, what does “personalization” mean? Is it about: Addressing your customers by their first name?Collecting explicitly expressed preferences and treating them accordingly?Customizing emails and landing pages based on customer profiles and behaviors?Knowing when to contact and through which channel?Keeping in touch with your customers constantly?Suggesting more of the same products that they just purchased via collaborative filtering? None of these are wrong answers. Of course you have to address your customers by their names and remember their preferences. It would be a disservice to collect data and not use it. The Real Deal on SMS Marketing Let’s say you clearly stated to a hotel chain that you prefer a room with a king-size bed near an elevator on a not-so-high floor. It would be annoying if the hotel regularly ignored your preferences while calling you a loyal customer—“Stephen, thank you for being a loyal member of our club for such a long time! Here is the key to a corner room on 21st floor.” If all you get out of the long-term relationship is just an extra bottle of water upon check-in, well, you may not feel so special after all. Personalization: A starting point Communications that connect through all imaginable channels—including offline touchpoints—must reflect the customers’ preferences. You don’t even need any analytics or modeling to satisfy your customers’ specific demands. All of this is just a starting point of basic personalization. But you know things are not that easy. Even without having to “predict” what the customers are about, displaying custom contents to each individual isn’t a simple task. First, you’d need a great variety of creative content. Second, you must possess display capabilities to show different things to different people. Let’s imagine a landing page of a retail site. Through past purchase history and various preference indicators, you may able to tell your target’s favorite items without any machine learning. But does your site support such dynamic display? Do you have enough creatives for such efforts? Many companies think that it has reached the goal line when it can show unique contents to specific visitors. However, reacting to known explicit information is more of a prerequisite to holistic personalization—not the end of it. Usually, only a very small fraction of customers actually let marketers know what they are about. The burden is on the marketers to figure them out, employing both explicit and implicit data to deliver the requisite insights. Further, in a multi-channel environment, in addition to knowing who to contact and what to offer, marketers must also consider the channel through which they're contacting the customers and when they're contacting them. Knowing what to ask Too many marketers think that “all the time” should be the default mode in 1:1 marketing. I happened to have opted out of many of my favorite brands, as most just didn’t know when to leave me alone. I may like the products they offer, but not enough to hear from them two or three times a day through various channels. Knowing when not to bother customers is part of personalization, too. Going beyond explicit preferences requires advanced analytics. Real insights into the consumers’ openness to receive messages would give marketers clear answers to questions such as “what to mention,” “when,” “how often” and “through what channel.” There really is no other way to escape that “known information only” trap. When I say analytics, I am talking about something more advanced than product-centric collaborative filtering (as in, “Oh, you just bought that item, so you must like this one, too”) based on what other customers did. If you leave such an engine on an autopilot mode, you may end up annoying your customers in the name of personalization. “Ok, I bought an air filter from you just now—is it too much to ask not to push another air filter for some time?” is an example of what goes through the minds of consumers. Product-centric personalization is better than nothing. But can we please go beyond what we’ve been doing for past 10 or 15 years and be a little more sensitive about it? We are dealing with live human beings—and we all know these modern shoppers are easily annoyed. People are not just mere sums of their product purchases. Personalization should be a series of gentle nudges towards something they might like, not pushing what marketers want to sell down their throats. This type of paradigm shift is not simple or easy. But for now, remember that all the data that you collected must be converted into “descriptors of individuals (or personas)” through analytics first. Otherwise, you will always be stuck at a rudimentary level of personalization where you’re ultimately seen as more annoying than helpful. How Small Brands Turn Customer Feedback Into Marketing Strategies