Trick or Treat? Five tips to avoid the pitfalls of personalization

Executive Director & Partner
New York

The world is saturated with choices: There are over 30 million songs on Spotify; 25 million people looking for love on Match.com; over 2,500 channels on Emirates’ in-flight entertainment system and more than 14,000 titles on Netflix. So how on earth does one make a decision without an educated nudge in the right direction?

The answer is personalization: it is one of most effective ways for people to differentiate between strong or weak products and services and good or bad experiences.

However, there’s a fine line between personalization being a valuable, useful tool for customers, and it being potentially overbearing — or even worse — creepy. Companies potentially know quite a lot about their customers, so how that information is surfaced and used is very important to gain and retain trust.

For example, MagicBands at Disney resorts are considered a highlight to those who pay for them, but if all employees automatically called kids by their first names without an introduction, it might be magical to kids — but probably creepy to parents. Perhaps noting their favorite Disney song or character, or indeed phrase, could make it an even more personal and thus valuable, while omitting some of the inherent creepiness.

Indeed, nobody likes seeing the same online ads stalking them across multiple platforms; or crawling chatbots that can unintentionally alarm people, in turn creating an overall negative experience.

Here’s how to utilize personalization without sending shivers down the spine of your customers.

1. Beware of making a monster

Whatever product or service you’re creating, it needs to work for busy — not necessarily tech-savvy — but privacy-wary people. It’s worth carefully considering what level of interaction is appropriate; what information is needed; and whether you are simply bolting on unnecessary features that actually add no value at all.

People expect a certain level of personalization with digital products but there’s a difference between passive and active gathering of information and how you use it, be aware of what is acceptable for each situation. You don’t expect Mickey Mouse to know everything about you when you arrive, but if you’ve told Disney and you know your data will be used that way and you expect not to tell them again, then it’s Ok.

2. People are seldom straightforward

People don’t just change their behavior on a full moon. Be careful how you profile and categorize your customers because a bad personalization experience can be exasperating and unforgettable for all the wrong reasons.

The Hilton Hotel is an example of good personalization practice. In fact, the high-end hotel chain wants to know everything from guests’ ideal room temperature to the shows they like to watch on Netflix to the drinks they’re most partial to — all for when their guests arrive. This, at least, is what they’re now testing and it’s a step in the right direction. Crucially, what Hilton want to do is allow their guests to customize their own experience by storing their preferences. This is where customization and personalization cross paths for the good, giving a certain amount of power to the customer.

Likewise, categorizing your customers in wide groups is not always a good idea, context must always be considered with a side note that not everyone is the same; the time of day can be a crucial element (how one feels in the morning may be very different in the evening). Furthermore, it’s always worth building up a level of trust with your customer to understand who they are and what their needs are. Using personal data well will always trump the use of more generic demographic statistics or worse — assumptions.

3. Be helpful, not creepy

The whole idea of personalization is that it’s meant to improve a customer’s experience, making it easier for them to navigate their way around a multitude of available choices, rather than merely grabbing a load of data and carelessly throwing it back at them. Getting the right balance can be difficult, but it’s worth investing the time and often money — in getting it right.

The alternative can be like an overwhelmingly creepy digital friend who knows too much personal stuff about you while at the same time not actually caring to find out what you like.

A chatbot from a bank, for example, that addresses you by name whilst telling you how much money you spent the night before can feel a bit too invasive, a bit like you’re being watched by a nagging parent. Alternatively, a bank that offers decent money management tools adapted to address your spending habits could be far more useful and valuable. It would certainly be more preplanned, and as a result more thoughtful and less reactionary (people don’t always like to look in the mirror).

4. Not all surprises are alarming

True personalization puts the customer at the heart of a service, and customers want surprise as much as they want familiarity and consistency.Never design out surprise from your personalization plan as it can be a brilliant experience for customers if executed well.

To get the surprise element right, it may be good to rule out all data garnered from past behaviors, and instead, pool suggestions based on behaviors of customers with similar tastes or something completely random to avoid any echo chamber consequences. A good example can be found at Spotify, which regularly serves up suggested content to its users that the users will — crucially — enjoy. It also needn’t be a sea of surprises; with personalization, less is often more.

5. Avoid being a data snatcher

This can be a real cliffhanger for your customers; few things disappoint people more than being intentionally scammed, which is exactly how many do feel if data collection is shrouded in murky terms. Put simply: be transparent about how you use data.

Unroll.me — a free service that allowed customers to exorcise unwanted, spammy emails — found this out the hard way. Unknowingly to many of its customers, they monetized their business by selling their users’ data on to competitors. The devil was in the small print. Of course, there was an uproar when it was discovered that Uber, the car-hailing app, was buying information on Lyft customers, its rival.

What unroll.me could have done, is perhaps allow its users to choose which services or products — adapted to their interest — they’d like to share their data with, in exchange for purging all of the junk that ends up in their inbox. Instead, they betrayed their customers’ trust — and that, ultimately, is the business that most are in: the trust business.

At Smart, we have been working for many years designing solutions to optimize the customer experience, and what we’ve learned is that personalization should never feel like a stalker who offers ill-informed advice and solutions that will for sure further alienate people. Ultimately, a good personalization experience should feel like a connection with a friend, someone who tries to understand you and who intervenes thoughtfully to give sound, timely advice that is both valuable and trustworthy.