top of page

Personalization is broken: Why it is time to redefine the approach

Working in the field of personalization and data analytics for the larger part of my professional life so far, I saw a lot of different, a lot of good attempts to personalization. But what I also witnessed is that all the attempts stumble into a common pitfall and this is the time to change our approach of using personalized data.

Don’t get me wrong, personalization is great. In the era of endless amount of information, products and contents, in the era of global digital economies, personalization is one way to provide a guidance for the user. But I am also convinced that personalization solely executed in its pure form is broken.

And Here is Why

You do not only want to see things you like or things other statistically comparable people like. It is great yeah, because humans tend to have a very consistent taste and interest in information. It is the way our brains work. And peer groups based on common interest are also the basis for our societies and have been since the beginning of humanity.

But the problem is: we are not only people focused on our peer groups. We are also individualist who like to explore, to experience something new. Expect the unexpected.

Mirror this human behavior to the web and you will see why personalization is broken. In a hyper personalized world, I can only see the information that conforms my view, I can only buy products that suits my style, I will only be surrounded by things that I am already interested in. You see where this is going? Yeah, you will be stuck, stuck in your own little filter bubble. And I tell you what: this will get really boring. Where is the fun in that, where is the discovery in that?

Pure Personalization Can Lead to a Higher Churn-Rate in the Long-Term

Now look at the business side of this. If you are a publisher or an online retailer or every other online business model, chances are that you focused the latest years on collecting more and more data to build better user profiles, to better predict the behavior of the user. And this is good, because it required you to rethink the way you build products, tackle marketing strategies and provide content and shops. Data-driven or data-informed business processes replaced the gut-feeling decisions with evidence-based decision.

But the more and the longer you personalize, the higher there is the chance that the customer will get bored and jump to another website, to another shop or to another tool. Instagram did this with Facebook a few years back. Instagram provided a new way to discover information because Facebook became more and more a filter bubble. Instagram provided a fresh, algorithm-free newsfeed based on recency. And it has long declined to change the way its newsfeed worked. Sadly, a year ago Instagram has ditched the chronological order for an algorithm (but this decision is part of another story).

Exploration vs. Personalization

For me personalization is one pole on a pendulum and the other one is exploration.

Exploration is “the activity of searching and finding out about something” — Cambridge Dictionary

Since a few years the pendulum was swinging in the direction of personalization. But it is still a swinging pendulum and, hence, it is swinging back and forth. It will develop like every other trend.

Example: Look at flat Design and the Microsoft approach a few years back. Microsoft was the promoter of a true minimalistic and flat design as the antipole to Apple’s sceumorphism. The years passed and all major players adopted some trends of flat design. But look at it today: flat design is not only flat, it is almost flat, it is still touchable (best seen in Google’s material design and Microsoft’s refined design guidelines). It is combining the best of both worlds. Because usually the radical approaches aren’t the one to survive. They help defining a new way of thinking, but in the end the different poles converge into best-of-breed approaches.

How to Fix Personalization?

So, the question we should ask is: how can we fix personalization by leveraging the whole exploration-personalization-spectrum?

In short: It is balancing product development, marketing etc. between personalization and exploration.

In detail: It is a two-step guideline.

First: Declare the User to Be of Age

The first step to a solution is to declare the user to be of age. Give the user transparency about personalized areas on your site. Let the user decide if he wants to temporarily opt out of the filter bubble. Let him see how things would look if we didn’t know anything about the user. But give him also the ability to opt back in. The pendulum does not have to be on the side of exploration or personalization, it can be on both sides and in the middle as well. Give the users the possibility to choose.

Recent studies have shown that users have enough experience to know that data collection and personalization provide a lot of advantages for the user. But let it be their decision. Let them be the bus driver of their own online life.

Let the user be the bus driver of his own digital life.

Second: Implement a Three-Layered Approach to Exploratory Personalization

If you are working with video content you may be familiar with the concept of hygiene, hub and hero content. The summary of the model is that you must provide three types of videos in order to engage and retain the user: hygiene is content that you provide en masse. It is the basis of every content strategy. Hub content is more regular produced content that is pushed to the user and hero content is the highlight content, the icing on top of the cake.

Let’s take this approach, adapt this to personalization and you will get a new three-layered approach to exploratory personalization.

Exploratory personalization with a three-layered approach

A. Exploration (aka Hygiene)

This is the basis of your shop or website. No matter how much data you have about the user, it is not used for the means of personalization. Exploration is the area of your site with a truly unfiltered view. An example for this would be the old Instagram feed where you could see everything that was happening in a chronological order.

B. Segmentation (aka Hub)

This is the first step to personalization based on user segments. This area of the website or the shop is adjusted to the needs of your most relevant user groups. Based on behavior and other data you provide a better way to discover your content, products and tools. Examples for this are the classical first attempts on company’s personalization strategies like differentiating between men and women, new and returning visitors, visitors interested in tech vs. books vs. …

C. Personalization (aka Hero)

This is where the real magic happens. This is the view were the website is perfectly tailored to the specific needs of the single user with hyper-personal shopping advice or content recommendations.

Providing those different layers of exploratory personalization does not mean that you have to divide your website into different, distinctive categories. It can be done fluently throughout the whole user experience on your site or your shop. Just follow the first rule “declare the user of age”, show which parts of your site are powered by personalization data and why you think that this will be a benefit for the user. If you are really using the data to optimize the user and shopping experience, the users will thank you for this.

Final Thoughts

It goes without saying that this attempt is not a short-term approach and requires new ways of thinking about the user experience. But setting a guideline for the long-run will help building better products using personalized data. And I guess we can all agree on: personalization is here to stay.

How do you feel about the state of personalization and where it is heading?


bottom of page