Design for Personalization: How Companies Personalize Products

Have you ever noticed how perfectly a certain product ad appears in your Instagram feed right after you mentioned the trigger word briefly in a conversation? Or texted in a messenger, or even just thought of it… That’s the symptom of personalization.

People appreciate a personal touch. Even bad coffee in a paper cup tastes better when somebody draws your name and a heart on it. What’s more,research from McKinsey shows that businesses that are particularly successful in personalization generate 40% more revenue from such activities than their average counterparts. Most customers expect personalization and get frustrated when they don’t find one.

That’s why nowadays product design is tightly connected with personalization. Let’s see why, where, and how to apply personalization in UX design. But first, a quick definition.

What is personalization?

When the app interacts differently depending on a user profile or other variables, such as location or time, it’s a sign that personalization techniques were used.

Note that personalization is not the same as customization. The latter is related to changes in UI that users are making themselves to fit their preferences or needs, such as changing color themes. To find more, read our article about personalization vs customization.

Personalization, just like targeted ads, can be very helpful sometimes and very annoying and even nasty in certain cases. An example of the latter is flight booking services that show higher prices to users who enter the website from an iPhone. Websites with dynamic pricing can use customers’ personal data to adjust prices and earn more.

Personalization can be simple, like your name added to the welcome screen, or more sophisticated. You can use basic data such as location, or do it like Netflix ,which applies a variety of Machine Learning (ML) algorithms and statistical techniques such as Causal Models, Contextual Bandits, and Neural Networks. 

Companies with large budgets can afford all of it, but even when your resources are limited, personalization is still worth an investment.

What are the benefits of personalization?

1. It saves users’ time. When your browser places links to the six websites you visit most often on the starting page, it spares you the time of entering the website address in the search field. In terms of time, it saves from 3 to 30 seconds, depending on the user. But many things in UX design are about saving those few seconds here and there. In the end, these seconds matter to people, as I can remember any time I see someone rushing to cross the road when green light takes too long to wait.
2. It helps improve user engagement, conversion, and retention rates. Personalization allows a product to provide relevant info to the users, as well as better recommendations. When people feel that the app knows what they want, they are likely to interact more, remain loyal, and pay for it, as they feel extra value in personalization.
3. Gives competitive advantage. In the world of hundreds of similar products, good personalization is much easier and can be more efficient than an attempt to build unique functionality or design. A personalized experience makes users remember your app and choose it among the others.

Types of personalization

While targeted ads are the most recognizable kind of personalization, there is more to it, and it might not always be visible. Here are the main types of personalization in UX with examples.

Segment-based personalization

Many apps ask users about their objectives at the beginning of collaboration. That’s a typical solution for segment-based personalization. From the very first steps, the product would suggest to users the features that might be the most suitable for them.

This type of personalization has a high chance of being non-intrusive and efficient. People usually don’t feel like they are sharing too much personal information when they say that they will use the product to learn a language to travel or work. That’s the case of the Busuu language learning app: the vocabulary and exercises change depending on the learner’s objective, their knowledge level, and the time dedicated to learning.

And here is how one of our clients, TextMagic, organized a personalized marketing system: you can segment users by different categories, such as their location, dates of subscription, or last interaction with an email.

Location-based personalization

There are many obvious use cases for location-based personalization. Many e-commerce websites won’t let you make a single click until they can locate you, and would ask you to choose the country of delivery manually in case the website couldn’t detect it.

When talking about examples, Seek, an app for defining animals and plants, suggests users learn something about the animals common in their area to keep them engaged.

Time-based personalization

Time is not perceived as personal information, but it can be used for personalization as well. Meditation apps like Calm will suggest you do a sleep meditation at night and a “Daily Calm” one in the morning.

Cross-selling personalization

When you are watching a page of a product, you see suggestions of things that “would look well with this”, or “what other users bought with it”. Online sellers like YOOX have mastered cross-selling, and now it has become a kind of standard option for e-marketing platforms. The difference, as always, is in the level of sophistication. The suggestions can be prepared manually or half-automatically by a content editor, or based on complex algorithms, or even an AI-powered system.

Individualized personalization

This kind of personalization goes an extra mile from segment-based. Relying on all the available data and user behavior, the system makes conclusions on users’ characteristics, “labeling” them as aspiring photographers, in search of a job, singles, vegans, Star Wars fans, and so on.

Let’s take the famous personalization master, Netflix. Their objective is to create “a fully personalized homepage experience”. From the outside, it looks simple: for those who like dramas, they show dramas, those who like vintage thrillers, get vintage thrillers… But there is a lot more to it. 

For example, the same film can have different artwork shown to different users. If a user is a fan of Uma Thurman, for instance, they would see a Pulp Fiction artwork with her picture, while a person who has just watched several films with John Travolta, would see him instead.

So, let’s imagine we have a generic fitness app and we want to provide personalized user experience. What options do we have?

1. We can split users based on their fitness objective: those who do fitness for rehabilitation, those who want to lose weight, yoga lovers, those who want to be fit, those who love/hate running, and so on. We tag each workout for a specific segment and show users the ones that would help reach their objective.
2. Location-based personalization. Typically, fitness apps need access to users’ location to track their running and walking progress. To motivate them to stick to their running routine, we can show a beautiful picture from the park where they run often on the welcome screen.
3. Time-based. In the morning, the app would show “5 minutes stretch to wake up”, in the middle of the day “5 exercises to do in the office during a lunch break”, and so on. But it’s not the best idea to remind users to go jogging at 7 PM in winter if the app knows they don’t like jogging in darkness.
4. Cross-selling. If app owners want to make income by selling sports goods, they must use personalization to show users relevant goods: yoga mats under yoga workout, rain vest next to the graph of running progress (ideally, on a rainy day — remember about location!).

These are some of the easiest cases of personalization in UX. Examples are based on the 100% available data. If users granted us access to other, more private information, the personalization could go much further, but we don’t insist.

We don’t have any fitness apps in our portfolio yet, but we are open to requests, so when you’re looking to design a personalized fitness app – you know where to find us ;)

How to create design for personalization

Rule number one is to base your decisions on data from user research. Be it complex ML algorithms or just a basic element of personalization, you always have to bear in mind that personalization has to work for all users and not just for you.

The basic research for personalization would start with in-depth user interviews. It helps to learn what level of personalization your target audience would appreciate and what would be too much for them.

To verify and adjust design solutions, A/B testing works best. It helps designers to avoid bias and check the results with a wide array of users.

Personalization design must be checked and revised from time to time. That's why  you have to check on your users through a survey or some interviews once in a while. In modern products, constant changes and adjustments are key to success.

Ethical personalization

There is a number of challenges that designers face when working on personalization in product design. First and foremost is the previously mentioned privacy concern. To prevent people from those unpleasant surprises, design must rely on principles of transparency and reciprocity. To make it easier, we have formulated a few rules that help you make personalization ethical.

1. Let users know where you take their private info from and how you plan to use it. Many people are familiar with that scary feeling when in-app ads are so well targeted that they seem to listen not only to all that you are saying but even what you are thinking of. To make targeting more ethical, YouTube lets people know why they see certain ads (at least to those who can find the tiny button in the down left corner).

2. Don’t oblige people to provide all the permissions at once. Often, the first thing you see when opening a new app is a pop-up asking you to provide access to literally everything. Even if its sole function is to light a candle at a concert. It makes users hesitant to approve and is likely to cause annoyance. The smart one is to ask for permissions and access when it is essential to complete a certain task and go one by one, not bundling them all together.
3. To make people feel better about sharing their private info, provide some value first. Often, apps ask users to give access to the data from their phones the moment they open it for the first time. Naturally, many people feel uncomfortable about giving so much info, especially when they don’t see clearly how it will be used.
4. Let them opt out easily. In many cases, finding the privacy page in settings is a real challenge. Try to make it easy to find, as hiding the information doesn’t help build trust with users.

Smart design would ask for that data later on the user journey, when they have already got some value from the product and thus are more willing to give something in return. Reciprocity is not just an idyllic idea: it is a scientifically proven psychological principle that is used in UI a lot.

A case study on Brave browser from growth.design shows how to place the pop-up strategically: when the user sees the value of the product.

These pieces of advice will work only if the inner architecture of a product is ethical and it doesn’t use private data for unauthorized motifs.

Avoid overpersonalization

Most solutions on personalization politics are the responsibility of product managers and developers. Still, there is one thing we can say for sure as UX professionals: too much personalization hurts. It can create mistrust and annoyance, finally resulting in an overall negative user experience. To find out where the thin line between just enough and too much is, read our article about overpersonalization.

Is personalization the future of UX?

For sure, the answer is YES. Though functionality has higher priority, personalization is what lets products gain a competitive advantage once they get to the product-market fit. With the fast advance of AI and ML technologies, personalization becomes more sophisticated, precise, and available. To get an idea of how AI can benefit your product, read our article on AI-Powered Functionality for SaaS.