The backbone to building a well-loved product is knowing your end user. Many are familiar with user persona descriptions, well-adopted by marketing teams but often are not as well received by product teams. The biggest criticism is they can feel too superficial - focusing on high-level goals, motivations and demographics, complete with a backstory and profile photo but less rooted in product behavior.

The original purpose of personas was to help guide decisions and to make sure teams were taking the end customer into account. They have been taken under the wing of marketing teams, focusing on market segments to help with landing on the right messaging and marketing initiatives which resonate with the end customer. The problem is different teams have different drivers behind the use of personas. For the marketing team, the personas will be optimized for commercial purposes, whereas the teams working on the product will have other factors to account for around user experience. This makes traditional marketing personas less useful for product, design and technical teams, as they are somewhat removed from in-product behavior and interaction data. 

This shift away from product behavior to broader profile segments has also led to a loss of nuance. Users in a segment have shared characteristics but they are not identical. A single market segment will be made up of many user profiles, each with their own behavior characteristics - some overlapping, some not. Often differences which might seem small or not significant enough to warrant a user being assigned to a different segment can cause divergent behavior down the line in response to a product change. Equally, there may be users assigned to different segments who respond in the same way to a change. At the end of the day, there’s no such thing as ‘an average user’.

Also, nowadays, speaking directly with users has become more widely adopted, and teams are constantly learning and gaining new insights. Having set personas can feel too final and resistant to change - resulting in decisions that no longer align with the initial userbase snapshot as soon as there’s a shift in behavior or attitudes or a new user profile emerges. This is particularly restrictive in a time where the demand for more responsive and assistive products only continues to rise with the adoption of AI co-pilots. With this comes an expectation of more intelligent products with a deeper understanding of situational context and behavior drivers.

So, what does the future look like for user personas?


Personas have the potential to be much more powerful than we give them credit for. Instead of solely relying on top-down company decisions where the default is the “highest paid person’s opinion”, in recent years, there has been a shift towards product-led decisions from product-led growth to product-led sales. This change has put even greater importance on customer and user behaviors. The closer we can get to an individual level, the more helpful user context will be across the organization. Some examples, include;

  • Customer service: An in-product customer service bot can better respond to questions knowing a user’s historic behavior as well as the current session behaviors which led up to the user’s request. 
  • Sales: This behavior context can also be leveraged by sales to better personalize offers - generating thousands of micro-variations of sales messages rather than cycling through a handful of pre-set notifications.
  • Product: With richer behavioral profiles trained not just on a few funnel events but on your entire event stream and every micro-interaction within your product, suddenly you have a much stronger springboard for anticipating and predicting user needs when it comes to making product changes.


For this to happen, it’s important to get as close to real-time as possible, this is where it’s helpful to make sure user personas are living and evolve over time along with data drift. Of course, we want to learn from historic behaviors but that also means we shouldn’t ignore deviations and contradictions which break the pattern and could be a sign of a new behavior pattern emerging.

Often teams will set up rules-based workflows to handle some of the complexity and scale around decision-making such as, user segmentation logic, triggers for messaging and “if-then” logic but it’s key that user personas and the consequent business decisions reflect more nuanced behaviors. This is where there are benefits to model-driven personas over static written profiles. In the flurry of taps and clicks, it’s easy to find yourself drowning in data signals, trying to decipher what’s worth listening to. Data points on their own can only go so far but in combination, they start to reveal distinct behaviors. As products become more sophisticated and data input continues to become more multi-modal, making sense of complex data patterns is going to get harder particularly when trying to detect latent behaviors. Machine learning helps uncover the more nuanced behaviors in product usage data patterns. Combine this with behavioral science, and teams are able to completely reinvent the way they think about user personas. Beyond detecting hidden insights, a model can also handle the scale of data looking at every possible combination of behavior without getting bored and at a much higher processing speed than a human.

At Blok, we use model-driven user personas, designed to be dynamic and offer richer context for making faster product decisions optimized for the individual end user.

‍

Photo by Ben Sweet on Unsplash