User Personas Can Be Dangerous

User personas have let us down. Personas represent users, which isn’t inherently bad, but it isn’t great either. Representation, which is what personas are, can be dangerous.

Representation waters things down and generalizes perspectives into more manageable groups for us to think about, strategize around, and act on. Representation sounds good on the surface, but too much representation versus engagement and participation from constituents can get us off track.

I have written previously about how the voice of customer initiatives are false attempts at actually being close to customers. User personas run the same risk. User personas can fake us into believing we understand and resonate with users. We have generalized the users into identities detached and disconnected from who they are and what they value.

Representation tends to lead to extrapolation and often premature extrapolation at that. We are so eager to get validation and alignment with what we are working on that we will start representing what we want out of too few examples and data points. The extrapolation means we start filling the blanks for users in personas because we need the pieces to fit. Personas can become less about users and more about who we believe who the users are, what they need, and what they value.

Personas are now too often becoming representations of what we believe about users versus what we know about users. Personas based on generalities that product creators forge are no more useful than features a product team has come up with in a vacuum. Product teams, especially with new products, are much better off creating a product with a small group of invested users and forgoing time and energy on personas that might loosely represent the users. Personas can provide more value as a product evolves as distinct groups of users can be identified and captured, but this is dangerous. As soon as we cross the threshold of representing users, we automatically begin to drift away from them and generalize.

Creating for someone is always better than creating for a sizeable nebulous group. Still, we’ve convinced ourselves that we need to represent the larger group even initially because we risk creating something too narrow for a small group of people. Product market fit, total addressable market, and scaling are all guilty of driving out the perspective of creating for an engaged few and then going broader and deeper. Personas help to pervade this thinking and action by representing instead of engaging. Personas are every product team’s voice of customer initiative.

Personas didn’t start badly. Product teams needed to capture what they were learning about early, engaged users’ personas fit the bill. Personas started as more retrospective than prospective. They captured the attributes of users after observation, data, and conversations. Personas migrated to being prospective because of the rush to build quickly. The need for speed and push to design and development flipped the value of personas from clearly positive to questionably negative. Unfortunately, as with many aspects of a process or methodology, what starts pure gets bastardized and misused over time, and personas have fallen victim to this.

Personas can still be valuable and useful, but when leveraged with too much representation and not enough practicality, they no longer serve the users as much as they do product creators. And when the pendulum around personas swings too far to the side of product creators, getting placeholders in place around users, personas are doing a disservice to product creators even though it doesn’t seem like it. Personas provide the most value when used in the best interest of users, not in the best interest of product teams. Product teams must keep in mind that personas are not for their benefit as much as they are for the user’s benefit.

The trap too many product teams fall into now is believing that is continuously and iteratively engaging with users is can be replaced by personas. Engaging directly with users is irreplaceable.