In Defense of Personas

defenseAlan Klement has been on a recent war against personas. I enjoy his posts, but after reading his most recent, felt the need to step up in defense of personas.

Alan’s primary beef with personas is that “they are made up and not real”. Because of this, he claims that they will ultimately cause teams to make erroneous decisions in the design in development of features. Instead of using user stories with personas, Alan proposes writing “job stories” that omit the persona and focus purely on situation, motivation, and expected outcome. I have no beef with the elements of his job stories (not sure I see the need for a new name), but I do take issue with omitting personas – and I’ll tell you why….

Persona’s provide valuable context. Without this context, we are left in the dark as to important aspects of the user such as their technical acumen, domain knowledge, frequency of system use, etc. Each of these drives very important concerns when designing for a new feature. Consider the following example:

My user wants to input a new order into our system. If I were to write one of Alan’s Job Stories for this, I might say: “when I need an order placed, I want to enter that order into the system so that it is promptly and accurately fulfilled.” No persona, but I know what needs to be done, why, and the desired outcome. Alan would have us believe this is a good story. It’s not. In fact, it sucks. It sucks because we know absolutely nothing about our user. Would you implement this feature differently if the user was the end customer vs. our employee in a call center? Of course you would. What if it’s an employee who handles orders for us and 50 other companies, so is not all that familiar with our products? Again, hell yes, it gets implemented much differently.

I think the problem here may be that Alan has never seen a good persona. Personas are not made up. A persona is, by definition, an archetype user. This means that we do real research with real customers to create personas from shared relevant attributes of these real customers. We don’t make up attributes and we don’t generalize. The only “made up” aspects of personas are the little bits of color we give them to help them to develop a personality, e.g. name, favorite band, kind of pet, car they drive – but it’s critical that these fictionalized attributes of our personas not be relevant to our product set. E.g. we don’t make up their favorite band if we are on the Spotify team. The relevant attributes then are quite real – and extremely important as we are saying they are shared across a significant number of our user base. These then serve to inform our designs. So going back to my example, if I create a persona John like this:

John has worked in our Dallas call center for 12 years. He can tell you the part number of nearly every product we’ve sold in that time as well as the available sizes and colors. He works five days a week, eight hours a day takes approximately 800 orders in any given work week. During peak hours (9A-12P) he is 98% utilized, meaning of these 180 minutes, he is on the phone with a customer for 176. Calls with customers can range from 1-15 minutes and John is motivated to keep calls short during peak hours….

I won’t go on, but are you starting to think about some of the requirements implied by this persona? How much do we miss out on if we don’t know these things? If I add to this persona that John drives a 1976 Gremlin, listens to Black Sabbath, and frequently participates in Civil War battle reenactments… this is kitsch – everyone knows not to take it into account when developing our order entry system.

Don’t give up on Personas. If you aren’t finding them useful, it’s likely you aren’t doing them correctly. Spend some time getting them right and they will serve you well.

Share

If you enjoyed this post, please consider to leave a comment or subscribe to the feed and get future articles delivered to your feed reader.

Comments

Enjoyed your article. A few things.

“It sucks because we know absolutely nothing about our user.”

You don’t need to ‘know’ the user. You only need to know the situation they are in and what they are trying to get accomplished.

“Would you implement this feature differently if the user was the end customer vs. our employee in a call center?…What if it’s an employee who handles orders for us and 50 other companies, so is not all that familiar with our products? Again, hell yes, it gets implemented much differently.”

Those are different situations. So you would need to write different stories for each of them. I would suggest grouping similar POVs & situations together to define features.

“This means that we do real research with real customers to create personas from shared relevant attributes of these real customers.”

The problem with Personas still exists. Describing a persona as an ‘archetype user’ doesn’t take away the fact that they are not real – it’s just a mash up of attributes. Consider this:

A billionaire and a construction worker will have completely different personas; however, they both stopped the corner pizza joint to get a slice. It doesn’t mater if the construction worker was walking by after working 10 hours or if the billionaire had her chauffeur pull over, both were in the same situation:

They were in a rush, very hungry, like the taste of pizza, were passing by and wanted something to eat ‘on the go’.

Any other information about them is irrelevant.

Ryan Singer from 37 Signals talks nicely about the problems with persona’s in this interview:

http://insideintercom.io/an-interview-with-ryan-singer/

Thanks for the comment – appreciate the honest discourse.

“You don’t need to ‘know’ the user. You only need to know the situation they are in and what they are trying to get accomplished.”

This is where you and I disagree. Good systems are not built just for specific situations, but for specific users in specific situations.

Your billionaire / construction worker example I think identifies your lack of understanding regarding personas. If, in fact, there was no tangible difference in the needs, expectations, motivations, and capabilities of these two folks when it came to pizza consumption, then they would not represent two different personas – only one. That said, I would argue that there are likely significant differences between these folks, e.g. the construction worker is perhaps more price sensitive than the billionaire and the billionaire is possibly more health conscious… These details are not in the situation, but in the persona – and they definitely impact the pizza shop’s offering.

Leave a comment

(required)

(required)