Over the last few days, I’ve taken part in (and facilitated parts of) an intense workshop meant to define the user experience of a new product. In the room we had representatives from pretty much the entire team — software engineers, hardware engineers, industrial designers, interaction designers, marketing, brand, and even the CEO.
At the end of the first day, we found ourselves a little unmoored — even though we had talked about our presumed users (this project is to launch a brand new product into the market, so there are no existing users), the discussion was nebulous. We needed an anchor.
So on the morning of the second day we dove into a discussion of personas — those archetypal users of the product. We had three personas (Casey, Jessica, and Eric), and we talked about (and occasionally argued about) them for quite a while, until we arrived at a shared understanding of who they are, and what they would get out of the product.
This discussion proved enormously valuable — it lead to some coherence around who the product was for, and it helped focus our discussion of desired experiences, and, in turn, functional requirements. We referred to these personas for the remainder of the workshop, and they came in handy for resolving conversations that got stuck in “Well, I think…”
So, I was a bit surprised last night to read Steve Portigal’s article in the latest Interactions magazine, “Persona Non Grata.” (The first element of surprise was how it had the exact same title as Dan’s Adaptive Path essay from over two years ago). What most surprised me was the vitriol Steve cast at the practice of persona development — he essentially derides it as a waste of time, an exercise that purports to build empathy but in reality distances us from our users.
The thing is, when you read the article, it becomes clear that Steve is talking not about personas, but poorly conceived personas. Like any tool, personas can be wielded effectively or not. Steve is right in that the bulk of personas out there don’t serve their desired purpose, because they get too cutesy with alliterative names, or label people as types as opposed to individuals, or become cliched. But that’s not the fault of personas, that’s the fault of bad personas.
In our practice, we haven’t seen a tool for building empathy as effective as a well-constructed persona. We’ve used it numerous times to great success.
My frustration with the article is two-fold. First, because so many personas are bad doesn’t mean that we should throw out the practice. That’s like saying we should stop making movies because most movies suck. Steve commits Jakob’s Fallacy, perhaps most famous in his “Flash 99% bad“, wherein you dismiss a whole category of activity because the bulk of its practice is poor.
Second, no constructive alternative is presented. Steve could have taken two paths — either delineate what it takes to create a truly productive persona, or present other tools that successfully accomplish the objectives that personas fail to meet. However, he does neither, so at the end of the article, you’re simply left wondering, “Well, if personas suck, how do I make sense of my user research? How do I build empathy across a product team?”
Amen, brother. (It’s probably not very surprising that I’m right there with you.)
It’s not the tool, but what you do with it.
I believe the conversation you want to have went like this:
“90% of science fiction is crap.”
“Yes, 90% of everything is crap.”
Reminds me of 37signals’ post on the topic of last year (Something that we – Etre – responded to here).
Personas for the most part are bad because they are synthetically manufactured composites that are not real people. What kind of “empathy” are you building across product teams with “replicants”?
What you do with your user data is your analyze it, you look for underlying meaning and you share that with your product teams.
I think that one of the points Steve is trying to get across is that the temptation to synthesize together a set of behaviours into a single “meta-person” causes a disconnect and this unreality allows clients to pretend to connect with “real” consumers whilst actually ignoring them.
For us a common problem is that many clients have already created things that look very similar to “personas”. But they call them “target market segments” or something similar. Ad agencies and clients have been doing this for years if not decades – they were often characterised by alliterative names and pseudo-real moodboards and so on. Clients are understandably confused when we wander in and suggest creating personas.
One way we are combating this is by actually using real people as our personas, and essentially documenting a real person, using their name (or very similar) and their actual habits and likes, rather than shaping them in any way. This a purist approach, I admit, but it also pretty powerful.
At the other extreme we have recently had clients ask us to create personas based on their existing segmentations. This is a level of abstraction too far…
Hi,I enjoyed your response in response to that article. It is the fault of some bad personas that they in general have a bad rap going around. I believe that you have to ground them first in some actual data (so you aren’t disconnected) and then build in things that help shape them. I love how you talk about empathy as that is a particular focus for me right now in studying theater education for the purpose of using the techniques to drive organizational development and process. In particular focusing on how to lend that more empathetic layer to personas and scenarios by instilling exercises centered around character development and building narrative into the traditional processes. I strongly believe when built and used effectively personas can serve a great purpose in design and communication.
“…wherein you dismiss a whole category of activity because the bulk of its practice is poor.” is especially fine. Gets right to it.
I don’t need to spend a lot of time defending personas, so your article helps me find the words and then get back to training and mentoring my staff.