November 6th, 2007Can we only build software for ourselves?
37 Signals has a post today where they poo-poo personas
We don’t use personas. We use ourselves. I believe personas lead to a false sense of understanding at the deepest, most critical levels.
Every product we build is a product we build for ourselves to solve our own problems. We recognize our problems aren’t unique. In fact, our problems are probably a lot like your problems. So we bundle up the solutions to our problems in the form of web-based software and offer them for sale.
So I buy in that you can/should build software that scratches an itch. I also can understand why it is powerful when it’s your itch but are they really saying we can only build software for ourselves? Note to self: good thing I consume a lot of RSS.
Seriously though, a lot of useful software would never have been built if this were the case. However, I suppose you could argue that there’s a lot of software built that shouldn’t have been.
I wonder what my former colleague Robert Barlow-Busch thinks. Bobby, what say ye?
November 6th, 2007 at 9:05 pm
If you haven’t “lived” the problem to some extent then you will probably not be able to solve it well. Of course you can do that by working closely and intensely with people that have the problem, but it will be a lot simpler if you share the problem.
November 6th, 2007 at 9:12 pm
Oops! Forgot to mention that I’ve always thought that personas were crap. Give me one real person with the problem rather than five theoretical composites anyday. Ca n anyone name a single great software tool that was designed around personas rather than people?
November 6th, 2007 at 10:23 pm
You obviously can’t develop just for yourself, since in many cases target users aren’t like us. Don’t all applications ship with options that no one on the development team ever uses? But, if we generalize the comment to mean that you should get feedback from real users in addition to ones you’ve made up, then I don’t think anyone will dispute that either. You do both. I’m not sure it’s much of a debate.
November 7th, 2007 at 1:35 am
The 37Signals post basically said if it’s not for themselves, they’re not going to build it.
Personas can be effective to anthropomorphise your potential users and associate them with scenarios or stories that allow for variances and edge cases of your feature set. It also helps with prioritizing.
To say you can only develop features for yourself is ludicrously dogmatic.
November 7th, 2007 at 1:35 am
PS Gary, glad someone reads my blog
November 7th, 2007 at 6:26 pm
Yeah, I was a bit surprised at what the 37signals gang had to say here. Clearly, it’s just not realistic for developers to always build for themselves (who would build radiology software, for instance?), and I disagree with their criticisms of personas, captured in their comment that “personas don’t talk back.” They bloody well DO talk back!
Personas are *characters*. We are hard-wired to make meaning from narrative, and character is a powerful narrative construct. Ever read a book? Watched a movie? Did you feel engaged? React emotionally to the characters? Good characters leap off the page or the screen and become real. Same thing with personas. You can project them into new situations and they WILL talk to you.
November 7th, 2007 at 11:25 pm
By “yourself” I mean your organization. I am not assuming that a particular developer needs to have experienced the problem. That would be silly. Someone in the company must have a fairly intimate knowledge of the problem in order to solve it well.
I’ve solved numerous diverse problems in my career, but first I worked intensely with someone who lived the problem.
Also, you can certainly theorize about a few edge cases, but I would first solve the problem for the bulk of the users. That way you can start making sales because early adopters won’t care about edge cases. And frankly, it might not even be worth solving the edge case problems.
November 7th, 2007 at 11:26 pm
Also, I’m somewhat concerned that I don’t count as someone…
January 9th, 2008 at 11:54 am
Hi Jeff
I think I agree with the “real people” thesis. The best software I have delivered has been when I have genuinely understood what people are trying to do with it. Which is very different to what they say they need to do with it.
To me this is why so much enteprise software is so bad. Enterprise Software is not written by people who want to use to software. A lot of information gets lost in translation between the real users and the people doing the building.