Wednesday, January 24, 2007

Revisiting Personas

I opened this blog back in 2001, when the service had just started to become well known. At the time I had no idea what blogs were, but it was a big buzzword at the time, so I figured it was the next big thing.

Since then I've never really had a chance to come back and get my blog started. Well, it's about time. And what better topic to start with than Personas, probably my most well-known piece of work to date. Ironically, this wasn't my master's thesis, or even part of my master's program - it was a small essay I wrote during my undergraduate CS373 Software Engineering course, back in 1998! Since then, I've continued to receive feedback from new students, letting me know that Personas is still alive and kicking.

I think it's time to revisit the Personas concept, and to look at how it fits with modern software design methodologies and development lifecycles.

First, let’s recap on the original ideas. I borrowed the Personas concept from a book by Alan Cooper, The Inmates are Running the Asylum. The idea is that rather than following a generalized approach to analyzing requirements, a more targeted approach should be used. This involves defining a set of Personas, characters with specific characteristics, on which design decisions can be based.

The motivation behind this idea is that we often try to meet the needs of the "average" user, which results in "making everyone happy some of the time" - but with software that's not good enough. With the Personas approach, we are focused on "making some of the users happy all of the time", and combined with good decisions about the targeted characteristics of the Personas, this can result in making the majority of users happy with most or all of the software.

It's amazing to think that many years later, the most modern and popular software products still suffer from this problem of generalization - how many applications still seem to blindly target generalized needs of some average user, while not actually meeting the needs of anyone? One recent case that springs to mind was highlighted recently in an insightful post on the Joel on Software blog, where a usability problem was pointed out with a major feature of Windows Vista. It makes me wonder if a Personas-style approach would have eliminated this problem.

Back in 1998, RUP and UML were very new concepts, and we were just starting to learn about them and put them into practice. Today UML has lost that new-acronym scent, but the use-case technique still has a familiar ring for Personas, and the idea of creating a Persona storyboard seems to apply a more targeted technique that can improve a standard use-case approach.

Has the more recent Agile Development approach rendered Personas obsolete? The Agile approach is in direct contrast to a planning-driven methodology. But there are aspects of the Agile approach, including Extreme Programming, that have similarities to the Personas approach. The XP approach encourages the use of note cards to engage in “story telling” – clearly this can benefit from the Personas approach to target the scenario to specific user needs. As noted on the IBM website, one goal in XP is to “motivate the team to sit together in a room and talk about what the proposed system needs to do and how it needs to do it” – and this is exactly the idea with the Personas approach.

In a way it seems that newer Agile approaches are the next evolution of the old Personas idea – in fact, while that idea was almost revolutionary in 1998, today it can almost seem obvious. But new Agile approaches would benefit from looking back and the real motivations behind the Personas approach, and attempting to incorporate those ideas. While an Agile approach can drive a more user-focused and rapid delivery approach, the end result can still be wrong if based on user requirements that are too generalized. Moving from an “average user” approach to a more targeted user approach to analyzing requirements will ensure that not only are results delivered to meet some interpretation of the requirements in an “agile” manner, but also more likely that those results will meet a targeted, and more correct analysis of those requirements.

By the way, Apel Mjausson has a blog entry linking to my essay, and also to a number of other interesting papers on the topic. Also, there is a Polish site referencing my essay.

No comments: