"Mark Plant: Experiments and experience from the UX/Agile divide" at London IA

On Tuesday night we held the first London IA event of the year. Sense Worldwide's Senseloft in Soho has become our regular venue, and sponsors Zebra People were joined by new sponsors Rosenfeld Media, who provided some UX books for us to give away in a prize draw. As ever, Matthew Solle did all the hard work of organising it.

One of the evening's talks was from Mark Plant, who was looking at how user experience design fitted into the agile software development process.

His opening point was made with a couple of graphs. The first showed the classic waterfall project outline of wireframes -> design -> build. Or, as he said it sometimes is in marketing-driven projects: design -> pointlessly reverse engineer the wireframes from a design that has already been 'signed off' with the client -> build.

The second graph showed how many people make the shift to agile - which was effectively still wireframes -> design -> build, but just now called 'sprints', with the word design replaced by 'skin', and phases of shorter duration. All the old problems of waterfall projects were there.

Mark works at Lab49, and they've been trying to make the process truly agile, and so now work on a 'co-design' sprint. Here the user experience architect and the visual designer work together. They've scrapped wireframes, and instead start with collaborative sketches, which are transformed into visual design. They are also trying to take a pair programming approach, where the designer and the architect share one computer, and take it in turns to work on the design.

One problem with the agile approach is that you can end up with a product broken into lots of little stories, each of which have small designs, but you have no coherent vision. Mark explained that they try to avoid this pitfall by having a 'concept' stage up front, which can last from a day to a month, depending on the size of the project. This involves the technical team as well as the user experience team.

Mark also spent some time looking at the concept of "good enough". He gave an example of a hideously complicated 'super combo menu box' that had once been designed from a brainstorm on one of his projects.

To build it would have taken a lot of time. Instead, they looked at it in a product roadmap way, and broke the features down into their constituent parts. Thus they found that if they started with a simple scroll-box, they had something which wasn't the finished design, but which was functional enough to use.

Of course, if you constantly simplify and reduce features, you have to guard against removing so much that you no longer have a compelling product. Although, as I tend argue, what can be less compelling for your end users than shipping nothing at all?

You can view a version of the presentation that Mark gave at UX People at the end of last year: Coping strategies, UX in an Agile world

Next...

I'll be blogging my notes from the other talk on Tuesday night, where Jason Mesut was looking at what user experience designers could learn from the world of music software and hardware design.