Re: [agile-usability] Card Sorting in agile development
On 02/18/2010 12:02 AM, Osama Sohaib wrote:
Ok, we cant use card sorting instead of stories because it delivers value. But can we use it in release planningwith involvement of actual user so that it increase the chance for successful usable software. of course it cant replaced stories..
Sorry I wasn't clear, Osama.
I'm entirely in favor of using card sorting, or anything that will help make better, more usable software. That's really important work. I wouldn't list it as a story on its own, but instead use upcoming stories to drive it. Generally, there are four ways I see UX research techniques fitting in:
- Before we have decided that we'll do a story. Here, we try to figure out enough about the story to know how valuable it is, and what other stories should go with it.
- When we have decided to do a story. After we've put the story in the plan and before we actualy do it, we need to flesh out our understanding of the story enough that we can estimate it accurately and deliver it reliably.
- After a story is complete, but before it is fully released.
Some places release gradually (e.g., starting with 1% of the
population); others release in batches. Either way, there are
opportunities to test completed stories with real users before wide
- After release. Agile processes are iterative, and that works best when you can tell how well released features are working for users.
Often, I'll see a number of these kinds of research happening at once. Just yesterday I saw a product manager bringing in test users. As long as he had them there, he had them try out some things in the existing product, used them to check out some things coming up soon, and interviewed them broadly to get an idea of what unserved needs he should be looking at for longer-term product planning.
So yes, you can definitely use card sorting to help with release planning. Or afterward, to help get more detail on stories that have been scheduled for release. The only time I wouldn't do it is when working on the exact story that needs it. I recommend 1-week iterations and stories that are in the 0.5-3 day range, and I imagine it would be hard to squeeze the card sorting plus the interface changes into a couple of days.