For the past year I’ve been working on projects in a traditional waterfall model. I’ve come to realize that this favors the design team since we have lots of time to research, conceptualize and design. This model puts a lot of pressure on technology to make assumptions based on documentation. The reality is that once the coding begins is when our back end team gets a real sense of how large a beast they are dealing with. Each project presents new challenges and getting into the sandbox as early as possible helps technology make better estimates.

Fair enough, so now we are trying to work agile into our projects but it seems we’ve went from one extreme to the other. Where agile puts little stock in up front research it puts pressure on design to create an overarching vision for our products without properly understanding the domain or the users.

We had general training on agile and the principles I got from it I loved:

• less documentation
• working code
• people over processes
• a lot of collaboration

Although I left the training with some less than desirable feelings. My gut was telling me that without the upfront research we only have a partial plan of attack and as things change in future iterations they affect all of the previous iterations. As the evolution of interaction evolves by iteration 5 it means we need to refine iteration 4,3,2,1.

Anyway I’m still open to agile but I think a hybrid approach makes the most sense.

• A good amount of upfront research which will provide a solid foundation and a plan for the site.
• The equivalent of stubs for wireframes with enough detail or a guess at the interaction
• Usability testing of the foundation and then iterative testing as more user stories are incorporated.

In the end we need to service these three areas:

With waterfall we focus on users and business, with agile we focus on technology and business. Fingers crossed in my hybrid world we get it right in the middle.