Mark Detweiler from SAP has an exciting article – Managing UCD within Agile Projects – in the May / June issue of interactions. Last year I used SCRUM to manage a small development team on a high-profile project and one of the lessons we learned was that agile development does not work well with disciplines that are not also agile. The development is managed and completed in a series of sprints (we worked three, 30 day sprints) and the application was twitching before any formal UCD work had completed. This challenge applies to more than just User Centered Design (UCD); accelerating development is only so valuable. Detweiler’s UCD centric view offers a series of tips on making a better go at it, but the truth is this problem needs a more radical conception.
Detweiler offers three iterative UCD phases, Understanding Users, Defining Interaction and Designing the user interface. One observation is that UCD teams need to focus their energy on work that influences the current development sprint. Every profession has more they want to do. Delivering it apart form the product development positions their brilliance it to be ignored.
Understanding users is important, but it is not going to hold up development, so what is the minimal amount of work that needs to occur in order to have the development team be more considerate of the user context. Since we are iterating, maybe each sprint has a refinement of the user definition. One thing for sure, the developers will create a user persona if one is not provided.
Defining user interaction has to happen in a more collaborative manner with the development team. It is often just as easy to code the application, as it is to write the use case documentation. This does not mean use cases are a waste of time, but we need to ask what about the document is going to influence development. The time may be better spent with the UCD professional in the initial task breakdown meetings to help define interactions in detail. Having access to a UCD professional as the development occurs helps accelerate decision-making, but the goal is not to have the solution be perfect – it is to have a thoughtful solution, where lessons learned inform the next sprint’s interaction design. The UCD professional needs to own their role representing their stakeholders, but not constantly checking in to see if they are meeting expectations – there is simply no time for checking in and reporting back at every turn.
Designing the user interface as part of an agile development project is difficult. Short of starting this work prior to development, we need to find the opportunity in having the user interface be quick and dirty. An existing design system can address the developer’s need to assemble the application without having to worry about the overall application style. Assuming that a custom user interface is required, the designers need to be in the boat of iterative design. Save the Mona Lisa for a few iterations from now. If you want to influence this sprint’s user interface, worry less about the pixels and more about the styles. Design elements that telegraph where the design will be going. Development is iterative – it is assumed things will change so influence at each point, but save the masterful work for a time when it is not needed yesterday. Great work often takes time; why pretend it can be done in less?
All of this feels a little sloppy and this is why I think what we need is to revisit UCD and maybe the larger thought of user experience and how it is most affective in an agile development environment. This issue is exacerbated in innovation teams where often prototypes catch like wildfire and suddenly a solution with little to no formal UCD thought is at Version 1.0, causing pause to the value and role of UCD. We need to do more than apply more resource or slow down development to accommodate external dependencies. We need to do more than just the same thing in a slightly modified way. To be successful, we need to reinvent UCD or stop being agile.