BACK TO BLOG

Nov 25, 2009

An Agile Moment

Posted by: SethG

Recently, we had a great agile development moment while revisiting our PayPal implementation for Giving Impact. Sometimes a feature and/or approach simply doesn’t work out. We tried something that we hoped would be an improvement; however there was one aspect of the solution that proved to be so disagreeable that we decided to shelve that approach. At first I caught myself feeling like that disagreeable detail should have been identified earlier; however after a deep breath, I realized that this was a just another reaffirming agile moment.

If I were to have taken the time to think through every detail of the solution we explored, I likely would have spent the same amount of time and money as I had moving to development rapidly and would have ended up with the same result of shelving it. In hindsight, I actually think there was far more value in actually having the experience of using a working solution and determining with a degree of certainty that it was not a workable solution, rather than coming to that conclusion only via discussion and planning.

When building applications, I have found that through an iterative development process we end up with great results more efficiently and rapidly. For each iteration we begin with some brief planning meetings with design and dev to discuss a particular feature(s). This planning phase may result with any combination of design, dev, and product management being tasked with work. Occasionally both design and development feel more planning and clarity is needed, so then we devote some additional planning time which may result in a wireframe or some form of guiding document. In other words the upfront planning and amount of time needed for a given feature is determined through team collaboration with the goal to only do as much as is needed to get designers designing and developers developing. At the end of the iteration there is working code that then is evaluated. This cycle continues until the project reaches a point of completion where it is deemed ready to go live.

Every once in a while, I have experienced an issue like the one that inspired this post, where the working code reveals a show stopper of a problem. In this case, after some additional discussion and research, we could not identify a clear next step to address our issue. We made the call that it was not worth additional effort at this time since we could not readily identify a path to address the issue.

Could I have spent more time talking and researching and maybe found the right question that would have revealed our show stopper before development, maybe. Did it pain me for a few minutes that I had not thought of it? Yes. Was I better off trusting the process and having a working solution to try that revealed the issue definitively? I believe so. The result has provided us additional confidence, certainty, and experience that only strengthens our application.

Tweet

MOD-Lab