SWOGC: How We Built a Web Portal in under 48 Hours
October 28, 2012 2 Comments
This past weekend I volunteered my time to the Southwest Ohio Give Camp for the 3rd consecutive year. All in all, it’s a great event, a local chapter of GiveCamp.org, where regional software developers, designers, and technologists gather together in teams to create working solutions for charities. In the past, I’ve merely volunteered my time and expertise, but this year I undertook a higher challenge, leading my own team. In this post, I discuss my plan for building an entire web portal in less than 48 hours and the lessons learned.
Leading a GiveCamp Team – The Plan
I had compiled 8 tips for myself through observation in past years. Going into the weekend, these were my keys to success:
Before the Weekend
- Prepare with the charity representative to obtain the majority of requirements
- Present the charity with a “Wish List”, what we need the charity to provide (e.g. logins or files) to be successful
- Begin visual design collaboration and discussing philosophies
- Decide on a CMS, such as WordPress or Umbraco
During the Weekend
- Be pragmatic and cognizant of the critical path!
- As leader, make sure there are no tasks during the weekend that only I can do
- Stay clean and flexible
- Create a project plan first
Putting it all together – The Execution
Knowing I would be leading my own team, I recruited 2 .NET developers to help over the weekend. In addition, the event organizers assigned me 2 more developers and a designer. Our charity’s representative was also there to help for almost the entire time.
From Friday evening to Sunday afternoon, our team of technologists contributed about 120 hours to the project. We implemented an internal employee portal for the charity made to match the style of the existing website as much as possible. I would say it was generally a success.
Lessons Learned – a Retrospective
“Is that it? That’s all you’re going to say about how the weekend went?”
Well, the details of the project weren’t that interesting. But now that you know what happened, I can remark on what I found to be most important about our strategy and implementation.
Getting Started was Difficult
As you might imagine, assembling a team of mostly strangers, who know very little about the task at hand, can present a specific set of problems. Namely, everyone struggles to dive into a task because of a lack of knowledge about the overall requirements, plan, and others’ skill sets while pitted against the aggressive time constraint.
I think this is somewhat unavoidable. I did a good job of following my own advice, numbers 4 and 8 from above. I also had a list of known tasks pre-written on post-it notes that I used as discussion points to explain the scope of the project. I think having had some of the bigger decisions made going into the weekend helped the team to focus on delivering value to the charity. However, the team didn’t seem to naturally mesh and become fully engaged until Saturday afternoon.
How could I have changed my approach to reach that point earlier? I could have gone further with my research for the weekend making sure that any technical roadblocks were overcome before we started. This became most obvious in relation to my choice of CMS. We were using Joomla!, the CMS that the charity already had for their existing website. Personally, I did not know enough about this CMS to get started in a reasonable timeframe. I was dependent on team members to get our site up and running and the members who did not know Joomla! did not have much to do. Looking back, I would either choose a CMS that I was already more comfortable with or make sure that I knew how to get started with what I chose. Alas, I realized I got lucky with my team member assignments by acquiring some Joomla! knowledge along with them.
Iteration is Awesome
Having the luxury of a charity representative throughout the weekend, we leveraged a tight feedback loop to ensure we delivered a valuable solution. In fact, around noon on Saturday, we all realized that a significant portion of the planned tasks would only be of marginal benefit. Because we adhered to tip number 7 above, we were able to have an ad hoc discussion with the charity about what would provide the most value. As a result, we scrapped many of the broad functionality on our project plan and replaced it with one deeply functional requirement. Had we not had an iterative approach we would not have been able to fulfill the charity’s actual needs.
Get a Designer
Every year, the most in-demand skill at these GiveCamps has been design. There are usually not enough designers for every team and this year was no different. We were lucky to get assigned one (who also was fairly knowledgeable in Joomla!). Having a designer on the team is important because there is no easier way to feel like your solution is worthless than to have it look subpar. This is especially true at a GiveCamp weekend, when all the teams present their results on Sunday and give you benchmarks to compare against. On a time constraint and non-existent budget, I don’t know many developers who could make a user-interface that meets the new standards of the web. Even though our charity was not looking for a “fresh” design on their internal portal, it was nice to have someone who could troubleshoot display issues as we put it together.
Would I Do it again?
In a word, absolutely. While I’ve found that the solutions created for these charities over the past 3 years have not been the most interesting technically, the delight and value we are able to provide for non-profit organizations is quite fulfilling. Sure, the weekend comes with a certain level of anxiety. At one point on Saturday, the Web Host for the charity’s site (and all our work in progress) went down temporarily rendering our team useless. Still, it is a great way to meet other local developers who do not necessarily work on the same platform in their day jobs. Leading a team of volunteers who just want to contribute their skills in the most effective way possible provides exactly the right amount of eustress, and I’d recommend it to anyone.