Here are some of my thoughts on Startup Weekend and in particular, Atlanta’s Startup Weekend. I suggest everyone read Micah’s accounts of ASW as well. He wasn’t there when we were finally able to turn it around Sat. night, but his account was pretty accurate up to that point.
So here goes:
- StartupWeekend is not your traditional startup. Most Startups usually start with an idea before building the team, which means the team would have already bought into the idea. Traditional Startups also have a core group from which the idea originates from. In SW, there is no initial idea to begin with, and everyone has an input on what the ultimate idea should be. It’s now a week later and I’m still surprise that we actually chose an idea. It’s very hard to run a company as a democracy. We probably should have elected a person in each of the subgroups to help come to a final decision.
- Unless your normal job involves delivering a product from conception to production in 2 1/2 days, don’t expect any ‘This is the way I’ve always done it’ to fly. It didn’t seem like many of us were working in a mindset of having to finish something in 2 days. I admit I was guilty of that as well.
- The whole team was split up into separate sub-groups (dev, marketing, biz-dev, etc.), and I didn’t get an opportunity to work with them as much as I would have liked to. There were a lot of interesting, heated discussions going on everywhere, but there was no realistic scenario where I can participate in them and get the development that needed to be done finished by the Sunday. There really should be a ‘MockupWeekend’, where people get together just to find, build, and mock up the idea in a weekend.
- From a development standpoint, we got ourselves into trouble by often classifying things as ‘easy’. Our gauge of what ‘easy’ is is normally in the context of no less than a week. In a weekend, nothing is really ‘easy’. So what problems did this cause? For one, we were constantly pushing the easy things til the end, which is smart in the traditional development sense, but when you have an overabundance of developers, it makes sense to identify those easy tasks and get people started working on them. Calling things out as easy too often also opened up the flood gates to additional features that weren’t necessarily vital to the launch of the product.
- ASW would have benefitted if Andrew brought more structure into group, but I think ultimately Startup Weekend would suffer if was ran that way. This was Atlanta’s opportunity to show what we could do, and from a self-organizing standpoint, we did pretty poorly. Ultimately though, we did learned how to work together, and no doubt if we did this again we would be much better at it. If there was one thing we did do well, and that was when it came down to crunchtime, we rocked it.
Keys to Success
Like Micah said, at some point on Saturday night, the momentum started to turn in our favor. As always, there were probably many things working together that helped turn the ship, but I felt there were a few things that were key. This is from development side — I had to shut myself from the other groups otherwise I couldn’t have focused on doing actual development.
- Jason and Alan was finally able to get on top of the situation and make themselves gate keeper of what development was supposed to be building. Before then, we were getting conflicting requirements left and right.
- Jeff, Amro, and Andrew from Appcelerator jumped in a took the task of the developing the widget using appcelerator, while the rest of us worked on the website with traditional Rails. I was one of those that was originally wary of using Appcelerator, but in the end it looked like appcelerator was a very good fit — at least on the widget side. More importantly, it put together a team that was already used to with working with one another hammering away at probably our most complex task.
- There were a lot of devs that really came through and filled in the gaps where help was needed. I can’t stress how important this was, to have people filling in doing sysadmin, CSS/layouts, and other small but definitely-not-insignificant details that really helped. I see these things as the 20% that always takes 80% of the time, and having people fill in where ever they can really helped a lot.
In later posts, I’m going to a recap of the events. I’d also like to talk a little more about Skribit, which I have high hopes for.