RE - It's OK to be afraid

A while back, Kodel wrote a post about a team and their fear before going live. We all know this I guess. Being in the app development business, this pops up every time before we push a new app to the public. And every time I ask my team the same question “Are we good to go? No major bugs?”. Up to this day they respond with “Not sure, I couldn’t find anything…but that doesn’t mean I might not have missed something…”. And the uncertainty rises.

I could have chosen to not ask and proceed with the deploy, but then I would be taking all responsibility instead of them. And that is, at least for me, a very important difference. Notice how I do not ask them if everything is bug free. I do not believe in such a thing exists, but I do believe you can work towards a certain degree of predictability. A point where the response from my team would be “Sure, let’s go”.

There’s even a platform out there which assists in the uncertainty by allowing staged rollouts. Far too much has this been used in the past as a safety net. The problem is not the feature, there are situations where it can be really useful (think of scaling up at a certain pace), but when developers use this as a sort of testing mechanism to quickly adjust stuff on the fly while deploying, I can’t tell if they are being extremely agile, or just down right stupid.

So I’m helping my team build better apps, make sure we find and use great frameworks that help us write code which can be tested, assist in improving the way we deploy our apps to the stores and up the frequency for beta versions so we can get feedback earlier on and fix what we missed. We started to write more unit tests by using test driven development and do cross platform code reviews to learn from and assist each other.

Because the only way to do way with the fear is by knowing what’s in front of you. It’s about writing great code and feeling proud of what you do. Only then will you be able to deploy and focus on what’s next instead of ruining an entire sprint on fixing what should have already worked. Are we there yet? Almost. We ruined a deploy a few weeks ago which resulted in this post by me. I asked the very same question to the team and the response was indeed not what I wanted it to be. But we all agreed we would help out to get to that point, because no-one can do this alone. It’s a team effort. So next time when I ask them, the answer will be a lot more optimistic.