Post-mortems of game development should instead be an on-going process, says Tim Heaton
Game post-mortems are like the Catholic church’s practice of confession.
They’re an opportunity to face your sins and disclose them to ‘the boss’. Your producer or publisher maybe, who may like to think of themselves as omnipotent, but possibly aren’t yet the spiritual leaders of a formal games business religion.
Penance may involve repeating ‘we won’t do that again’ multiple times whilst throwing all the old worthless schedules into the bin. Afterwards you can feel cleansed and start a new project afresh. If you’re lucky.
Most team members will have participated in a post-mortem at some stage in their career. Sometimes it consists of simply listing the five most significant things that went right, and the five that went wrong. They’re printed in the games development press and published on websites, and give an often-entertaining insight into the tortured process of getting a game to market.
But they’re a bit hopeless these days don’t you think? One problem is encapsulated in the name. I’m no doctor but if you’re only figuring out what’s gone wrong after death you’re probably too late. It may be interesting to know, and further deaths may be avoided, but you are still dead. Or worse – you’ve overspent on your budget, you’ve shipped late with a 60 per cent Metacritic, and your offices are being measured up by the latest internet bubble company with venture capital to burn.
In addition, when you read a post-mortem the feeling of déjà vu is overwhelming. Let me tell you the contents of your next post-mortem; you went into production too early, your tech wasn’t ready for the content, the key design feature didn’t deliver when implemented, and you didn’t take multiplayer seriously enough. I know that because I’ve read all of them, and committed most of the deadly sins myself.
Nowadays there’s rarely a ‘Wow – let me note that down’ revelation for us all to learn from.
And finally, waiting until the end of a project doesn’t really address the fact that in most of game development the devil is in the details. It’s human nature to diminish some of the trauma and aggravation of prior experience; it’s how we move on.
I think it must be how mothers are prepared to have another baby, and may explain why we keep ending up with governments we don’t really like. So some of the real transgressions are lost in the haze of the last few months of crunch – and you’re probably six months late because you didn’t start fast enough, not because you didn’t finish fast enough.
So, what’s to be done? Firstly you need to create an ethos within a team that is open and honest. One where anyone can be happy to voice opinion, without fear of ridicule or being called a troublemaker.
Secondly, it needs to be a frequent and easy process during development, not at the end. The lower the barrier to entry the more it will be used in an instinctive way. Here at The Creative Assembly we are experimenting with an online system, allowing easy entry of issues, transparency, and a voting system that hopefully highlights the importance of the current issues.
We’re doing these monthly but can easily increase or decrease the frequency of the voting. Issues are then discussed at team meetings, to give immediate feedback and action work to be done. Anyone in the team can propose an issue, but the wording is very important and, like all ‘customer surveys’, questions or statements need to be both finessed and clear. Issues range from the scope of the game to the temperature in the studio.
Thirdly, and this is the big one, you need to act upon it. Even when an issue is logically and pragmatically appraised and a decision is made to not confront the problem, the team needs to understand that, and trust that the decision is made in the best interests of themselves and the game.
And for those issues that need resolving you need to move fast, and be seen to act. That way the team buys into the process more and more as a positive force for change. Ignoring issues or letting them drift will slowly kill any future willingness to contribute.
So, I’d like to be able to say we have the perfect pre-mortem system in place at Creative Assembly, but we’re still experimenting. The online system provides a great opportunity for tracking and evaluating what’s really disrupting the team, while it’s happening. We’ll continue to try and learn from our mistakes, and atone for our sins. One thing’s for sure, none of us are infallible.