Butterfly Effect

One of the very important lessons from that was how ever much you give out to your external teams is what you're going to get back in. It sounds dead simple to the point of dumb, but if a product came back that was not up to quality par, after looking at it maybe we weren't providing them with enough information. That's one of the key things we learnt – you can't expect them to succeed if you do not invest time into them.

On Mass Effect 2 we spent a lot of time focusing on those aspects and seeing how can we ensure the communication channels are really tight, how can we improve the quality of the art that went out? And we also tried to catch all the really big things early on. The whole philosophy is how can we help them succeed rather than shove this on them and wait for it to come back. The basic way of looking at it is here's a contract, here's some money and bring us back our assets. But the other approach is how can we work together? And that pays off because we have a huge initial time investment but then over time you can walk away hands free. And if you get that machine running they will turn out the work. We were able to break things down.

Looking back we saw [external partners] were making extra textures that we didn't really need or use and therefore the assets were less efficient in game. So we said okay, let's make a standardised shader and by creating this repeatable set of actions for them we set a certain expectations from them and they get used to making it and they're able to make it 200 times. So long as the base pipeline is really efficient we found out that we're able get far more high-quality assets out, they're spending less time on them, and we can move much more quickly. Internally we might be spending between five and ten minutes per asset. But if you do that 200 times it adds up really quickly. So it's taking those small gains and sending it back out. As a whole picture we're able to look at small areas that we're able to get maximum gain from and tighten up on the areas that need working on.

We shared a lot. Some of the principles sound absolutely silly but once we see the benefits of it it becomes obvious. One of the things that we didn't do with the first game was we didn't share tools and engine. We were secretive and you have to be careful of brain drain as well because you're training up and your knowledge base is sometimes what keeps you ahead of your competitors. On the other hand, if you're sending out work and it's not meeting your internal quality you don't want that discrepancy.

So for Mass Effect 2 we decided that for trusted vendors it was very important to be as transparent as possible and that philosophy actually worked out very well. We were actually quite surprised that the more we shared, the better the work that we got back. The realisation came when there was a huge ramp-up time and we were working with a studio called Blur, who do a lot of high-end CG work for trailers, and we gave them a lot of access to all our high-res models and all the information we had and worked with them very closely. The interesting thing that happened was artists are in a very small circle and our artists new their artists so we ended up learning from them. So we looked at some of the changes and were able to put them back into our work. It was a very symbiotic relationship were the philosophy was very much don't treat them as a bastard child. Even if they are only working with your for two or six months, treat them as you would like to succeed.

That's something we learned from the first game. The second time around we looked at the risk factors. We still kept a lot of the key things in house, but we were able to send out a certain type of work and there was little risk that someone would steal that and make their own Mass Effect. The benefit was ten-fold back to us, you gain so much back. We worked with a Scottish outfit called Axis Animation too, and we were really happy with their work. They represent a new studio that we didn't work with before but we were happy to be transparent with them and they stepped up to the plate and did an amazing job. They were able to bring in interesting things. In the end it opened our eyes because it's a very fast-moving business world and your level of engagement can be very short sometimes. So even if your company has a short contract you want to get the best out them in a short time. And if we go back for future work the people might not be there. You have to make the most of what you have.

I think we're lucky in the sense that with DLC you always want to put in a small amount of innovation each time to keep things fresh and avoid becoming a production line. So with the Kasumi DLC we've tried some new game elements, few introduce spy elements to the gameplay. That was fun for the internal team because it's a new set of challenges. As we do that sort of work between the two major projects it's a good way to test out new ideas.

At the same time we can never veer to far from understanding what made the game great. The hard thing internally is that everyone wants to make change but we have to recognise that a lot of things got us up to that 96 score on Metacritic. So it's very dangerous because any changes to the formula might be good and might be bad and you're walking this edge. We recognise that it's not a perfect game and there are still areas we have to work on. I think we're carrying forward the same kind of design philosophy that we did on the first one to the second one. When the first came out we broke down every bit of feedback we got, good or bad. And eventually we broke it down and addressed those issues. Really it's no different for the next iteration. It's about evaluating the work and the process and then making slight adjustments and testing it out as scientifically as possible. The project director and I went to the same engineering school so we think along the lines of a feedback loop. You have a hypothesis of the direction you'd like it to go, you put it out, see what the reception is, see how it actually ended up and then you make your subtle tweaks and come back and do your next iteration. Each time you're really leveraging against that foundation and time and investment you've put into it and hopefully edge towards higher and higher sales and critical response.