"Psychological safety: Can we take risks on this team without feeling insecure or embarrassed?" - from re:work
There's a great NYTimes article about it that looks at your hunch about the weirdness of the team members being the key to awesomeness. But Google ultimately realized that it wasn't so much about how different the backgrounds of the team members were, but was instead that that internal diversity could produce psychological safety.
http://www.nytimes.com/2016/02/28/magazine/what-google-learn...
Makes no sense to me. Why should the team need to be different? Why should the team need to be the same? This is Voodoo recruiting - some belief that a company has formed that really is meaningless but becomes an important part of the companies recruiting culture.
Will our team be different? Are they weird? Shake the chicken bones, dance around the fire and we'll get a good hiring decision.
It wasn't always a deal breaker, but it was very important that the team would be different after hiring someone. We hired quite a few non-CS grads, whether they came via a boot camp or were just self taught.
We ended up with a ragged team of misfits and it was awesome. I've never worked on a team that was so effective at challenging assumptions and biases and shipping features faster than anyone else. I'm convinced it wasn't through Herculean efforts by individuals, rather the product of clear communication and trust within the team.
Ive tried to explain why this worked. One observation: Very rarely did someone make implicit assumptions. Such assumptions are often wrong, but they happen because individuals are similar enough that extrapolation a partial understanding into a full one implicitly happens. We assume that since we got from A to B on the same path that the path from B to C is likely the same. On a team of misfits, you have to clearly communicate the entire sequence of events because everyone is on an entirely different page to start with. The result is that the final product is 100% on target, whereas normally there are a few deviations as the result of implicit assumptions.