You know almost immediately if any problems occur in the process as well as in large batches you waste much more time sorting and organizing the different segments of the process.
Hardware becoming software - This has already begun on our team but using CAD to prototype concepts, something we haven't done within CAD however that could rapidly speed things up is designing using parametric constraints which I could explain in person or using a computer but I don't know how to explain here... basically setting CAD dimensions equal to other dimensions so that if you change 2-3 base dimensions the entire object will update to the proportional size (would only be useful sometimes but would save a lot of time when it is).
Fast production changes - If we designed a robot to be flexible to multiple ideas beforehand it would be much easier to adapt quickly.
Rapid prototyping tool - If we build a protobot extremely early in the season and are able to rapidly develop (go through bml loop) it our final robot will be much better as a result.
They use a feedback loop much more often in their businesses than we do in school or even robotics. And they are able to work more effectively because of it. They use their feedback as soon as possible to reiterate sooner.
Our team has a sticky growth engine, we rely on getting members and retaining them. In our specific case next year acquiring people will be more of the problem than retaining others, but in following years retention will become more important as many of the people who were in FLL will join us at least for 1 season and its up to us to make their experience good.
We lose people for a variety of reasons which only some of we can control, we lose people to things we cant control such as them moving (can't control) , others we can such as them having a bad experience or commandeering too much of their time (can control). Sponsors might stop funding us because we don't show appreciation enough (easy to control), or because it just doesn't work for them (can't control).