It would be pretty interesting to do A/B testing starting out with a group of people who regularly drink a lot of Soylent. I haven't heard of anyone doing that sort of testing, though.

The experimentation that the Soylent folks are doing themselves looks like the fairly uncontrolled do-it-yourself kind, not like a real scientific study. It seems like it would be difficult to do properly because the formula of Soylent changes quite often.