You are right. I worked as BI consultant before. In our company the three main challenges we had were :

* collecting data * having clean data * extracting the right indicators and analytics from the data

I think the problems remain the sames with QS. Actual QS solutions lack of interoperability, do not really allow experimentations and do wrong about privacy. That's why Cozy could help: it allows to build data collectors in a few hours, a data browser is available to clean wrong data, and the best of all it allows anyone to hack new stuff on top of this new set of data without compromising privacy.

disclaimer: Cozy member