GrowthBook

"We decided it was a better use of our time to get teams to run experiments, as opposed to continuing to build a thing that would allow them to run experiments." That's how Mikhail Popov, data scientist at the Wikimedia Foundation, explains the decision to stop building their own experimentation tools. Wikimedia serves 15B+ page views a month across 300+ language communities, and it has promised those communities that personal identifiers never reach its servers. So GrowthBook runs entirely on Wikimedia's own infrastructure, cut off from outside networks, and analyzes data where it already lives. When Wikimedia needed GrowthBook to work with its authentication system, we built that with their engineers and added it to the open-source product so everyone can use it. Analysis now takes about an hour per experiment instead of four. PMs and engineers set up their own metrics without waiting on SQL from data science. Read the full story: https://buff.ly/PEXGhmW