Katie Bauer!

Trying to remember where I read abt a framework for experimentation where we set a threshold for acceptable # of negative experiment results to ensure teams arent too risk adverse & avoiding bold bets. Like an OKR targeting 30% negative A/B tests. Anyone know what Im thinking of?

The tldr was to treat experimentation as a learning exercise & make teams feel supported when they have to deliver bad news. I think it might’ve been written by someone at Slack circa 2019? @imightbemary @shaayohn @GisselleXie @robinson_es @seanjtaylor @bjoseph @far33d @j_houg

sounds like something @ronnyk would say - maybe in a deck like this https://exp-platform.com/Documents/2017-05-17EmetricsControlledExperimentsPitfallsKohaviNR.pdf … ? His book, Trustworthy Online Controlled Experiments, has some nice sections that discuss a "crawl, walk, run" model of using data as a business that are relevant. It's worth reading.