Methods don't seem robust enough to establish Rippling was true cause of 1/2 headcount.
Potential stronger follow-up study: aggregate a set of pre/post case studies of real new Rippling clients to asses if they reduce headcount by X%, Y mos post onboarding. Perhaps in progress!
@parkerconrad I'm curious about two aspects of the methodology BSG used in the study:
1. Did BSG adjust for industry or firm age across Rippling and Non-Rippling cohorts? Headcount norms will vary by both, would assume the Rippling cohort skews tech-adjacent, newer firms.
2. It looks like LinkedIn data was used to establish Non-Rippling headcount data. What was done to verify this data? If only LinkedIn data was used not actual company records I'd guess this dramatically overstates active HR/IT headcount of firms.
@AkshayNarisetti Any chance you’d open source the code and allow @betaboxlearning to allow kids to build a similar project in school? Looks super fun 🤩
@SievaKozinsky This happens IMO because far too many startups are better understood as a baby cashflow machine embedded inside a status signaling machine. Raising money is the optimization function of the signaling machine but just an input for the future cashflow machine.
Are the recent developments in AI showing us that the only way to achieve a socialist utopia is through ruthless, capitalistic technological advancement?