| Well if for every X product area you need: - N people to just juggle live incidents/emergencies. Actually triage, live, why the Desantis campaign launch on Twitter isn't working. Give the server more RAM, spin up more instances, pray it works. - M+N people to do the work to prevent the incidents/emergencies (maintenance). Actually ensure the next big giant event on the streaming service doesn't cause failures. Create automatic scaling up of new capacity as needed. But avoid overprovisioning and costing the company a bazillion $$. Get to the bottom of what triggered the incident. Build practices, runbooks, etc the next time this happens. - M+N+L people to actually create new stuff. Create a new feature on the live streaming service to comment or somesuch. (and all the work M+N people do to handle incidents with this new feature) So X * (M+N+L) people if every product area is trying to create new stuff. But the reality is most teams have T people where M < T < (M+N+L). Which is very company dependent. Spotify has dozens of consumer facing product areas. Probably many more for creators, record labels, admins, etc etc |