Thoughts on engineering promotion process @Pinterest Inc.
Pinterest’s engineering org is around 350 yet the promotion process is ridiculously heavy weight inspired by behemoths such as Google. Let me give you a stat to kick off the discussion — “Close to 1000 hours is collectively spent in creating the packets, reviewing the packets and appeals by the promotion eligible candidates, promotion committee members and facilitators” every six months! Let that sink in a bit. Thats a staggering amont of engineering effort for an org. of 350 engineers. The effort that could have went into improving the product. You may think this ensures fairness and objectivity but unfortunately the system is skewed towards engineers who are eloquent at writing essays than writing code. I have read packets where a particular engineer caused grief multiple times to backend engineers by breaking their stuff repeatedly but cleverly hand-picked his immediate team members as peer reviewers who would support him/her and wrote a great essay (and eventually got promoted) and an engineer who made meaningful contributions but struggled to showcase his/her work. His/her packet was full of grammatical errors, typos and poor sentence formation that clearly showed the engineer has struggled to write their packet. Why is Pinterest forcing this on engineers? Why can’t the managers take this burden like most other companies that are successful? Doesn’t Pinterest trust the decisions of its managers? If not, how about hiring an external company to handle the promotion stuff? The committee members anyways heavily rely on whats written in the packet for making a decision. Why waste engineering effort for reading essays?
Another criticism about the entire promotion process: an engineer should demonstrate next level’s ability sustainably which means an engineer who is working at next level is underpaid for a full year.
I believe companies can operate without assigning arbitrary levels to engineers when they are relatively small. Scrap the levels, reward high performers with sizable bonus, fire the low performers and focus on iterating the product fast.