Skin in the Game

You can't have a seat at the table unless you have skin in the game

One of the most important things I learned during my product management rotation was the feeling of total responsibility for the outcomes of the product I was working on. As an acting PM, it was drilled into me that it’s not sufficient to diagnose problems, but I also needed to figure out how to fix them, and then see those fixes to completion. We were also told to take the blame of failures, and give the credit of successes. Even after being a data science manager, where I assumed I was already exhibiting these traits, this was a bit of a rude awakening during my time as a PM.

In general, my experience is that insights-generating functions like DS or UXR will take credit for the recommendations that turned out well, but don’t or can’t necessarily bear the responsibilities for the recommendations that failed. This, to me, is a big problem of not having enough skin in the game, and a hindrance towards such functions having a seat at the table or being co-equal functional partners to product and engineering. That is, if you’re praised when things go well, but aren’t “yelled at” when things don’t go well, then you’re probably sheltered from having full responsibility.

Within analysis-oriented data science teams, this issue holds true more for insights generation than for evaluation, because you tend to get quick and vocal feedback if your evaluation is bad. If your methodology or process is blocking a team from launching something good, there’s no question you’ll hear from them, very loudly. But for broad strategic recommendations that may take years to manifest, it’s very likely that a reorg or leadership change destroys the chain of accountability before the true impact is well understood.

At Google, where there was a lot of paranoia around the Peter Principle and of people failing upwards, I noticed many instances of high-level promotions for flashy product or process launches that turned out to be spectacularly or unexpectedly unsuccessful after a few years. I don’t think there’s an easy solution to this problem, because you can’t withhold promoting someone by virtue of their working on a very long-term project. That would disincentivize such projects which are quite necessary for the long-term health of a company. Furthermore, there is significant learning value that comes especially from making costly mistakes.

What I do think is required is that insights-generating functions need to be exposed to both the shared successes and failures of the product. This means such functions require some longer-term accountability mechanism that eventually allocates credit according to true, long-term impact. I suspect that this is feasible for directors and VPs, given that their strategic initiatives are typically measured on multi-year timescales, though there’s undoubtedly incentive to push for assessing shorter-term impact. For individual contributors, perhaps the best that can be done is to ensure something is learned via postmortem or otherwise.