What does it look like to be successful here? How is my performance evaluated?
Performance management in our Product organization
This is copy/pasted from our internal blog so some links will not be functional.
We focus on inputs over outputs when we evaluate performance.
Our team standards encapsulate what we mean by ‘inputs’. You may also hear us say, “We evaluate the quality of thinking.” ‘Quality of thinking’ is an abstraction that managers may refer to when thinking about the judgment of our team members, as well as how they interpret information, communicate what we should do in light of that information, and act (in other words, how our standards are applied day to day). Even though ‘quality of thinking’ is abstract, managers should exude clarity when suggesting that your quality of thinking could be improved.
Why do we avoid using output targets to evaluate performance?
I’d put downsides in three buckets:
Culture
There are three beliefs about culture we try to hold in our heads at once:
Culture is the set of behaviors that are perceived to be rewarded or punished.
We want to build a customer-centric culture.
It’s virtuous to strive for a sharply defined culture.
Given (1), it’s hard to imagine being successful on (2) or (3) if you reward progress on outputs.
On (2), small things like helping a customer onboard or resolving a worker’s support inquiry will not show up in aggregate output metrics (especially at our scale), but we want to reward those behaviors. Of course, we want to be biased toward higher leverage solutions, but everyone on our team should be put in positions to viscerally feel our customers’ pain.
On (3), you lose cultural coherence if you start rewarding behaviors that you can’t control. We can’t control outputs.
Instead of operating in that world, we prefer to hire ambitious, curious generalists and then give them business context. We then reward their inputs like “solving our core customer problems, quickly” (in service of crafting a sharply defined culture).
Tunnel Vision
You may have heard us reference Goodhart’s Law. If we start turning metrics into targets that influence performance management on our team, we’re worried that: i) the ground can shift out from under you – maybe you chose a good target two months ago, but the marketplace changed, and ii) you no longer wear your “owner across the business” hat and focus relentlessly on your target. Both can lead to suboptimal business decisions (especially in marketplaces).
Risk Aversion
We don’t want to create incentives to be slightly more risk averse. In many areas of Product work, slugging percentage is more important than batting average. We don’t want team members to be nervous about swinging for the fences.
But aren’t we here to grow output metrics?
No. We do want to build the largest business possible, but we are here to create value for customers. If we execute on that, we are confident that we can capture value.
If this is the case, what role do metrics play at Clipboard?
Metrics still play an essential role in the day-to-day.
We expect team members to be deep in our data, as well as marry that understanding with customer conversations. Metrics point us towards what we should be focused on and throw off clues for further digging.
Additionally, we are big on “bet-level” metrics – both in terms of “success criteria” (what impact tells us that this new work is worth the maintenance costs?) and “guardrail metrics” (what could we inadvertently mess up in this initiative?). These are essential inputs into prioritizing our portfolio. Our ability to craft good metrics here is also a component of our “quality of thinking”.
Do we get rewarded for meeting our success criteria?
No. While we want to meet our success criteria, we celebrate excellent inputs. We have many examples of bets that failed on outputs, but we commended (and rewarded) team members because of their inputs.
I assume this isn’t a free lunch. What are the tradeoffs to keep in mind when we prioritize inputs over outputs?
You create dependencies on:
Clear-enough product strategy that team members can act on/make independent decisions with
Combining “communicating in the open” with normalized debate so that we can quickly iterate on how the organization is prioritizing
Strong hiring machine that can identify the aforementioned ambitious, curious people
Groups of managers with extremely strong judgment
What does it look like to ‘outperform’ on this team?
Team members can see org-wide performance score definitions here. To restate, a ‘3’ means you are performing excellently and meeting our high standards. When I think of outperforming, I think of not only meeting/exceeding each of our standards, but also pushing the org forward in one or more dimensions outside of your immediate scope. That might look like driving high quality improvements across company-building activities (eg hiring or onboarding) or helping push the business forward in another function (eg Ops) or accelerating an unrelated product area (eg Pricing).