6 goal-setting techniques from Mixpanel’s VP of Product and Design
One of the hardest things about using metrics to guide and strengthen your product and your team is that there is just so much information available. Choosing where and how to focus is key to making successful decisions. In this article, Mixpanel’s VP of Product and Design, Neil Rahilly, outlines 6 crucial steps to set, achieve, and celebrate your business’s data-driven goals.
1. Focus on one goal.
If you get up in the morning and think about one metric and one metric only; if you work all day on one metric; and if, before you go to sleep at night, you’re still thinking about that one metric—you’ll operate with greater focus and with clearer, more immediate, results than if you’re trying to improve 25 different things at once.
This hyper-focused approach depends on deep pre-planning. You have to do the hard work of deciding what’s actually the most important thing to do before you do anything. If you prioritize several goals at once, you can rationalize anything, or never agree about what should be done.
Here’s an example. Our customer education team used to measure their success with lots of different metrics, and they spent time on lots of different activities (like helping the marketing team create content for SEO). But when we narrowed down their goals to just one—reduce the load on customer support teams through one-to-many support—we agreed that they would concentrate on reducing the number of support tickets by providing customers with the information they need to self-serve.
2. Bake in priorities.
For a long time, our performance engineering team measured query latency—the time it takes for a request to travel from the browser to the server and back again—as an average, or percentile, across all requests for all customers. This had the effect of weighting all queries from all accounts as equally important. When we reduced the 90th percentile latency from 2 seconds to 1.5 seconds, we celebrated improving things by 25%. The problem was that, meanwhile, the queries of some of our very largest (and highest-paying!) customers were taking minutes, not seconds. A huge chunk of ARR (Annual Recurring Revenue) was in this long tail, so it was better to eliminate the awful experience of waiting a minute than to marginally improve what was already decent.
So we changed the metric to “FastARR,” which is the percentage of ARR associated with accounts for which all users are experiencing acceptable latencies. This tells us how much ARR is at risk because Mixpanel is too slow. That’s what really matters, and it orients the team toward solving the biggest problems for the biggest customers when that’s what makes sense.
3. Let your customers measure success.
When you use metrics like product adoption rate or churn, you’re aligning your goals with the success of your customers. This is different from tracking deadlines or the number of features shipped — things that might make you feel more in control but ultimately don’t help your customers succeed. A goal like reducing customer churn, on the other hand, forces you to ask why your customers churn, and to look for a solution that will make the most difference.
4. Be realistic.
Gather your team and, together, answer this question: What’s the least we need to do to achieve one huge success? Then focus on doing exactly that. And always remember: You have three parameters—your target, your time, and your resources. If you give a team more goals than they can reasonably achieve, you’ll get either mediocre and half-finished work or a burnt-out team—or both. When you’re not realistic about the resources that you have, you don’t do the last mile of planning.
5. Don’t go backwards.
Say you ask your team to lower data center costs by 20%. The team goes away and completes the task; the data center costs will indeed be 20% lower next month. But suddenly, someone bursts into the room: “Everyone’s queries are timing out! Mixpanel.com is so slow! It takes 30 seconds to load!” You turn to your team and they say: “But YOU said to lower the data center costs.”
In other words, don’t let your one clear goal blind you to system complexities and necessary constraints. (I call it the Universal No Regression Principle, or UNRP). Your goal should come with an implicit assumption that nothing else will regress as a result of your work unless such a trade-off is explicitly agreed upon. Actually, make that an explicit directive, just to make sure that the whole team is on the same page.
6. Build in the reward 🎉
Okay, you’ve identified your goal. What next? Build in the reward. When we achieve X, we’ll celebrate with Y. In too many cases, people hear messages like “You’d better hit that goal or there’s going to be trouble.” Not exactly the way to motivate a team to put in the hard work you need them to do….
At Mixpanel, we have a fund called “Dollars and Sense,” which is topped off with money we’ve saved through data-driven cost reductions. If we reduce a monthly recurring cost, one month’s worth of those savings goes into the account. We spend it on ourselves—we have fun, we hand out bonuses, we buy people presents, we reinvest it into the things that our team wants. In my experience, it’s the very best way to focus our collective energy on achieving our goal.
Metrics are one of the most powerful tools that we have to guide our business decisions—what we change, what we maintain, what we value. Taking a disciplined, focused approach that involves and motivates your team is the best way to harness the power of data-driven decision making.