Making decisions with hard trade-offs
Today's memo is sponsored by Flatfile: Untangle the data your customers need to import with Flatfile Embeds. Enable 1-click data imports for your customers with HIPAA, GDPR, SOC 2 Type 1 compliance out the box. Cut onboarding time-to-value with Flatfile. View the platform.
The job of a product manager is to make good decisions. The most difficult decisions are the ones that require tough tradeoffs. In a perfect world, all metrics go up -- but in reality, some metrics go up and others go down. Especially if another team or priority is getting the short end of the stick, you can’t justify a tradeoff in your head -- you need a clear framework to convince the other team that the tradeoff is the right thing for the company. It’s no surprise that tradeoff questions frequently show up in PM interviews, including Facebook’s PM Execution interviews. Here are some guiding principles to tackle the Execution interview, as well as in your day job when you face a tough tradeoff question.
There are two types of tradeoffs decisions you’ll encounter:
You have to choose between feature A and feature B, when outcomes for either are speculative. For example:
You’re the PM for the new user sign-up flow at Facebook. Profile photos are mandatory today. Would you make them optional?
You’re the PM for Facebook Notifications. How do you determine whether to show a friend’s fundraiser or an event a friend is going to?
Metric X goes up but an equally important metric Y metric goes down. For example:
You launch a feature for bicyclists on Google Maps. The # of bicycle trips goes up but the # of car trips goes down.
You launch a feature to improve watch time on Facebook Watch. Watch time goes up but likes and comments goes down.
1. Anchor on the “why” before diving into the options.
Example: You’re the PM for the new user sign-up flow at Facebook. Profile photos are mandatory today. Would you make them optional?
Bad PM: Anchors in the action.
Jumps into how they’d run an A/B test to increase sign-ups.
Good PM: Anchors on the why.
Facebook’s mission is to bring the world closer together. And the more people there are on Facebook, the more useful it is. The goal of the new user registration flow is to enable people to enter in just enough information to help friends identify them, while not introducing so many requirements that people drop off when signing up.
2. Define a good scenario, a bad scenario, and ways to measure them.
Example: You’re the PM for Google Maps. You’ve launched a new feature for bicycles. The # of bicycle trips goes up but # of car trips goes down. What do you do?
Good PM: The overall goal for Google Maps is # daily active users per day. Even we see a decrease in car rides, it doesn’t matter because our goal metric is overall # trips.
Great PM: Suppose we want to encourage more people to bike because it’s better for the environment than driving, so we launched a feature that make it easy to find bike routes without steep inclines.
Great scenario: More people are riding bikes instead of driving, because we’ve made it easy for them to discover a route that’s easy to ride. Our feature launch has been successful.
Bad scenario: We introduced some bug that is affecting car trips.
I’d want to look at data that tells me where the replacement activity is actually happening. Constraining a variable can be helpful to root cause your problem. Specifically, for distances that are feasible for biking, are drivers using Google Maps less, or does the drop in usage include distances across the board?
3. Remember the solution doesn’t have to be binary. Brainstorm variations that create a win-win.
Example: You’re the PM for new user sign-up at Facebook. Profile photos are mandatory today. Would you make them optional?
Bad PM: Makes a binary recommendation.
If the experiment shows that making profile photos optional leads to more fraudulent accounts being created, then no, I would keep the photo as mandatory.
Good PM: Explores ways to capture the upside and minimize the downside.
While requiring a Facebook photo may make peoples’ profiles easier to identify and more engaging, it may also be a burden, especially for people with limited internet bandwidth. There are some alternatives here we could explore, such as picking from a set of placeholder profile photos, or making the photo optional on sign-up, but encouraging people to upload a photo once they’ve created an account.
When you face tough trade offs as a product manager, try zooming out from the two-dimensional view of “should we do X or Y?” Instead, consider the “why,” define good vs. bad scenarios, and remember that the solution may not be binary. These guiding principles can help.