How to choose a north star metric in a PM interview

“What's the north star metric for this product?” It sounds like a quick question with a right answer. Most candidates say daily active users or revenue and move on. Then the interviewer asks “why that one?” and the answer starts to wobble.

Picking one metric is hard on purpose. Every choice leaves something out and the interviewer wants to see whether you know what.

What a north star metric is for

A north star is the one number a product team agrees to grow because it best captures the value users get. It's meant to settle arguments. If two projects compete for the same engineers, the team backs the one more likely to move the north star.

That means a good north star has to do three things at once:

1. Reflect real value to the user, not just activity.
2. Lead revenue rather than lag it, so the team can act before the money moves.
3. Be something the team can influence week to week.

Why the obvious answers fall short

Daily active users measures attention, not value. People can open an app every day because they're stuck or confused. Revenue is real but slow to react and easy to grow in ways that hurt users later, like raising prices or adding fees. Time spent sounds good until you remember that some products are better when they're faster.

None of these is wrong in every case. They're just not automatically right. Naming one without a reason is what costs marks.

A worked example: Fennel

Fennel is a fictional recipe and meal-planning app. Users pick recipes for the week, Fennel builds a shopping list and they cook at home. It makes money from a monthly subscription.

Start with the value. People don't really want recipes. They want to eat well at home without the hassle of planning. The moment of value is a meal actually cooked, not a recipe viewed.

Option three wins. It ties directly to why people pay: a household cooking with Fennel twice a week is getting what it signed up for, so it's the household most likely to renew.

The threshold is a choice you should defend too. Once a week might be luck. Twice starts to look like a habit. Say you'd check the threshold against renewal data once you had it, rather than pretending it's proven.

Counter-metrics: the half most candidates skip

Any single number can be gamed, even by a well-meaning team. If Fennel's team is rewarded for cooked meals, they might send more notifications, nudge people to tap “cooked” or plan easier recipes nobody enjoys. The north star goes up and the product gets worse.

So pair it with two or three counter-metrics that would catch the damage:

1. Notification opt-outs, to catch nagging.
2. Average recipe rating after cooking, to catch quality slipping.
3. Monthly cancellations, to catch growth that doesn't last.

You're not trying to grow these. You're watching them to make sure they don't move the wrong way while the north star climbs.

“My north star would be households cooking two or more Fennel recipes a week. Alongside it I'd watch notification opt-outs, recipe ratings and cancellations, so we don't grow it by annoying people or making the food worse.”

How to answer in the room

1. 01Say what value the product gives, in one sentence.
2. 02Name the moment that value happens.
3. 03Offer two or three candidates and reject the weaker ones out loud.
4. 04Pick one, with a threshold you can defend.
5. 05Add counter-metrics and say what each one guards against.

That takes a few minutes and answers the follow-ups before they're asked. A common next push is “what if the north star goes up but revenue falls?” If you've already named cancellations as a counter-metric, you're halfway to the answer.