Adoption Should Never Be Your North Star Metric

A product manager came to me with a metrics question that four other people had already answered for her.

Her team was rolling a new meeting platform out to the whole company, and she was the product lead for the AI summarizer that shipped with it. She did not get to choose the platform. She did not get to choose the constraints it shipped under either. None of that was hers. Then leadership handed down the last one, which was the one that actually mattered to her career: the project would be measured on adoption.

By the time we talked, adoption was sitting between 45 and 55 percent of eligible employees, and divisions kept walking up to ask her what her adoption North Star ought to be. She had run enterprise products for the better part of two decades, and she still could not name a number that meant anything.

She could not name one because there isn't one.

Why adoption feels like the right number

Adoption is the number you can produce on day one, before the product has had time to do anything at all, and that is most of its appeal. It moves. It goes up when you ship, it goes up again when you fix onboarding, and it drops into a quarterly business review without anyone having to defend a claim about value. Every other number in that room has to be argued for, and this one only has to be reported.

She had already found the crack in it, which is the part worth noticing. Once the rollout finished in October, everyone eligible would have the summarizer by default, so the population and the number would stop moving on the same day. A metric that flatlines the moment you finish shipping was never measuring the product; it was measuring the rollout.

That is what a vanity metric actually is, and the family is larger than most people think. There is adoption, which tells you a login happened. There is engagement, which tells you an app was open, possibly in a background tab, possibly during a meeting somebody else was running. There is usage, which tells you a button got clicked by a person who may well have been hunting for a different button. Each one is real data about a real event, and not one of them says that a human being was better off afterward.

Adoption tells you the door opened. It tells you nothing about what happened in the room.

The sentence you write before you measure anything

Here is what I told her. It is the line to take out of this letter. Your adoption should never be a North Star metric. Your North Star should be about the value your customers get out of the thing and how you maximize that. Adoption is a leading indicator sitting somewhere in the middle of the chain, and it is not the destination. Put a leading indicator at the end of the chain and you get a team that spends three years optimizing the middle.

The work that fixes this happens before any measurement, and it is one sentence, said out loud, in front of the people who will be reading the number later. We believe that if these specific people use this thing in this situation, they will get this outcome, which produces this result for their business, which produces this result for ours. Five links. The last one is not optional, because a chain that stops at the customer's outcome is three-quarters of an argument. For her summarizer the sentence ran close to this. Teams who use it to capture their notes leave the meeting with more clarity and more speed, so they hold fewer follow-up meetings. That means the same decision at fewer paid hours, and more decisions per quarter out of the same headcount. Every link there is measurable. Not one of them is adoption.

Similarly, picture a gym reporting new memberships. Sign-ups are genuinely useful and the owner should absolutely watch them, but the gym does not make its money on people who sign up. It makes its money on the people still showing up in March, because those are the ones who renew. A gym that hangs sign-ups on the wall as its north star will spend its entire budget on January promotions and none of it on the reason anybody stays.

DO NOT start by renaming the metric on the dashboard. A PM who swaps the North Star line in the deck without changing the conversation underneath it has just given leadership a new word for the same number. Four minutes later somebody in that room asks what happened to adoption, and now you are defending a relabeling instead of an idea.

Do not hand a leader the metric, get them to name it

Her harder problem was never picking the right measure. It was that several levels of leadership had already agreed on the wrong one. Walking in to correct a decision they have already made is the fastest way to lose an argument you were right about. So I do the opposite, and I am not especially proud of how well it works: I act a little dumber than I am. I ask how people are really going to use this thing. Sure, they'll adopt it. What do we think somebody gets out of it in a training session, in a weekly decision meeting, on a call with a customer? And then, the moment anyone in the room describes an outcome, I get visibly interested and ask whether we could measure that.

At that point the new North Star is their idea, which is very often the only form in which a senior leader will accept it. Then you repeat it back with the conditions attached. So if I understand you right, if we could confirm we're saving those teams hours every week, we'd know this is worth scaling. They can agree, they can push back, they can reshape it, and every one of those is a win, because the conversation has moved off adoption without anybody having to be wrong in public.

Run this with the person who talks the most in your reviews, not the person who agrees with you the fastest. The agreeable one costs you nothing and buys you nothing, and the loud one is the whole reason the metric is stuck. And do not run it in the meeting. Run it in the meeting before the meeting, one-on-one, where they can think out loud without an audience to perform for.

What to do this week

Write the sentence. One sentence, five links, for whatever you are building right now: these people, in this situation, get this outcome, which does this for their business, which does this for ours. If you cannot finish it, you have just found the real problem, and it is not your dashboard.

Then take the front half of it to one person who will be reading your metrics next quarter, and ask them to finish it. Do not bring a proposal. A proposal is something they have to approve or reject. Bring the question, and let them say the outcome part out loud.

Then look at the number you reported last month. If it went up and you cannot say what got better for a human being because it went up, you did not measure your product, you measured your rollout. Adoption only ever tells you that the door opened, and your job has always been what happens in the room.

-Brennan

Want to get your metric accepted instead of overruled? Join the next cohort of "The Influential PM: Strategy, Stakeholders, Significance."

Join the live course to become an Influential PM in three weeks.