Vanity Metrics: Why the Numbers That Feel Good Are Not the Ones That Pay
What it actually looks like
You know your follower count to the exact number. You could not tell me last month's retention without opening something.
The launch went well — nine hundred signups in two days, a post that kept moving, four people you admire sharing it. You screenshotted the graph. Three weeks later you still have not checked how many of those nine hundred came back a second time, and the reason is not that it is hard to find out.
Look at your dashboard. The number your eye lands on first, top left, is the one that only goes up: cumulative signups, total registered users, impressions. It is the number that structurally cannot deliver bad news, which is precisely how it earned that slot.
When something you shipped gets attention you feel it in your chest, in a way that money arriving in the account does not quite produce. That is worth noticing rather than being embarrassed by.
You refresh. Not once — you check the same post four times in an afternoon, and the fourth check has no decision attached to it.
And when someone asks how the business is going, you reach for the growth number rather than the revenue number. Not to mislead them. It is genuinely the first thing in your hand.
Who this happens to
This is common among founders who build in public.
If your distribution runs through an audience, you are looking at engagement numbers all day for legitimate reasons. The problem is not that you see them; it is that seeing them daily is what makes them feel like the score. Attention becomes the water you swim in, and the slower measures — did they come back, did they pay, did they stay — happen somewhere you have to deliberately go and look.
It is common in solo founders and very small teams, because nobody else is holding a different number. In a larger company someone owns retention and will interrupt you with it. Alone, the only number in the room is the one you chose to display.
And it is especially common in people who are early. Real business metrics need volume to say anything at all — with forty customers, your churn number is noise. Engagement numbers work at any scale and give you an answer immediately. When the honest measures cannot speak yet, the flattering ones are the only ones talking.
What sets it off
Something going unexpectedly well. A post that outruns its usual reach, a spike in signups after a launch, a mention from someone with a big audience. The spike is genuinely information — it just is not the information you then act on for the next two weeks.
A quiet week on the real numbers. Revenue flat, nothing shipped that landed. This is exactly when the applause number becomes most attractive, because it is the one you can still move today.
Somebody else's screenshot. A competitor posting their user count, a peer's milestone in your feed. You now have a number to be behind on, and it is always the number that was easy to share — nobody screenshots their churn.
Investor or advisor conversations, where growth is the thing you are expected to arrive with. The pressure to have a rising line to show is real, and rising lines are easiest to produce where they cost the least.
And the deadest trigger of all: a slow afternoon. You open the dashboard with no question in mind. That check is not analysis. It is a mood adjustment with a spreadsheet attached.
Why it keeps happening
Because these numbers are fast, public, and unambiguous, and the ones that matter are none of those things.
Retention takes weeks to say anything. Revenue is lumpy and confounded. Whether the product is actually good is a question you can carry for months. Impressions resolve in ninety minutes. Given a fast, clear, flattering signal and a slow, murky, threatening one, attention goes to the fast one — not because you are shallow, but because that is how reinforcement works. Anything that reliably pays out on a short delay gets learned quickly and repeated without deciding to.
Then the second mechanism, which is more consequential than the first. What you look at daily is what you optimize toward, whether or not you ever choose to. When a measure becomes the target, it stops being a good measure — the standard version of this observation is Goodhart's law, and it applies to a founder watching a chart just as much as to an institution. Once your eye is on reach, you start writing the post that travels rather than the post your actual customers needed. Nobody makes that decision. It happens in the margins, in a hundred small choices about what to spend Tuesday on.
Underneath both sits the thing that keeps it in place: an anxious founder needs proof they are not wasting their life, and the applause numbers are the only ones offering proof on a daily schedule. Wanting that reassurance is entirely human. The cost is that it aims your effort at the scoreboard that is not keeping score.
Which leaves the arithmetic. A hundred thousand followers and no retention is a dying company with an audience. Two hundred customers who renew is a business. Both of those are real outcomes, and the number on your dashboard's top left cannot tell them apart — it goes up in either case, right up until the end.
What actually helps
Pick one number that can go down. That is the whole test. If your headline metric is structurally incapable of reporting bad news — cumulative anything, total anything — it is a trophy, not an instrument. Weekly active accounts can fall. Month-two retention can fall. Revenue can fall. Choose the one that best represents someone getting real value from what you built, and write it somewhere you will see it on a bad Thursday.
Demote the rest to the same screen, at the same size. You do not have to stop looking at reach — it is real distribution data. You have to stop letting it be the first thing you see. Move the applause numbers below the fold of your own dashboard. What you measure daily is where your effort drifts, so the layout is not cosmetic.
Check on a schedule, not on a feeling. Set one time a week to look at numbers, and hold to it. The urge to check on a slow afternoon is a mood, and answering it with data does not make it analysis. If you find yourself opening a dashboard with no question in mind, close it and write down the question first. Half the time there isn't one.
Convert one spike into one fact. After the next good post, do exactly one piece of follow-through: of the people who arrived from it, how many were still around fourteen days later? One number, ten minutes. This is the habit the whole pattern is missing — not less celebration, just one honest question asked afterward, before the spike gets filed as proof.
Say the real number out loud once a week. To a peer, a co-founder, anyone. Not the reach figure — the one you would rather not lead with. Numbers you say to another person stop being decorative, and the ones you never say are the ones that quietly drift.