9 Jun 2026 / Essay
The one number you can't buy.
Open any content team's dashboard and you are looking at a marketplace. Impressions have a price. Reach has a price. Clicks, followers, views, the broad category of motion that gets reported as engagement — each of these can be acquired for money, and the rate is known to anyone who has ever run a campaign. You do not have to earn them in the sense that matters. You have to fund them. A line on a dashboard that climbs after you increase a budget is not telling you that the work got better. It is telling you that the work got more expensive.
This is the quiet fact underneath most content strategy, and it is not said out loud. The numbers a brand reports on are, almost without exception, numbers a brand can buy. That is precisely why they are reported. They move when you spend, they move in the direction everyone wants, and they move on a schedule that fits a quarterly review. A metric that behaves like that is comfortable. It is also, for the same reason, close to meaningless. Impression-led content optimizes for what is cheap to count, not what is hard to earn — and what is cheap to count is cheap to read.
So before asking which numbers a team should chase, it is worth asking a sharper question. Which of these numbers could not be bought, no matter the budget?
Start with what a buyable metric does to the work over time. A number you can purchase is a number you will, sooner or later, optimize by purchasing. This is not a moral failing. It is the path of least resistance, and organizations follow it the way water follows a slope. If the goal is impressions and impressions can be bought, the most reliable way to produce more impressions is to spend more, not to write better. The editorial question — is this good enough that someone would choose to read it? — becomes optional. It can be skipped entirely and the dashboard will never notice.
What gets produced under that incentive is content engineered to be counted, not content engineered to be finished. It is built for the metric, and the metric rewards arrival, not completion. A headline that earns the click has done its whole job the instant the click registers; what happens after — whether the reader stays, whether they reach the end, whether they remember any of it tomorrow — is invisible to the number being optimized. Volume metrics select for content that is forgettable on purpose, because forgettable is cheaper to make and the system pays out the same either way.
There is a second-order cost, and it is worse than the first. A team that optimizes a buyable metric long enough stops practicing the harder skill. Earning a reader — holding someone through a difficult middle, closing in a way they remember the next day — is a craft, and crafts decay without use. The buyable metric never calls for it. So the dashboard stays green while the capacity to make something worth finishing erodes, because nothing in the measurement ever required that capacity to exist. By the time anyone notices, the team has years of healthy charts and no muscle left for the one thing that was ever going to matter.
This is how a content operation can run for years, post relentlessly, report growth every quarter, and produce nothing anyone would miss if it stopped. The dashboard was green the entire time. The numbers were real. They were numbers that money moves, and money was moving them.
There is one exception. Completion is the only reader-side metric that cannot be inflated by spend.
You can pay to put a piece of writing in front of a million people. There is a price for that, and you can pay it today. You cannot pay any of those million people to reach the end. Finishing is not a transaction you can enter on their behalf. It is a decision they make, sentence by sentence, and the only currency that buys it is the work being good enough to earn the next line. No media budget reaches that decision. No amount of distribution spend converts a skimmed first paragraph into a finished essay. The reader either keeps going or they don't, and what determines it is on the page.
That is what makes completion different in kind, not just in degree, from every other number on the dashboard. Impressions measure how much you spent on placement. Reach measures the size of the room you rented. Engagement, in most of its reported forms, measures how well the opening was engineered to provoke a reflex. Completion measures something none of those touch: whether the thing was worth a reader's time all the way through. It is the one signal that reports back on the quality of the work itself, because it is the one signal the work itself has to produce.
"The only" is not a flourish. It is the claim. Soften it to "one of the harder metrics to buy" and the argument dissolves into preference — just another team explaining why its favorite number is special. Held exactly, it says something structural about the attention market: there is a single reader-side measurement that capital cannot reach, and it happens to be the one that correlates with the work being good. That is not a coincidence to optimize around. It is the place to stand.
It helps to be precise about why finishing resists money, because the resistance is mechanical, not mystical. Every buyable metric shares a property: it can be satisfied by an input the buyer controls. Impressions are satisfied by ad spend. Reach is satisfied by audience acquisition. Even engagement, in its cheaper forms, is satisfied by formats engineered to trip a response. In each case the brand supplies what the metric needs.
Completion requires an input the brand does not control and cannot supply: the reader's sustained attention, given voluntarily, held to the end. There is no budget line for that. You cannot procure it, cannot accelerate it, cannot make up for its absence with more of something else. The reader's continued attention is the one resource in the entire funnel that money cannot stand in for, because it is the reader's, not yours, and they spend it only when the work keeps earning it.
This inverts the usual relationship between budget and result. For buyable metrics, more spend produces more number; the relationship is close to linear and that is exactly why those metrics feel safe. For completion, spend does nothing past the threshold of getting the piece in front of someone. From there the curve is governed entirely by the writing. A larger budget puts a weak essay in front of more people who will abandon it at the same rate. The money buys a bigger audience for the same unfinished read. Past the point of placement, completion is indifferent to how much you spent — and a metric indifferent to spend is the only kind that tells you the truth about the work.
This is the difference between a vanity metric and a fitness metric, and the line between them is exactly the line between what you can buy and what you cannot. Volume is a vanity metric. Finish-rate is a fitness metric.
A vanity metric makes the team feel productive and the quarter look healthy. It rises with effort and budget, it is easy to present, and it asks nothing hard of the work. A fitness metric measures whether the thing is actually fit for its purpose — whether it does the job it exists to do. For editorial, the job is to be read, all the way through, by someone who chose to. Finish-rate measures precisely that and nothing else. It cannot be padded, cannot be bought, cannot be made to look better than the work deserves.
Which is also why finish-rate is uncomfortable, and why most operations avoid building around it. A number you cannot buy is a number you cannot rescue with budget when the work is weak. It will tell you, without appeal, when a piece did not earn its readers. That is not a flaw in the metric. That is the entire reason to trust it. The numbers that flatter you are for sale. The number that can correct you is the one no one can purchase.
So we built the operation around it, on purpose, before we built anything else. We measure completion, not impressions. The gate is 65% finish on long-form editorial, median per piece, four-week rolling, sustained two months.
That sentence is a standard, not a scoreboard. It does not say our readers finish at any particular rate; it says what bar a body of work has to clear before we will call the method proven. The number is deliberate in every part. Sixty-five percent of a long piece is a real finish, not a polite scroll. Median, not mean, so a single viral outlier cannot carry a quarter of mediocre work. Four-week rolling, so no single strong week flatters the picture. Sustained two months, so the bar is met by a pattern and not by an accident. Every clause exists to make the standard harder to fake — which is the point of choosing a number that cannot be bought in the first place. There would be no sense in picking an unbuyable metric and then defining it loosely enough to game.
The specific number is a choice we can defend. A lower bar would clear too easily and prove nothing; a piece can hold a third of its readers without being good, on novelty or headline alone. A much higher bar would select for short, safe work that finishes because it never risked losing anyone — which is its own way of failing the thesis. Sixty-five percent of a genuinely long piece sits where it has to be earned by the writing and cannot be reached by trimming the ambition. It is high enough that money cannot rescue a weak essay into clearing it, and reachable enough that strong work clears it honestly. The bar is set where only the work can move it.
The gate sits before the work ships, not after. It is not a report we read at the end of a quarter to find out how we did. It is the condition the work has to satisfy to count as the thing we claim to make. A metric you cannot buy only matters if you are willing to be held to it, and being held to it means letting it gate the work rather than merely describe it.
Which raises the obvious question, and we answer it the only way that is honest: we hold ourselves to it first. Pagecut is our own first client. The publication is the demo. We are operators, not vendors.
This is not a positioning line. It is the structure of how we work. We do not arrive with a deck about completion and a finish-rate standard we have never personally stood under. We run our own publication against the same gate we describe, on the same cadence, under the same bar, and the results of that — the real ones, not the flattering ones — are the evidence. The argument of this essay is one we are required to live inside before we are allowed to make it to anyone else. An agency that sells a standard it has not met itself is selling a metric it has bought: acquired through assertion rather than earned through work.
Holding the unbuyable number yourself is the only way to prove you mean it. Anyone can recommend completion as a metric. Recommending it costs nothing. Submitting your own work to it, in public, on a fixed rhythm, with no budget that can rescue a weak piece — that is the part that cannot be faked, for the same reason completion itself cannot be faked. The standard and the practice are the same shape. You cannot buy your way past either.
Which returns us to where this began, and to the diagnosis underneath all of it: the publish button was never the finish line. The internet rewards what people finish, not what brands publish. Publishing is an input you control and can therefore buy your way to more of. Finishing is the reader's verdict, and it is the one verdict no budget can overturn.
Almost every number a content team optimizes is a number that money moves. They climb when you spend and they tell you very little, because anything you can purchase your way to was never a signal in the first place — it was a receipt. Completion is the exception, and the exception is not incidental. A metric that resists money is the only one that reports on the work instead of the wallet. It is the number you cannot buy, which is exactly what makes it the only one worth chasing — and the reason we keep our surfaces, and our standards, rare on purpose.
The rest of the dashboard is for sale. That one number is not. Build for the one that isn't.