There is a reliable pattern in how people work: whatever they measure is what they end up optimising for, whether or not they decided to. Put likes on the dashboard and, over months, the work drifts toward likeable. Put revenue on the dashboard and the work drifts toward profitable. The drift is not conscious, which is exactly why the choice of metric matters so much.
For most founders the metric was never chosen at all. It was inherited from whichever screen the platform opens on. That screen shows impressions, likes and followers, because those are the numbers the platform can see and the ones that keep you coming back. They are not the numbers your business runs on.
The metric becomes the goal
Watch how this plays out in practice. A founder notices that posts about the struggles of building get more engagement than posts about the problem their product solves. They write more of the former, because the feedback is immediate and pleasant. A year later they have an audience of fellow builders and a product nobody in that audience needs.
Nobody in that story made a bad decision. Each individual choice was rational given the information on the screen. The failure was upstream: the information on the screen measured attention, and attention was not the goal.
This is worth stating plainly because it explains a lot of otherwise confusing behaviour. Content that is popular and unprofitable is not a mystery. It is the predictable output of a feedback loop wired to the wrong signal.
Vanity, leading and lagging metrics
Three categories are worth separating. Vanity metrics move independently of business outcomes: follower count, impressions, likes. They are not useless, but they are not steering material. Leading indicators predict outcomes: clicks to your site, email signups, trial starts. Lagging indicators are the outcomes: purchases, revenue, retention.
Good practice is to steer by leading indicators day to day and validate against a lagging one monthly. Steering purely by revenue is too slow for weekly decisions, especially with long sales cycles. Steering purely by clicks is how you end up with excellent traffic and no sales. You need the pair.
The test for whether a metric belongs in your leading set is simple: has it ever, in your own data, moved ahead of revenue? If email signups rise and revenue follows a month later, signups are a real leading indicator for you. If impressions rise and nothing follows, impressions are decoration.
Why platform analytics steer you wrong
It is not malice. A social platform genuinely cannot see your revenue. It has no access to your payment processor, so it reports the only thing in its view, which is behaviour inside its own app. Every dashboard is limited to what its owner can observe.
Your payment processor has the opposite problem. Stripe knows exactly who paid and when, and knows nothing about which post they read three weeks earlier. Each half of the picture lives in a system that cannot see the other half.
That gap is why so few small companies can name the post that produced their last customer. It is not carelessness. The default tooling makes the question genuinely hard to answer, so people stop asking it and go back to counting likes.
How to pick your one number
Pick the metric that is closest to money while still being fast enough to act on. Work backwards from revenue until you reach something that moves within a week, and stop there.
- Selling a product with a trial: trial starts attributed to content, validated monthly against conversions.
- Selling a service: qualified enquiries, validated against closed deals.
- Selling through a newsletter: subscribers who match your buyer profile, validated against list-driven revenue.
- Selling nothing yet: replies that describe a real problem in the reader’s own words.
One number, made visible, beats a dashboard of twelve. A wall of metrics produces no behaviour change because no single figure is uncomfortable enough to act on.
Putting revenue per post in front of you
The setup is more tractable than it looks. Give every post a tracked link with a UTM that includes the post identifier. Record the identifier somewhere durable rather than trusting memory. Then join the two data sets on a regular cadence, either by hand in a spreadsheet or automatically.
You will not capture everything. Dark social is real: people copy links, screenshot posts and recommend you in private messages, and none of that carries a parameter. Treat the number as a floor rather than a complete picture, and add a single question at signup asking how someone heard about you. Between the two you get enough signal to steer by.
If you would rather not maintain the join yourself, that is precisely the connection seenpaid makes: it schedules and cross-posts to 21 networks and reads your Stripe account read-only, so each post arrives in your dashboard with an amount next to it rather than only a like count.
When the number disagrees with you
The point of measuring properly is to be corrected, and being corrected is uncomfortable. Expect to discover that a post you were proud of earned nothing, and that a plain, almost boring post about a narrow technical problem produced three customers. Almost everyone who runs this exercise finds at least one result they do not like.
Follow the data anyway, with one caveat: give it enough volume. A single post is an anecdote. Twenty posts across a couple of months is a pattern worth acting on. Do not rebuild your strategy around one week of results in either direction.
The habit to build is a monthly half-hour review. Sort by revenue, look at the top five and the bottom five, and write one sentence about what separates them. That sentence, repeated twelve times a year, is a more effective content strategy than any framework.
Keep one applause metric, deliberately
None of this means deleting engagement from your life. Engagement is a decent early-warning system for reach problems, and a sudden collapse in impressions across every post usually means something worth investigating: a flagged link, a shadow-limited account, a broken scheduler.
The rule is one of hierarchy rather than exclusion. Revenue and its leading indicator decide what you write. Engagement is a diagnostic you consult when something looks broken. Problems start when the diagnostic gets promoted to the goal, which happens by default because it is the number the platform shows first and refreshes fastest.
A simple structural fix is to make the commercial number the one you see first each week and the engagement number something you have to go and look up. Whichever number requires less effort to check is the one that will shape your behaviour. That ordering is the whole design principle behind seenpaid, where the amount a post earned sits next to it by default and the like count is secondary.
Whatever tooling you use, the principle stands on its own. The number you keep in view becomes the work you produce. Choose it on purpose, put it somewhere you cannot avoid it, and let it be the thing that decides what you write next.