How to Read Your Social Media Revenue Numbers

Setting up revenue tracking is the easy half. You put tracked links on posts, connect a payment provider, and a number appears next to each post.

The hard half is knowing what that number means. Two posts show $245 and $98, and before you conclude anything you need to know which attribution model produced those figures, whether refunds are in them, how long the system waited before deciding, and how much of the real revenue never appeared at all.

This post is about the judging, not the wiring. If you need the wiring first, connecting posts to revenue has the Stripe and Shopify setup, and how to use UTM links covers naming the links themselves.

What a post-level revenue number actually claims

When a tool says a post produced $245, it is making a narrow claim: someone clicked a link belonging to that post, and within some window they paid you $245.

It is not claiming the post caused the purchase. It is not claiming that is all the revenue the post produced. It is a record of a traceable path, and traceable paths are a subset of real ones.

That sounds like a caveat and it is actually the most useful thing to internalise, because it tells you what the number is good for. It is good for comparing posts against each other, where the same undercount applies to all of them. It is bad as an absolute figure you put in a board deck without a footnote.

The two attribution models worth starting with

You will hear about first-touch, last-touch, multi-touch, view-through and incrementality. Two of them will carry you a long way.

The two disagree, and the disagreement is the interesting part. A post that scores high on first touch and low on last touch is doing reach work. A post that scores the other way is doing closing work. Both jobs are real, and a calendar made entirely of one or the other underperforms.

Most tools, including mine, default to last click over a fixed window. That default quietly favours the closing posts, which is worth remembering before you delete a discovery post for producing no revenue.

The window, and why 30 days is a guess

Attribution matches a payment to a click only if the payment happens inside a window. Thirty days is the common default, including ours.

Thirty days is not a fact about buyers. It is a compromise: long enough to catch considered purchases, short enough that a click in January does not claim a sale in June. Under it, an impulse buy is caught easily and a six-week enterprise decision is not caught at all.

If your product has a long sales cycle, know that your tracked revenue systematically understates the slow deals, which are often the large ones.

Decide what counts as revenue before you report it

This one causes more arguments than any model choice, because two people can look at the same week and produce different totals honestly.

Revenue could mean gross sales, net after refunds, first payment only, monthly recurring revenue, annual contract value, affiliate commission, or pipeline value. Pick one, write it down, and put the definition next to the number.

Revenue = successful payments, net of refunds, first payment only

Two specifics are worth deciding early. Netting refunds off is the honest choice, and it means a good week can go negative a month later, so warn people in advance rather than when it happens.

The other is recurring revenue. If a post brings in a subscriber who pays $19.99 a month for two years, is that post worth $19.99 or $480? Both answers are defensible and only one of them can go in the column, so pick before the number matters to anyone.

If you report to a client, agree the definition before the campaign, not when the invoice is questioned.

How long to wait before calling a winner

Judging too early is the most common mistake I see, and it is expensive because it kills formats that were working.

Some posts convert the same hour. A founder story builds trust today and shows up as revenue three weeks later, by which point most people have already decided the format does not sell. Give a post at least two weeks before you rank it, and longer if your product costs more than circa $100.

The related trap is ranking on a single post. One post can get lucky, land in a good slot, or catch an algorithm. Three posts on the same topic converting at a similar rate is a pattern. One post converting well is an anecdote you should test again before rebuilding the calendar around it.

Every total is a floor

So compare rather than total. Post A against post B is a reliable comparison, because both are undercounted by roughly the same factor. "Social produced $1,200 last month" is a floor with an unknown gap above it, and the gap is not stable enough to estimate.

The same logic makes the trend more trustworthy than the level. If tracked revenue doubles month over month, the doubling is real even though both figures are too low.

Expect the size of the undercount to vary by platform, too. Instagram loses far more than LinkedIn, because Instagram will not let you put a clickable link in a feed caption. That is a platform design decision showing up in your report rather than a problem with your content, and I went into it specifically in how to track sales from Instagram.

A report that says "at least $1,200, and here is what we could not see" is more credible than one that implies precision it does not have. Clients notice the difference, usually in your favour.

Reading a real week

Five posts for one launch, all pointing at the same landing page, each with its own tracked link:

The obvious read is that the tactical LinkedIn post won and the YouTube Short failed. Both conclusions are probably wrong on their own.

The Reel got 420 clicks and converted at 0.5 percent, which means it did reach work well and selling work badly. That is a reason to keep making Reels for the top of the funnel, not a reason to stop. The tactical post converted at 5.2 percent on a tenth of the traffic, so it sells to people who already know you, and it will not grow the audience on its own.

The Short at zero is the one you cannot judge yet. It has 74 clicks, which is too small a sample to mean anything, and video often converts late. Run it three more times before deciding.

The useful output of this table is not a ranking. It is next week's mix: more Reels for reach, more tactical posts near the offer, and a proper test of Short-form before writing it off.

The weekly review, in about ten minutes

1. Which posts produced revenue, and do they have a topic in common?
2. Which posts produced clicks but no revenue, and is that a landing page problem or an audience problem?
3. Which posts produced neither, and did they even have a link?
4. What changes in next week's calendar because of the above?

Question four is the only one that matters. A revenue report that never changes what you publish is a report you can stop running.

Where the number should live

The report is most useful next to the calendar, because that is where the decision gets made. Opening an analytics tool, building a report, filtering campaigns and matching events is a good way to ensure the review happens twice and then stops.

That is what I am building planpo.st around: the post, its tracked link, its clicks and its attributed revenue in one place, so the weekly review is a page you look at rather than a task you schedule.

The broader picture of how the attribution works, and what it does and does not claim, is on the revenue attribution page.