A telehealth pharmacy owner said this to me almost word for word: my sales are climbing every month, but the ad platform shows almost nothing, so these figures just can't add up. The platform was reporting about twenty sales in two weeks against a business that was plainly growing. He had already gone and bought a second tracking tool to get an independent read, because he trusted his bank balance more than his dashboard. That instinct was right. The second tool was not the fix, and I want to explain why, because the reflex to buy your way out of a trust problem is exactly how people end up with three dashboards that disagree instead of one.

This is not a review of any product. It is about what has to be true before any marketing attribution software gives you a number worth acting on, and why an owner who is scaling can watch the reported figures stay flat and know, correctly, that something is wrong.

Why the Platform Undercounts a Growing Business

An ad platform counts what fires inside a browser. A pixel loads, an event sends, a conversion gets attributed. Everything the platform reports is built on that one visible moment, and a great deal of real business happens where that moment never fires. A customer clicks the ad on their phone, thinks about it for a week, then buys on a laptop after the cookie expired. Someone lands from the ad and completes the purchase over the phone. Privacy settings and consent choices block the event entirely on a growing share of traffic. None of those sales vanish from your bank account. They vanish from the platform's count.

So a business can be genuinely growing while the reported conversions stay flat or fall, and the gap is not a bug you can file. It is the difference between what a browser can observe and what actually got paid for. The owner who senses the numbers are too low to be real is not being paranoid. He is comparing the report against the one figure he does trust, and they do not match.

Why a Second Tracking Tool Disagrees Instead of Settling It

The natural move is to add a tool that promises better tracking, and it feels like getting a second opinion. What you actually get is a second number produced by a different method, last-click in one place and multi-touch attribution in the other, and now two systems disagree and neither can referee the other. I have watched a single business end up with three sources reporting three different sales counts for the same month, and every one of them internally consistent.

The disagreement usually traces to a few ordinary things, none of which a slicker interface fixes. A purchase event gets labeled as a lead in one tool and as a sale in another, so they are literally counting different things. A payment processor feeds two related brands into one tracking bucket, so one brand's numbers are quietly inflated by the other's. And the same order can land on two pixels under two different names. As one owner's own ad-tracking specialist put it to me, when the same sale lands on two pixels the platform can count one customer twice. Add a tool to that situation and you have not resolved the contradiction. You have paid for another version of it.

The Number That Can Settle It Is the One You Already Own

There is one source in the business that does not depend on a pixel firing: the record of what got paid for. Your orders, your payments, your completed appointments. That data is not an estimate of sales. It is sales. Attribution you can believe is built by starting there and working outward, reconciling the platform's reported conversions against your own paid-order record, rather than trusting the platform to grade its own homework.

Doing that well is unglamorous. It means agreeing on one definition of a sale and using it everywhere, so lead, booking, and paid order stop being used interchangeably. It means breaking the numbers down by the service the customer actually bought, not by the campaign name that happened to be attached. And it means keeping the brands and entities separate so one is never counted inside the other. Most marketing attribution software can help hold that picture once it exists, and none of it can decide those definitions for you. The definitions are the work. The tool is where the answer lives after the work is done, which is the reverse of the order most owners try.

Coverage Belongs Next to the Number

Even a reconciled number carries a second fact that dashboards almost never show: how much of your real sales it can actually see. If you can trace thirty per cent of paid orders back to a traceable source, then a cost-per-acquisition or a return-on-ad-spend figure calculated on that data is a correct calculation over thirty per cent of the business, printed in the same confident type as a calculation over all of it. Nothing in how the figure is displayed tells you which one you are looking at.

This is the trap that makes attribution dangerous rather than merely incomplete, and it deserves its own reading: a number that looks authoritative while resting on a third of the picture is worse than no number at all, because no number leaves your judgment intact and a wrong one replaces it with false precision. I have written that case out in full in the answer that was wrong and looked right. For an owner deciding where next month's budget goes, the honest version states its own coverage next to the answer, every time, so you know how much of your business the recommendation is actually steering.

Before You Move Budget on a Number

You can test whether a figure is safe to act on before you trust it with real money. For last month, answer these.

  • Can you tie your paid orders back to where they came from, or only your leads? Attribution that stops at the lead is measuring interest, not revenue.
  • Does the platform's total and your own paid-order total agree, and if not, do you know the size of the gap? An unexplained gap is the finding.
  • Is a sale defined the same way in every tool that reports one? If a booking counts as a sale in one place and not another, your numbers were never comparable.
  • Are two brands, entities, or product lines sharing a single pixel or tracking account? If so, at least one of them is being reported wrong.

If those come back shaky, the fix is not a fourth tool. It is one reconciled definition of a sale, built on the order and payment data you already own, with its coverage stated honestly. That is the same discipline behind everything I build: untangling the systems a business already runs into one source it can trust, which for a healthcare or telehealth operation is exactly what makes the administrative and data layer worth automating in the first place. Get the measurement right, and then let a tool answer questions on top of it. The other order produces confident advice resting on a third of the picture, and it produces it fast, which is the part that makes it expensive.