I have worked inside more than one business where the board set a direction on the Monday, and by the Friday the numbers behind it did not hold. Not fraud. Just figures that fluctuated, would not reconcile, and fell over the moment someone senior actually looked. The uncomfortable truth about marketing reporting is that a great deal of it is not data at all. It is an opinion, with a chart on it.
A report is a story, and everyone telling it has a reason to tell it well
The person who assembles the weekly report is rarely the person who should. Manual reporting is treated as admin and pushed down the team, so the figures a board will bet on are populated by the most junior hands, with too little context to know when one looks wrong. That is not their failing. It is a broken delegation inside a stack of incentives that all point one way. A manager is judged on the line going up. An agency is paid to keep the account, justify the fee, and earns more when the numbers argue for more spend. Ask anyone to grade their own homework and they will give themselves a good mark, not from dishonesty, but because the structure rewards a rising story over an accurate one. Handed two versions of the same figure, few stop to diagnose the gap; they take the higher one, which buries the clue that something is broken and tilts every report upward. So the loss is left off the slide, the awkward figure is smoothed, and the number that does not fit the narrative quietly disappears. That last one is the real cost: the figure that does not fit is very often the one that mattered, the early warning, and massaging it away deletes your own smoke detector.
Before anyone spins it, the raw numbers often do not add up
Even with the best will in the room, the figures are frequently wrong before a single slide is built. Add up what Google, Meta and your email platform each claim they made you, and it comes to more than the business made, because every platform takes full credit for the same sale and nobody reconciles the overlap. Count "users" and you are usually counting devices, so one person on a phone, a laptop and a work machine is three customers who never existed. Attribution, the model that decides which channel gets the credit, is a choice rather than a measurement: change it and the same month turns paid media from hero to passenger. The conversion window, how long after a click, or even a view, a sale still counts, is another quiet dial, often set differently on every account, so two that look different may only have different settings. None of it needs bad intent, only that no one checked what the numbers are really counting.
The worst version counts your failures as wins
Inflation is not the worst of it. A conversion is only as honest as the moment that fires it, and unchecked, those moments drift. A form viewed rather than submitted. A thank-you page that can be refreshed. Your own office and agency filling in test enquiries. And in the worst cases I have seen, an error page or a declined payment still firing the conversion, so the system logs a failure as a success, celebrating the customer who tried to buy and could not. It survives for years because the number stays plausible. The only cure is to reconcile from the ground up: take the conversions the platform claims, count what actually completed in your own order and revenue systems, and study the gap. That gap is where errors, refreshes, test clicks and your own staff got counted as growth. Treat your tracking the way a security team treats a system: assume it will be gamed, tighten what counts as a real outcome, exclude your own locations, so you do not pay to convert yourself.
Blaming the market is an excuse until you have ruled everything else out
When performance dips, the convenient explanation is waiting: it is the market. Impressions up, clicks down, rising costs, more competition, a platform change. It sounds authoritative and lets everyone off the hook. An explanation you have not tried to disprove is not a finding, it is an excuse. The market moving at the same time as your numbers is a coincidence until you rule out the things you can actually fix: a change someone made, a broken tag, tired creative, a slower page, or the tracking quietly counting errors from months ago. The discipline is to attack your own conclusion before you present it, ruling the alternatives out one by one. Reporting stops at the first story that lets the meeting move on.
The standard has quietly moved, and most reporting has not
What good looks like now is not a tidier monthly deck. A monthly report is history: a trend that starts in week one is invisible until week five, by then a crisis or a closed window. Live measurement catches the shift while there is still time to act, and frees the team from the manual grind eating the hours meant for strategy. Tracking each step and each field, where people give up and which question kills the form, turns a mystery into a diagnosis you can act on. Measure everything for that diagnosis, and count only the real outcome as the result. When the data underneath is clean, it does more than tell an honest story: it holds every channel to a single definition of profit, so paid, organic and the rest are judged on the same honest measure instead of each on the one that flatters it. It sharpens the analysis, firms the accuracy, and feeds only your real, closed customers back to the ad platforms, so the bidding learns from genuine business rather than the errors.
The common thread is simple. If the number changes with the person who pulled it, you do not have data, you have an opinion, and you are paying senior people to build slides instead of make decisions. The remedy is to take the pen out of the hand of everyone with a reason to flatter the result: one live source of truth, built on logic agreed up front and locked, wired deep enough into your own systems that it reconciles to reality, holds every channel to one definition of profit, and is hardened so no channel can game it. That is what we built Crane Intelligence™ to be, and it is fast becoming the line between businesses that can see clearly and those still deciding on numbers massaged before they landed. The reports will always tell a good story. The question worth asking is whether anyone in the room could prove it.
