Before You Credit the Campaign, Check the Category

The case study in front of you reports a strong year over year gain, and the window is nagging at you. You half remember that stretch as a good one for the whole category. Here is the comparison the report left out, and a way to put a number on it.

The comparison hiding inside a year-over-year number

“Up 40 percent year over year” is a comparison against exactly one thing: this brand, twelve months earlier. That is the entire contents of the claim. It says nothing about what any other account did over the same period, because no other account was measured. The report is not withholding a second number. It never had one.

Which is the situation that made you suspicious. If every comparable brand also finished the year higher, some share of this brand’s 40 percent was going to happen with no campaign at all. Name the absent comparison: what did the untouched baseline do, the brands this campaign never reached?

Category growth is not the same animal as a seasonal spike

Category growth in the sense used here is secular: durable, not pinned to the calendar, and it does not reverse when a date passes. A platform’s user base expanding for several years running is category growth. So is a product category absorbing new demand across many quarters. No month arrives at which it switches off.

That makes it a different mechanism from a periodic, calendar-based tailwind, which recurs on a known window and is gone once the window closes. A reader who checks only whether the campaign ran near a holiday is running a date test, and a date test finds nothing when the tide has been rising for six straight quarters. A campaign can ride an eighteen month rise, launch in an unremarkable March, and pass that check.

A hypothetical worked example: netting one rate out of the other

What follows is a hypothetical. The numbers are invented for illustration, belong to no agency and no case study in this library, and none is a real reported result.

Picture a case study reporting a brand’s followers up 34 percent over twelve months. Strong on its face. Now suppose the platform it competes on grew 29 percent over the same months. Subtract one rate from the other and the excess is about 5 percentage points, a small fraction of the headline.

A second version. The brand reports 18 percent, which still reads as a win in a slide deck, while the category grew 26 percent. The net figure is negative 8 points: the brand went up and lost ground at once, and the reported number would never say so.

The limitation lands immediately. This subtraction is a plausibility estimate, not a rigorous decomposition, because category growth is rarely measured on the same basis as a brand’s own figure. It tells you roughly whether a headline is mostly campaign or mostly tide, and produces no defensible attribution number.

Where a reader can check what the category was actually doing

You rarely need insider data. Four places to look, in descending order of directness:

• The platform’s own public disclosures. Platforms reporting user counts, active accounts or revenue on a schedule give a growth rate for the surface the campaign ran on. Pull the window the case study used, not the nearest annual figure.
• Independent industry-trend reporting on the category. Coverage describing the category rather than a single brand, from a source with no stake in the campaign. A direction, expanding or flat or contracting, is enough.
• Several peer brands’ published numbers together. Where no formal category figure exists, three or four comparable brands reporting growth for the same window make a crude substitute. Building a rate from several points differs from adopting one as a target, because someone else’s case study result is not your benchmark.
• Search or public interest trend data. The weakest, for when nothing more direct exists. Interest is not demand, but a curve that clearly rose across the category is more than the case study gave you.

What netting out does not prove

A positive net figure is not proof of causation. Subtracting category growth removes one alternative explanation and leaves the rest standing: a price change, a distribution change, a product launch, a strategy shift unrelated to the campaign.

The honest version of that isolation is a real counterfactual, an untouched comparison group measured over the same window. That is the number no report can show you, for structural reasons, as the counterfactual no case study can show you sets out. Netting out is the cheap version you can run from outside, and weaker for it. It adjusts confidence rather than settling anything: a net figure that survives a generous category estimate raises it, one that collapses toward zero or goes negative lowers it sharply.

Signs a case study skipped this step

Three tells, visible without outside data:

• A year over year number with no mention of the category. The report says what the brand did and stops. Nothing about the platform, the industry or comparable accounts over the same months, not even to dismiss it.
• The period described only in relative terms. “Compared to last year” names a comparison without conceding that the whole space may have moved between its two ends.
• The comparison group changes partway through. The report opens against the whole industry, narrows to a vertical, then closes against a few named peers. Each framing may be defensible alone. Switched inside one document, they make netting out impossible.

Try the subtraction on a real one

The method gets sharp with practice. Browse the case study library, find a result reported year over year, and write down what you think the category did over that window before reading the conclusion. If your own work already states what the category was doing, submit your own.

FAQ

Is there an exact formula for netting out category growth?

No. Subtracting an estimated category rate from the brand’s reported rate gives a plausibility estimate, not a rigorous decomposition, because the two figures are rarely measured on the same basis or with the same precision. Read it as an order of magnitude rather than a measured quantity.

What if I cannot find any category growth data at all?

Treat the claim as unverified rather than assuming the category was flat. A missing comparison is not evidence that it would have favoured the campaign, and defaulting to zero category growth is an assumption dressed as neutrality. Unchecked is a different verdict from wrong.

Sources

• The campaign launched in December, and so did everyone’s sales, fetched 2026-09-06.
• The counterfactual: the number no case study can show you, fetched 2026-09-06.
• Someone else’s case study result is not your benchmark, fetched 2026-09-06.