Most performance reporting is a map of the things that are easy to count. Spend by campaign. Impressions by placement. CPM by audience. Every one of those is a media variable, and media variables are what ad platforms hand you for free.
The inconvenient part is that media variables are not where most of the outcome comes from.
The full split from that analysis:
Two things stand out. Creative is the single biggest lever by a wide margin. And targeting — the thing that absorbed a decade of industry attention, tooling and budget — comes fourth, at 11%.
This is CPG. It is US-centric, it is largely offline sales, and it is a decomposition of measured campaigns rather than a controlled experiment. Anyone quoting it as a universal law of digital DTC is overreaching. Treat it as a strong directional finding from a large sample, not as a constant you can plug into a model.
Why the gap persists
It is not that teams don't believe creative matters. Ask any media buyer and they'll tell you it's the biggest lever they have. The gap is that creative is the hardest thing in the stack to turn into data.
Spend is a number. A hook is not. To analyse creative the way you analyse media, something has to convert "a founder talking to camera in a kitchen, fast cuts, offer in the first two seconds" into structured fields you can group and compare. Most teams solve this with naming conventions — and naming conventions decay. They decay when a freelancer joins, when an agency hands the account over, when someone is shipping twelve variants on a Friday.
So the creative dimension quietly drops out of reporting, and what's left is a very precise account of the 51%.
What the 49% actually buys you
The practical consequence is about where marginal effort goes. If creative carries roughly half the outcome, then:
- A 10% improvement in creative quality is worth more than a 10% improvement in targeting precision, and it's usually cheaper to get.
- Knowing which creative attribute is working — the angle, the format, the person on screen — is worth more than knowing which audience segment converted, because the attribute is the thing you can brief again.
- A reporting stack that can't group by creative attribute can't tell you what to make next. It can only tell you what to turn off.
That last one is the real cost. Media reporting is good at subtraction and bad at addition. It will tell you an ad is underperforming. It will not tell you that every one of your winners this quarter was shot in a real location and every loser was shot in a studio.
Knack tags creative from the creative itself — the actual image and video frames — rather than from the ad name, so the analysis survives an account inherited from another agency. Every ad gets structured attributes: format, angle, hook, talent, setting, pacing, tone, CTA, who it speaks to, and whether it's owned or a partnership. The Patterns view then groups performance by those attributes, which is what turns "creative is 49%" from a statistic into a brief.
The honest limit
None of this makes creative analysis a substitute for incrementality testing. Platform-reported conversions are directional and they over-credit themselves; a creative that looks like it's driving revenue in Ads Manager may be harvesting demand that was already there. Creative analysis tells you which ads are doing the work relative to each other. It does not tell you whether the channel is profitable. Those are different questions and they need different instruments.
What creative analysis does do is stop you spending the 49% blind.
Sources
- NCSolutions, sales-driver analysis of nearly 450 CPG campaigns — creative 49%, brand 21%, reach 14%, targeting 11%, recency 5%. Reported by MarketingCharts.
- Nielsen, "When it Comes to Advertising Effectiveness, What is Key?" — the earlier study in the same series, which put creative at 47%.