Creative-Level Reporting: Measure the Video, Not the Ad

Your ad account is organised around the buy — campaign, ad set, ad. The creative is an attribute of an ad, not a thing the reporting counts on its own. So when a single creator video is re-cut into three lengths and run against four audiences, you get twelve rows and no answer to the only question that shapes your next brief: what did the video do? Creative-level reporting fixes that by collapsing every ad that carried an asset back into the asset.

This matters more in digital than anywhere else. Modelling nearly 500 campaigns, Nielsen Catalina Solutions found that creative accounted for 56% of the sales lift from digital advertising — roughly double the contribution of the media itself. If more than half your return rides on the asset, and your reporting cannot isolate the asset, you are optimising the smaller half.

Because the unit of reporting is the ad, and the ad is not the creative. One asset becomes many ads the moment it enters a real account: a 30-second cut and a 15-second cut, a square and a vertical, three audiences, two placements. Each combination reports separately, and none of those rows is the video.

The workaround most teams reach for — averaging the rows — makes it worse. An asset that ran against a cold prospecting audience and an asset that ran against a retargeting pool are not comparable, and averaging them produces a number that describes the media plan rather than the creative. The asset that looks weakest is often the one that drew the hardest audience.

This is the same failure that makes creator programmes look unprofitable in aggregate. As we have argued before, the usual explanation for disappointing influencer ROI is not the creator but the measurement around them.

Yes, and the gap between channels is the interesting part. Nielsen Catalina Solutions modelled the drivers of sales lift across nearly 500 campaigns spanning all major media platforms, and presented the results at the Advertising Research Foundation’s 2017 Audience Measurement conference. Across all media, creative contributed 47% of sales lift — more than reach, targeting, recency and context combined. Split by channel, the picture diverges sharply:

On TV, the media plan does the heavy lifting. In digital it is the other way round: the creative carries 56% and the media 30%. Creator content lives entirely on the digital side of that table. The asset is the largest single lever you have, and it is the one your reporting is least equipped to see.

Most reporting mixes the two without labelling them, which is how a creative decision ends up being made on an auction metric. The split is cleaner than it looks:

Meta’s own definitions are precise about the first two — 3-second video plays and ThruPlay —, and precision is the point: a 3-second play excludes replays, and a ThruPlay treats 97% of a video as its full length because people drop off as content fades out. Those are asset facts. CPM and frequency are not — they are what the auction charged and how hard the budget pushed, and neither moves because a creator opened on a better line.

Because they are the part of the funnel the creator controls outright. Everything that happens after the click is shared with the targeting, the landing page and the offer. What happens in the opening seconds is the asset, almost alone.

That makes the ratio worth building by hand: 3-second video plays divided by impressions gives you the share of people served who stopped long enough to register the video, and ThruPlays divided by 3-second plays gives you the share of those who stayed. Neither is a platform metric you can switch on — both are arithmetic over two numbers your account already reports, computed per asset rather than per ad.

Judge a creator on those two and you are judging their work. Judge them on CPM and you are judging your media buyer. That distinction is what a paid-ready brief is for, and it is why briefing for paid differs from briefing for organic — you are asking for a specific behaviour in a specific window, not for a nice video. It is also the earliest signal that an asset is tiring, which is the practical half of diagnosing creative fatigue: hold rate usually decays before CPA does.

You count it twice, and the blended number flatters everything. A creator post that earned 400,000 organic views and was then promoted does not have one view count — it has an organic figure, a paid figure, and a total that is only meaningful if you know the other two. Report the total alone and paid spend quietly takes credit for reach the creator earned for free.

The fix is structural rather than clever: keep organic, paid and total as three separate numbers on the same row, for the same asset, over the same date range. Social Native reports all three for the same campaign, with every metric card stating its own formula. Getting earned and paid onto one scorecard is the point several of these threads converge on, and we have made the wider case for it already.

This also has a prerequisite most teams discover late. An asset can only be compared across organic and paid if it is legally able to run in both, which is settled long before the reporting question arrives — rights to the audio, a live permission grant, an objective chosen before the creative was made. We covered what a post needs before it can run as an ad in a separate piece. Where the paperwork is right, the prize is measurable: Meta reports that adding partnership ads to a campaign delivers 19% lower CPAs and 13% higher click-through rates on average, figures reported from its own press materials.

Two routes, and they are not equivalent.

1. A naming convention. Put a stable asset identifier in every ad name and parse it back out at report time. It costs nothing to start and works until someone renames an ad, launches in a hurry, or joins the team without reading the document. Treat it as a way to begin, not a system.
2. A real asset identifier. Carry an ID with the file from the moment it is approved, so every ad that uses it inherits the reference and grouping survives renames. This is more work up front and it is the version that still functions in a year.

Either way the output you are building toward is the same: one row per asset, every ad that carried it collapsed into it, organic and paid kept apart, and hook and hold rates computed per asset. Once that report exists, the library becomes auditable — which is the groundwork for working out which assets are actually earning rather than which ones were expensive.

Approved assets in Social Native export directly into your own Meta Ads Manager or TikTok Ads Manager media library, and promoted creator posts carry an objective and destination you choose. Both are worth doing with the asset identifier already attached, because retrofitting one across a live account is considerably harder than starting with it.

1. Pick the identifier before the next flight. Naming convention or asset ID — decide which, and apply it to everything that launches from here. Backfilling is the expensive version.
2. Add hook and hold rate to the creative review. 3-second plays over impressions, ThruPlays over 3-second plays, per asset. They are the two numbers a creator can actually act on.
3. Split organic and paid on every report. If a single blended view count is the only number on the page, you cannot tell earned reach from bought reach, and neither can anyone reading it.

None of this requires new tooling. It requires deciding that the asset is the unit you measure, and then refusing to report at any other level.