You can rank every creative in your account by click-through rate and cost per install. Ask which creative actually drove paying users, though, and the trail goes cold at the install. The ad platform stops reporting there, your revenue lives somewhere else, and the two never get joined by creative. So teams optimise their most expensive lever, the creative, on a metric that stops at the click.
That gap is what creative-level ROAS is meant to close. This is a practical guide to attributing install and post-install revenue back to the individual ad, so the number you scale on reflects real payback rather than engagement.
What Creative-Level ROAS Actually Means
Creative-level ROAS is the return on ad spend calculated for an individual ad creative, measured by attributing post-install revenue back to the specific creative that drove each install rather than to the campaign.
It is a different measurement from the numbers most teams already have:
- CTR and IPM are engagement proxies. They tell you a creative earns attention, not that the attention converts to revenue.
- Campaign ROAS blends creatives together. A campaign can look healthy while one creative subsidises three that lose money.
- Creative-level ROAS keys on the creative ID. It follows one chain: impression, click, install, then revenue, all tied to the same creative identifier.
The measurement only works if that creative ID survives the whole chain. That is exactly where most setups break.
Why Creative Data Goes Dark After the Install
Ad platforms report richly up to the install and then stop. Everything that makes a creative worth scaling, the revenue, happens after the platform's visibility ends.
- The platform owns the pre-install view. Impressions, clicks and installs are reported per creative inside Meta, Google or TikTok.
- Revenue lives in your app and your MMP. Purchases and subscriptions are recorded post-install, in a different system, keyed by user rather than by creative.
- The creative ID is often dropped at the boundary. If the identifier is not passed through the SDK and postback layer, revenue gets attributed to the campaign, rarely to the creative that earned it.
The result is a reporting layer that can tell you which campaign made money but not which ad did.
How to Preserve the Creative ID Through to Revenue
Fixing this is mostly a plumbing problem, not an analytics one. The job is to carry the creative or ad-set identifier all the way into the attribution layer so revenue can be mapped back to it.
- Pass creative and ad-set identifiers into your attribution setup. Structured campaign naming and network-supplied creative IDs both work, provided they land in the MMP, not just the ad account.
- Map revenue events back to the creative that drove the install. Once post-install event tracking is live, each revenue event should resolve to the install, and the install to its creative.
- Get the postback configuration right. Misconfigured postbacks are a common reason revenue signals never reach the creative view. The complete postback setup guide for Meta, Google and TikTok walks through the mapping, and the underlying postback mechanics are worth understanding before you rely on the numbers.
Do these three and revenue stops being a campaign-level total and starts being a creative-level cut.
Building the Creative-to-Revenue View
Creative-level ROAS sits on two layers that do different jobs. Keeping them separate is what makes the view trustworthy.
- The creative-intelligence layer tags and scores the creative itself: hook, format, angle, the visual elements that drive performance. Dedicated element-level creative-analytics tools such as CreativeX, Foreplay, Motion and Segwise specialise in this, breaking a creative into its components and scoring them.
- The attribution layer is the revenue truth. It keeps the creative identifier attached from click through to revenue, so ROAS is measured on payback rather than clicks. This is the role an MMP like Linkrunner plays.
The minimum dashboard cut to run this weekly is small: creative ID, spend, installs, D7 revenue and the resulting ROAS. If you want a fuller reporting layout, the guide to building a creative performance dashboard your team will actually use covers the cuts that matter.
For scale context, this is not a boutique measurement. Jumbo Gaming tracks performance across more than 220 campaigns on Linkrunner, attributing around Rs25 crore in revenue through the platform, which is the campaign-and-creative density this kind of view is built to handle.
Reading Creative-Level ROAS Without Fooling Yourself
A creative-level number is noisier than a campaign-level one, so it is easy to over-react to it. Three habits keep you honest:
- Wait for enough installs per creative. A creative with a handful of installs has a ROAS figure that is mostly noise. Judge on a meaningful sample, not the first day of spend.
- Respect the revenue lag. Revenue does not arrive at the same moment as the install. In practice the first revenue event can follow the install by anything from minutes to a couple of days, which is why day-0 creative ROAS routinely misreads a creative that is actually fine. Read it at D7 before making cut-or-scale calls.
- Cross-reference with cohorts. A creative can win on early ROAS and decay on retention. Pair the creative cut with advanced cohort analysis beyond D0, D7 and D30 revenue before you trust it.
How to validate this in your MMP
Pull one creative you believe is a winner. Confirm its creative ID appears on the install records, then confirm at least one revenue event resolves back to those installs. If either link is missing, the ROAS you are reading is inferred, not measured, and the plumbing needs fixing first.
FAQ
What is the difference between creative-level ROAS and campaign-level ROAS?
Campaign-level ROAS blends every creative in a campaign into one figure. Creative-level ROAS attributes revenue to the individual creative that drove each install, so you can see which specific ad pays back and which is carried by the others.
Why can I see CTR per creative but not revenue per creative?
Ad platforms report engagement up to the install but not post-install revenue, which lives in your app and your MMP. Unless the creative ID is passed into the attribution layer, revenue cannot be joined back to the creative.
Do I need a creative-analytics tool and an MMP, or just one?
They do different jobs. A creative-analytics tool scores the creative's elements; the MMP carries the creative ID to revenue. For creative-level ROAS you need the revenue link, and the element scoring is what tells you why a creative wins.
How long should I wait before judging a creative on ROAS?
Wait for a meaningful install sample and read at D7 rather than D0, because revenue lags the install. Judging on day-zero payback is the most common way teams kill a creative that was working.
Closing
Creative is the biggest lever most UA teams have, and optimising it on click metrics leaves the actual decision, where to put the next rupee, running on a proxy. Creative-level ROAS fixes that by carrying one identifier from impression to revenue, so the number you scale on is payback, not attention. Pair a creative-intelligence tool for the element scoring with an attribution layer for the revenue truth, keep the dashboard cut small, and read it at D7.
If you want the revenue link handled for you, Linkrunner keeps the creative identifier attached from click through to revenue. Request a demo and start with one creative: confirm its ID reaches the install, and its revenue reaches back.