If you're spending money on more than one ad creative, you need to know which one is actually making you money — not just which one gets clicks. Learning how to calculate ROAS by ad creative is the fastest way to stop guessing and start reallocating budget toward the videos and images that pay for themselves, and cutting the ones that don't.
What ROAS by creative actually measures
Return on ad spend (ROAS) tells you how much revenue came back for every unit of currency you spent. Calculated at the account or campaign level, it tells you whether your overall ads are profitable. Calculated at the creative level, it tells you something much more useful: which specific hook, script, or visual approach is driving that profit, and which ones are quietly dragging the average down.
Two campaigns can show the same blended ROAS while hiding wildly different creative performance underneath. One video might be paying for itself three times over while another barely breaks even — and if you never split them apart, you'll never know which one to double down on.
How to calculate ROAS by ad creative
The formula itself is simple. The hard part is making sure the revenue and spend numbers you plug in actually belong to one creative, not an entire campaign or ad set.
The basic formula
- ROAS by creative = Revenue generated by that creative ÷ Ad spend on that creative
- Example: a creative spends €400 and generates €1,600 in attributed revenue. €1,600 ÷ €400 = 4.0, usually written as 4x or 400%.
- If you prefer a percentage view: multiply the result by 100. A 4.0 ROAS is the same as a 400% return.
Where to pull the revenue and spend numbers
Spend is easy — every ad platform reports spend at the individual ad level. Revenue is where things get messy, because the platform's own attributed revenue, your store's order data, and your own UTM-based tracking will rarely agree perfectly. Pick one source as your primary number (most teams use the ad platform's attributed revenue for day-to-day decisions and reconcile against store revenue weekly), and be consistent about which one you use when comparing creatives against each other.
Setting up your account so creative-level ROAS is actually possible
You can't calculate ROAS by ad creative if your ad account structure mixes several creatives under one ad or one tracking link. Before you can report on it, you need a setup that keeps every creative separable.
Naming conventions
Give every ad a name that identifies the creative on sight: product, hook angle, format, and version number. Something like `vitamingummy_hook-painpoint_v3_static` is far more useful six weeks from now than `Ad 4`. When you're pulling a report later, a clear name saves you from having to click into every ad to remember what it actually looked like.
UTMs and one creative per ad
- Run one creative per ad, even if that means more ads in an ad set. Mixing two videos inside one ad makes creative-level ROAS impossible to isolate.
- Add a UTM parameter that includes the creative name or ID, so your store analytics and your ad platform data can be matched up later.
- Keep the same UTM structure across every campaign so you can build one report instead of rebuilding it each time.
If you're still figuring out the right testing structure, it's worth reading how to build a creative testing framework for Shopify before you lock in your naming conventions — fixing the structure after you've already launched dozens of ads is painful.
Step-by-step: building your creative ROAS report
- Export spend and attributed revenue per ad for a fixed date range, long enough that each creative has had a real chance to spend (a few days at minimum, more if your sales cycle is longer).
- Group the export by creative name, not by ad set or campaign, so ads running the same creative under different campaigns get combined.
- Calculate ROAS for each creative using the formula above.
- Add supporting metrics next to ROAS: cost per click, hold rate or thumb-stop rate, and cost per purchase. A creative with great ROAS but a tiny spend sample is a different conversation than one that's proven itself at scale.
- Rank creatives from highest to lowest ROAS and flag anything spending well below your account average — that's usually where budget is being wasted.
- Repeat on a set schedule (weekly is common) so you catch fatigue or a sudden drop before it eats your budget.
If you want a deeper walkthrough of interpreting the report once you've built it, see how to read ad creative testing results.
A worked example
Say you're running four creatives for a phone stand brand over the same one-week window. Here's how a creative-level report might look once you pull it together.
| Creative | Spend | Revenue | ROAS | Notes |
|---|---|---|---|---|
| Hook: problem-first, UGC style | €620 | €2,480 | 4.0x | Strong, scale budget |
| Hook: product demo, studio style | €540 | €1,080 | 2.0x | Below account average |
| Hook: testimonial, static image | €410 | €410 | 1.0x | Breakeven, candidate to pause |
| Hook: offer-led, bold text overlay | €300 | €1,500 | 5.0x | Small spend — watch before scaling |
Without this breakdown, the account might show a blended ROAS of around 2.6x and look merely okay. With it, you can see that one creative is actually the account's best performer and should get more budget, one is draining spend at breakeven, and one promising creative simply hasn't spent enough yet to trust the number.
Common mistakes that skew creative ROAS
- Comparing creatives with very different spend levels. A creative that's only spent €50 can show an inflated ROAS purely by chance. Wait until spend is roughly comparable, or at least note the sample size next to the number.
- Mixing attribution windows. If one platform reports on a longer click or view window than another, their revenue numbers won't be comparable — pick one window and apply it everywhere.
- Letting one audience carry one creative. If creative A only ever ran to your warmest retargeting audience and creative B only ran cold, the ROAS gap might be about the audience, not the creative. Where possible, test creatives against the same audience before concluding one is better.
- Judging too early. A creative needs enough impressions and conversions to produce a reliable number. A single early sale can make a brand-new ad look like a breakout winner for a day, then settle back to average.
- Ignoring post-purchase costs. ROAS tells you about revenue, not margin. A creative with a slightly lower ROAS but a higher-margin product mix in its sales can still be the better one to scale.
What to do once you know ROAS by creative
The number is only useful if it changes a decision. Once you have ROAS sorted by creative, use it to:
- Shift budget toward the top one or two creatives, in small increments, watching for ROAS to hold as spend increases.
- Pause or rework creatives sitting near breakeven rather than letting them run on autopilot — how to know when to kill a losing ad creative covers the specific signals to look for.
- Feed the winning hook, format, or angle into your next batch of creative so you're not starting from zero each round — see how to choose winning ad creative variants for a repeatable way to pick what to test next.
- Cross-check ROAS against hold rate before scaling. A creative with strong ROAS but a sharp early drop-off in hold rate may be running on a small, fast-converting audience that won't scale cleanly.
Keeping a steady supply of creatives to test
Calculating ROAS by creative only matters if you have more than one creative in rotation. Many accounts stall out because producing a new video for every hook idea takes too long, so the same one or two ads run until they fatigue and the whole account's ROAS slides with them.
FrameNotion was built for that gap. You paste a product link, and FrameNotion AI writes and renders a 30-second vertical video ad from scratch — hook, problem, benefit, proof, offer, and call to action — in about 10 to 20 minutes, with a voiceover, captions, and music already in place. Because each ad is generated fresh rather than pulled from a template, you can brief a different hook angle for each one and feed several genuinely distinct creatives into the exact ROAS report described above, instead of five near-identical versions of the same ad. Every ad also exports in 4:5, 1:1, and 16:9, so the same creative is ready for whichever placement you're testing it in. Browse example ads or see how it works if you want to see the output before you try it — the first ad is free on pricing.
FrameNotion doesn't publish to ad platforms or report performance back to you — you'll still pull spend and revenue from your ad accounts and store the way described above. What it solves is the upstream bottleneck: having enough distinct, finished creatives to actually run the comparison.
Frequently asked questions
What counts as a good ROAS for a single ad creative?+
There's no universal number — it depends on your margins, average order value, and how much of your cost structure the ad needs to cover. Instead of chasing a fixed target, compare each creative's ROAS against your account average and your actual breakeven ROAS (the point where revenue covers ad spend plus cost of goods), and treat anything above that as a candidate to scale.
Should I calculate ROAS by creative daily or weekly?+
Weekly is usually more reliable for comparing creatives, since daily numbers can swing from a handful of orders. Check spend daily to catch anything clearly broken, but make scaling or pausing decisions on a weekly rollup unless a creative is burning budget with zero sales.
Can I calculate ROAS by creative without UTMs if I only use one ad platform?+
Yes — most ad platforms report spend and attributed revenue per individual ad natively, so UTMs aren't strictly required if you're staying inside one platform's reporting. UTMs become important once you want to reconcile that revenue against your own store data or compare performance across platforms.
Why does my creative-level ROAS differ from my store's own revenue numbers?+
Ad platforms and store analytics often use different attribution windows and different rules for what counts as a conversion. Pick one source as your primary decision-making number and use the other as a periodic sanity check rather than trying to make them match exactly.
How many creatives should I be testing at once to get clean ROAS comparisons?+
Enough to find a winner without splitting your budget too thin to reach meaningful spend on any one of them. Three to five creatives per testing cycle is a reasonable starting point to test, then narrow down once a couple of clear leaders emerge.
