Reviewing creative test results with your team means getting everyone who touches the ad account into one meeting, looking at the same numbers in the same order, and leaving with a written decision for each creative — kill it, iterate on it, or scale it. Done well, this takes 30 to 45 minutes a week. Done badly, it turns into a slide of screenshots, a vague feeling that "the UGC ones are doing better," and no actual decision — which means the same underperforming ad keeps spending money for another week.
Why most creative review meetings don't produce decisions
The usual failure mode isn't a lack of data — it's a lack of structure. Someone shares their screen, scrolls through the ads manager, and the group starts reacting to whatever catches the eye first: a low cost-per-click here, a comment that got a laugh there. Thirty minutes later the meeting ends with "let's keep an eye on it" for most ads, which is another way of saying nothing changed.
A review that actually produces decisions has three things the scroll-and-react version doesn't: a fixed list of creatives to review, a fixed order of metrics to look at, and a rule for what each outcome means. Build those three things once and the meeting becomes mechanical — which is exactly what you want from a recurring process.
How to review creative test results with your team: the 6-step process
This is the structure to run every time, whether you're a solo founder reviewing with a freelancer or a team of five reviewing with an agency.
- Set the cadence and invite list before the test starts. Decide the day and time weekly (for example, every Monday at 10am) and the standing invite list: media buyer, creative lead, and whoever owns the offer or product. Same people, same slot, every week.
- Pull the data into one sheet before the call, not during it. One person is responsible for exporting spend, impressions, hook rate, hold rate, click-through rate, and conversion rate (or cost per result) for every creative under test, into a single sheet everyone can see. Nobody should be alt-tabbing into the ads manager live during the meeting.
- Apply a minimum threshold before anyone discusses results. Agree ahead of time on a minimum spend or impression count below which a creative is "not enough data yet" and gets skipped this week. Reviewing a creative that's had two days and a handful of impressions wastes time and invites opinion-based arguments.
- Review the numbers first, in silence, for two minutes. Have everyone look at the sheet before anyone speaks. This sounds small but it changes the conversation — people arrive with a read on the data instead of a read on who talks first.
- Discuss only the creatives that are close calls. Ads that are clear winners or clear losers against your bar don't need debate — they need a decision logged and the meeting moves on. Spend your discussion time on the ambiguous middle, where instinct and data disagree.
- Write the decision down before the meeting ends. For every creative: kill, iterate (and what changes), or scale (and by how much budget). Assign an owner and a deadline for the next action. If this step gets skipped, the meeting didn't actually happen — it was just a discussion.
If you don't already have a consistent way to log results between meetings, it's worth setting that up first — see how to document creative test results for a step-by-step template your team can reuse every week.
What to look at, and in what order
The order you look at metrics in shapes the conversation. Start too broad (overall ROAS) and you'll argue about the whole account instead of the creative. Start too narrow (comments) and you'll overweight anecdotes. A workable order, top to bottom:
- Delivery: did the creative actually get enough impressions or spend to be judged? If not, hold judgment.
- Attention: hook rate or video-played-to-25% — this tells you whether the first few seconds are earning attention, independent of everything else.
- Retention: hold rate or thumb-stop ratio through to the midpoint — this isolates the body of the ad from the hook.
- Engagement signal: click-through rate — a proxy for whether the offer and call to action are landing, not just the opening.
- Outcome: conversion rate or cost per result — the number that actually decides spend, but one that's easy to misread if the earlier signals weren't checked first.
Reviewing in that order stops the common trap of declaring a creative "bad" because of a weak outcome number, when the real issue is a weak hook that never got enough people far enough into the video to see the offer. If hook rate is fine but outcome is weak, the fix is usually the back half or the offer. If hook rate is weak, nothing downstream matters yet — the opening needs work first. For a deeper look at isolating the opening specifically, see how to test a hook before spending ad budget.
Separating data from opinion in the room
Every creative review has a moment where someone says "I just don't love this one" about an ad that's performing fine, or defends an ad that's quietly losing money because they wrote the script. Both are normal — and both need a rule, not a debate, to resolve.
A simple house rule that works: the data sets the default decision, and anyone who disagrees has to propose a specific, testable change rather than a feeling. "I don't love this" doesn't override a creative that's clearing the bar. "I think the offer slide is too late — let's move it earlier and test that as a variant" is a real proposal and gets logged as the next test, not argued about in the room.
It also helps to separate two different questions that get tangled together: is this creative performing, and is this creative done being tested? A creative can be performing well and still be a candidate for a variant test (new hook, new proof point). A creative can be underperforming and still be worth one more iteration before you kill it, if the underlying angle hasn't really been tested yet. Keep those as two separate columns in your sheet so the team isn't debating both at once.
The kill, iterate, or scale decision framework
Every creative under review should land in one of three buckets by the end of the meeting. Set rough rules in advance so the room isn't inventing a bar on the spot.
| Decision | When to use it | What happens next |
|---|---|---|
| Kill | Below your minimum bar on the core outcome metric after a full test window, with no clear fixable reason (hook and hold were both fine, outcome still weak) | Pause spend, archive the asset, note the angle so it isn't retested blindly later |
| Iterate | Close to the bar, or clearly weak in one specific stage (hook, hold, or offer) with an obvious fix to try | Brief a specific variant — new hook, new offer slide, new ending — and queue it as the next test, not a guess |
| Scale | Clearly above the bar on outcome, with enough delivery data to trust it | Increase budget in controlled steps and watch for fatigue rather than doubling spend overnight |
For the "scale" bucket, pair this review with a plan for how new budget gets allocated across creatives so a single winner doesn't accidentally starve the next test — how to allocate budget across creative tests covers that in more detail. And before declaring anything a long-term winner, check how long should you run a creative test before deciding — ending a review a few days early is one of the most common reasons teams scale the wrong ad.
Running the review when the team is remote or asynchronous
Not every team can get everyone on a live call weekly, and that's fine as long as the structure survives the move to async. The fix is to make the sheet itself do the work a live meeting would do:
- Lock a submission deadline (for example, every Friday by 3pm) for the person pulling the data, so the sheet is ready before anyone needs to react to it.
- Add a comment column next to each creative where each stakeholder leaves their read before seeing anyone else's — this preserves the "look first, talk second" effect without a live room.
- Set a decision deadline separate from the data deadline (for example, Monday noon) so the sheet doesn't sit open indefinitely with no owner closing it out.
- Still assign one person to make the final call on close ones. Async reviews die when a decision needs five people to agree and nobody goes first.
Whichever format you use, keep the historical reviews somewhere searchable. If your team is testing several hooks or angles a week, it gets hard to remember which variant of which ad underperformed three months ago unless the files and the decisions are organized the same way every time — see how to organize ad creative files for testing for a filing structure that holds up over dozens of tests.
Common mistakes that derail the review
- Reviewing every ad with the same depth. Spend your discussion time on close calls, not on obvious winners or obvious losers.
- Changing the metrics you care about mid-test. If hold rate suddenly becomes the headline metric because it happens to favor someone's creative, the whole review loses credibility.
- No minimum spend threshold, so the team argues about noise. A handful of impressions can produce a dramatic-looking percentage that means nothing.
- Letting the meeting end without a written decision. "We'll revisit" is not a decision — it's a way to avoid one.
- Looking at results in isolation from the pattern across past tests. A single creative review is useful; a pattern across many reviews is what actually tells you what to brief next — see how to identify winning creative patterns for how to build that view over time.
Where FrameNotion fits into the review cycle
The review meeting is only as useful as the next batch of creative it produces. Once the team agrees a hook needs testing, or an offer slide needs to move earlier, or a new angle deserves a shot, that decision needs to turn into an actual ad fast — otherwise the review becomes a list of good intentions that never get tested.
FrameNotion is built for that handoff. Paste the product or landing page link, and FrameNotion AI writes and renders a 30-second vertical ad (9:16, 1080×1920) from scratch, complete with voiceover, music cut to the beat, and word-by-word captions, in about 10 to 20 minutes. If the review decided a specific hook or offer needs testing, you can add notes about that angle, upload product images, and get a draft ready before the next review instead of weeks later. After it's rendered, text and colors can be edited without starting over, and the same ad also comes out in 4:5, 1:1 and 16:9 for wherever the test is running. The first ad is free on signup, with no card required — see pricing or browse example ads to see the range of outputs before committing a test slot to it.
Frequently asked questions
How often should a team review creative test results?+
Weekly is a reasonable starting point for most accounts running regular creative tests. If spend or test volume is low, every two weeks may make more sense — the key is a fixed, repeated cadence rather than reviewing whenever someone remembers to check.
Who should be in a creative review meeting?+
At minimum, whoever buys the media and whoever creates the ads. If an agency or freelancer is involved, include them live rather than relaying decisions secondhand — misread context is one of the most common reasons good creative gets killed too early or kept too long.
What if the team disagrees on whether to kill or iterate a creative?+
Set the rule before the disagreement happens: data sets the default decision, and anyone who wants to override it has to propose a specific, testable change rather than a general opinion. That turns a debate into a new test, not a stalemate.
Should creative reviews include qualitative feedback like comments or DMs?+
They can, but treat it as context, not a decision-maker. Comments and reactions can explain why a number moved, but the quantitative metrics should still set the kill, iterate, or scale call.
How do we keep reviews consistent when the team changes over time?+
Write the agenda, the metric order, and the decision thresholds down somewhere everyone can reference, rather than keeping them in one person's head. A documented process survives staff turnover; an informal habit usually doesn't.