Preflight Plus Performance Is Here. The First Pre-Launch Score Built From Your Own Data.
Aug 24, 2026
This is the big one.
For years, every pre-launch creative score in this industry has been built on the same quiet assumption: that there is one right way to make an ad, and your creative either follows it or it doesn’t.
Brand safe zones. Logo in the first three seconds. Captions on. Hook under two seconds. All true. All generic. All of it describes what the industry says a creative should be, and none of it knows a single thing about what has actually worked in your account, on your KPI, with your audience.
None of that goes away. Best practice is the floor, and the floor holds everything up. But a floor was never meant to be the whole building.
Today we are adding the rest of it.
Preflight Plus now scores every creative on two signals instead of one. It is connected directly to your account, it benchmarks each new creative against every creative you have ever run, and it hands you one number before you spend a cent.
Two signals. One number.
Best practice covers what a creative should be. Platform specs, brand rules, the ABCD frameworks, the things that quietly cap your delivery when you get them wrong.
Performance covers what your own historical data says actually works. Not benchmarks. Not industry averages. Not what worked for someone else’s brand in someone else’s vertical. Yours.
One is the rulebook. The other is the receipts. Preflight Plus now reads both, and gives you a single score you can take into a creative review and defend.
Pick your lens
Upload a batch, choose how it gets scored:
Both (default). Final score = Performance x 0.7 + Best Practice x 0.3. Performance carries the weight, because your own results outrank a universal rulebook every single time.
Performance only. Your historical signal at 100%.
Best Practices only. Exactly what you have today, untouched.
Your choice drives everything downstream: which sub-scores appear, which recommendations render, which tooltips explain the math. No relabeled screens. Every explanation is built for the lens you picked.
The pipeline, in four steps
- Tag the creative with our proprietary taxonomy engine. The same engine behind the Creative Genome breaks your creative into its component parts: hook style, pacing, end card, on-screen objects, talent, text density, CTA timing and hundreds more. This is the layer nobody else has, and it is the reason everything after step one is possible.
- Score every tag on best practice, per platform. Multi-select Google, Meta and LinkedIn, and each gets its own independent read. A creative that sails past Meta’s bar can still miss LinkedIn’s, and now you will know before it launches.
- Score every tag on performance, against your leading KPI. One score, cross platform. Your data does not care which manager the file was uploaded to. It cares what the content did.
- Roll it all up. One headline number, both sub-scores visible underneath, so you can always see which signal is pulling and why.
- Get agentic recommendations, grounded in your own data. A score tells you where you stand. This tells you what to do about it. Alison surfaces the specific elements holding the creative back and what to replace them with, pulled from the tags that have actually won in your account. Best practice recommendations arrive per platform. Performance recommendations arrive once, cross platform, powered by the same iteration engine behind Alison’s creative briefs. You do not leave Preflight with a grade. You leave with a version two.
How the performance score is actually built
No black box. Here is the whole thing in plain language.
Every tag has a track record. For each tag, we pool all your past creatives that used it and measure their combined KPI. The blended ROAS of everything carrying “fast-cut hook,” for example. Higher-spend creatives count more, because they were tested harder.
Graded on a curve, not an absolute. Each tag is standardized against every other tag in your account. Direction follows the KPI automatically: higher ROAS is better, lower CPA is better. Higher is always better on the final score, on every account, on every metric.
Evidence earns trust. A tag seen on 40 creatives is trusted. A tag seen on 5 gets pulled toward neutral until it proves itself. Below the data floor, we tell you it is insufficient data instead of inventing a number.
Weighted by what actually matters. Tags combine by how long they hold the screen, how distinctive they are (a tag on every creative carries no signal, a rare one carries a lot), and how strong the signal is. Neutral tags get ignored so they cannot drag a strong creative back toward average.
Ranked 1.0 to 10.0. Your creative is ranked against the rest of your account and mapped to a one-decimal score. The full range is always used, so nothing clusters at 5 and every single score means something.
You choose the metric. It recalculates live.
Your account’s leading KPI is the default. Override it any time: CTR, CVR, ROAS, CPA, video completion rate and more, based on what each platform supports.
Switch the metric and every performance score in the session recalculates instantly. The active metric sits right beside the score, so you always know exactly what you are looking at: Scored by: ROAS.
Metrics that do not fit the creative type in your batch are disabled automatically. No scoring static banners on video completion rate.
And your existing context inputs (objective, placement, audience) now pull double duty. They filter the historical dataset behind the performance score, so a prospecting video is judged against prospecting videos, not against your entire retargeting library.
The honest part
Confidence means telling you where the edges are.
It reads content, not context. A creative can win on offer, timing or targeting and still score mid because the content itself looks ordinary. When that happens, the model is being honest with you, which is the entire point.
It runs on your data. Thin conversion history means thin tag records. Where the evidence is not there, Preflight Plus says so and falls back to New Concept rather than guessing.
It is account relative. A 7 in your account is not a 7 in anyone else’s. That is not a limitation. That is the whole idea. The benchmark is you.
What changes on Monday morning
The question in your creative review stops being “does this follow the rules” and becomes “does this look like what has already worked for us, and if not, do we have a reason.”
You will see exactly which elements are dragging a creative down before a cent is spent. You will catch the creative that ticks every platform box and contains nothing your audience has ever responded to. And when you want to run something bold and rule-breaking, you will be able to prove whether the risk has precedent in your own account.
That is not a scoring tool. That is a different way of working.
Preflight Plus Performance is live now.
Book a demo and we will score your creative against your own data, on the call.
