Ad creative testing
that teaches you something every round.

A repeatable way to run tests that teach: choose the variable from the angle and hook matrix, write a hypothesis, keep everything else fixed, name the ads by a convention, set the budget and dates, judge with one metric, write the lesson down and brief the next round from it. The framework matters more than any single test, because it is what makes results comparable across rounds.

Meta offers two routes. The creative testing tool compares up to seven variants inside an existing campaign using the Highest volume bid strategy, with a share of budget and a comparison metric you choose. An A/B test splits the audience so nobody sees both versions. Meta recommends one variable per test and a measurable hypothesis, and the creative testing tool makes no automatic changes to your ads when the test ends.

Usually the angle, then the hook, because they decide who stops and why; format, creator and offer tests then refine a message that already works. If the offer is weak, no hook will rescue it, so for a product that barely converts the offer can be the first test. Whichever you pick, change that one element and keep the rest of the ad identical.

Enough for each variant to produce a stable number on the metric that judges it, which depends on your cost per result and your audience size. Meta suggests keeping its creative tests to no more than 20% of the existing campaign or ad set budget, and recommends at least seven days for A/B tests, which run for up to 30 days. Write your own minimum spend per variant into the plan before launch.

Hook rate is 3-second video plays divided by impressions: how many viewers stopped scrolling. Hold rate is 15-second ThruPlays divided by 3-second plays, where Meta counts one when a video plays to the end or for at least 15 seconds. Some teams divide by impressions instead, so write your definition on the plan and keep it the same across rounds.

Use a fixed order of short codes separated by underscores: date, product, angle, hook, format, creator and version, such as 0914_SLEEP_problem_H2_ugc_maya_v1. Keep a legend of the codes next to the test plan and never rename a live ad. With every name built the same way, results can be grouped by hook, angle or creator in any spreadsheet or analytics tool.

Decide before launch: the metric, the minimum spend or results per variant, and the gap that counts as a win. Meta's A/B tests can report that no winner was found, and a creative test in Ads Manager names top performers without a confidence level, so the rule has to be yours. When the gap is small, call it a draw, keep the control and test a bigger difference next.

A running record of what each test taught, one sentence per test, tagged by variable, product and audience. It stops a team from retesting what it already knows and gives every new brief a starting point. In Storyflow it is a titled column of notes on the testing board, and the AI reads it when it drafts the next round's scripts.

No. Storyflow plans the tests and briefs the variants; it does not connect to Meta, TikTok or Google ad accounts, launch ads or import results. Launch and read tests in Ads Manager, and use a creative analytics tool such as Motion, which syncs results from Meta, TikTok, YouTube and LinkedIn accounts and groups them by naming convention or AI tags. Copy the numbers that decide each test into the results table.

It can draft them, and you choose. Give Storyflow's AI the product notes, reviews and last round's lessons, and its video script workflow drafts a script per variant with the hook written out, while the storyboard workflow lays out the frames. It reads the board you have open and up to five attached documents. It cannot know which hook will win; only the test can.

Not during early access. Storyflow is paid-only during early access, with plans from $7.99 a month billed annually ($9.99 monthly). A free plan launches before the end of 2026, and anyone a paid member invites to a board joins free. Your creators pay nothing to read their briefs.