Ad analysis for media buyers

The problem: your data tells you what's winning but not why

You've been in this situation: a creative comes out of nowhere, hooks at 18%, CVR is solid, and the account goes from $1k/day to $5k/day in a week. Now you need to scale it and brief variations before it fatigues. The account manager is asking what you'd test next. The creative team wants a brief by Thursday.

Your data tells you the ad is winning. It doesn't tell you which three seconds are doing the work, whether that mechanism will survive a budget increase, or which lever to preserve when you brief a variation. You watch the ad again. You notice the creator's energy, the hook's pacing, the testimonial overlay at the 12-second mark. You write a brief that says "keep the same vibe, test different hooks." The variations come back, none of them land, and the original fatigues before you've built a real creative stack.

This is the specific problem structured creative diagnosis solves. Not "which ad is winning" — your platform data already answers that. The question is why it's winning, at the creative-mechanism level, so you can brief variations that extend the winner instead of fragmenting the audience with surface clones.

The same diagnostic applies when you're studying a competitor's ad that's been running for six weeks. Long run time is the clearest signal that something is working. But watching the ad and trying to infer the mechanism without a fixed framework is how you end up copying a Hims ad's visual language without understanding that the mechanism was a BAB copy framework targeting Most-Aware retargeting traffic — and your cold prospecting brief was never going to work with the same surface.

How AdRevila fits your workflow

Paste the Meta or TikTok Ad Library URL — either your own ads (shared via public link) or a competitor's. AdRevila passes it to a single Gemini model via fal.ai, which watches the full video in one call with audio and visual context together. The hook archetype requires both. A Pattern Interrupt is a visual break against feed norms; a Direct Address requires the second-person framing plus a specific qualifier. Transcript-only analysis misses both.

The report comes back in 60–120 seconds: hook archetype, Schwartz awareness stage, body copy diagnosis, Cialdini lever tags, Hormozi value equation, CTA commitment level, and A/B hypotheses. Output as JSON, Markdown, PDF, or public share link for async review with your team.

For competitive analysis, the workflow is: find the ads that have been running longest in your competitor's Meta Ad Library, run them through AdRevila, and you have a prioritized diagnostic of which mechanisms are working in your category before you brief a single frame.

What you get back

Hook archetype. Classified as one of five types — Pattern Interrupt, Bold Statement, Question, Direct Address, Proof-First — with an explanation of why and a verdict on whether the hook earns the next five seconds. This is the brief input you need. Not "the hook is engaging" but "this is a Proof-First hook leading with a specific customer result, and it's earning the next five seconds because it opens a credibility loop the Problem-Aware viewer needs closed."

Schwartz awareness stage. The ad is placed on the five-stage ladder (Unaware through Most-Aware) based on what the opening assumes the viewer already knows. This is the match you're making when you scale: the mechanism that works at $1k/day against a cold prospecting audience will degrade if you scale into a more saturated audience segment without adjusting the stage framing. Knowing the stage is knowing the traffic temperature the ad was written for.

Body diagnosis. Copy framework (AIDA, PAS, BAB, or narrative) plus Cialdini levers tagged with the exact moment each fires. When you see Social Proof tagged at the 14-second mark because a customer count flashes on screen, that's brief-able. Your variation brief can say "preserve the Social Proof trigger at the body midpoint" instead of "keep the same energy."

Offer math. The Hormozi value equation scored per lever: Dream Outcome, Perceived Likelihood, Time Delay, Effort and Sacrifice. The CTA commitment level (low/mid/high) versus the awareness stage the ad is targeting. This pairing matters for scaling: a high-commitment CTA on cold Problem-Aware traffic is leaving CVR on the table, and the offer math section names that explicitly so you can fix it in the next variation.

Letter grade and A/B hypotheses. An A–F grade per the AdRevila rubric, plus two to four specific test variants framed as briefs. Not "test a different hook" — "test a Direct Address hook with the same Problem-Aware stage framing to see whether the explicit second-person qualifier improves the hook rate on cold traffic."

Which creative to scale, and why a hook is leaking

Two decisions media buyers make repeatedly where creative diagnosis changes the answer.

Scaling decisions. When you're deciding which creative deserves more budget, the mechanism read tells you something the CVR number doesn't: whether the win is portable. A creative winning on warm retargeting traffic with a Most-Aware offer structure and a high-commitment CTA is not a candidate for cold scaling — the mechanism is optimized for a different traffic temperature. A creative winning on cold traffic with a Problem-Aware hook and a low-commitment CTA is portable. You can scale it and expect the mechanism to hold. Without the diagnostic, you're scaling based on CVR alone, which means you'll scale the warm retargeting winner into cold traffic and watch the numbers collapse without understanding why.

Hook leakage diagnosis. When a creative's thumbstop ratio is good but CTR is weak, something between the hook and the CTA is leaking. AdRevila's body diagnosis and Cialdini lever tags show you where. If the hook archetype is earning the next five seconds but the body fires zero Cialdini levers and the copy framework collapses into description, that's the leak. The brief for the variation is specific: preserve the hook archetype, rebuild the body with two named levers, keep the CTA at the same commitment level. That's a testable brief. "Fix the body" is not.

The hook-body-CTA relationship and what Schwartz stage your winner is targeting are the two inputs that make every scaling and variation decision more precise. AdRevila extracts both from the video in the time it takes to read a Slack message.