The 6-category rubric, line by line

The first three seconds. Whether the visual + on-screen text + audio earn the next three seconds. Specifically: pattern interrupt (something the eye doesn't expect on the FYP), curiosity gap (a question the viewer needs answered), or a high-density payoff delivered before the brain decides to scroll.

A weak hook score (under 30) usually means: opens with a logo, opens with the speaker introducing themselves, opens with a slow camera pan, or opens with an offer/CTA before earning attention.

Can a sound-off viewer figure out, in 5 seconds, what product or service is being talked about and what the value prop is? Clarity measures the on-screen text density, the visual storytelling, the framing of the subject: does the ad read without audio.

Affiliates often lose clarity points when their script is good but the captions are missing or the on-screen text doesn't name the product until 8+ seconds in.

Does the ad ask for the action it wants? Where in the runtime? How clearly? CTA scores high when the ask is specific ("tap the link in bio to save 20%" not "check it out"), appears at least twice (once mid-roll for the engaged viewer, once end for the patient one), and matches the offer vertical (TikTok Shop CTAs read differently than DTC link-in-bio CTAs).

Does the creative read as on-brand for the product, or could you paste any other product into the frame and the ad would work just as well? Brand fit penalizes generic templates (the same UGC structure used by every brand on the FYP), rewards specificity: a recognizable creator voice, a vertical-specific aesthetic, an offer-specific shot list.

The reasoning behind the weight: differentiated creative is harder to scroll past and easier to remember, and an audience saves what feels like a discovery. The rubric treats generic-template creative as a risk to surface before spend, not as a guarantee about what any campaign will pay per acquisition.

How TikTok-native does the creative read versus how ad-coded does it read. Specifically: how polished the production looks (UGC vs studio), framing (vertical 9:16 is the expected default; other aspect ratios lose points), pace of cuts, voiceover style, brand chrome density.

In the current rubric, high polish caps the native-feel score at 50 even when every other dimension is strong. The viewer's instinct is "I'm being sold to" → scroll. UGC-tier production routinely scores higher native feel than studio cuts of the same script.

Cut frequency and density. In the current rubric, below 15 cuts per minute on a video ad usually scores low (the viewer's brain has time to wander), while 30+ cuts per minute, without feeling frenetic, tends to score high; the attention loop refreshes constantly. Pacing also weights the payoff curve: where the value lands, how often, and whether the peak comes early enough for the short-form viewer, which the rubric models as roughly the first 15 seconds.

The Quick Check verdict (overall 0–100) is roughly: weighted average of the six categories, with hook contributing about 30% and the others ~14% each. Quick reports surface only the categories; Deep reports show the per-score evidence (which frame, which word) that drove each number.

The point of breaking the rubric apart isn't academic. It's so the fix is obvious. A 42 overall says "this needs work." A 42 overall with hook=20, native feel=80, clarity=70 says "rewrite the first three seconds, leave the rest."

The Ad Bench methodology: Every weight, threshold, and cap on this page describes The Ad Bench's current scoring rubric, not a platform rule. The rubric is designed to identify creative risk before media spend, not predict or guarantee campaign performance, and its calibration can change between builds.

1. The Ad Bench methodology, current scoring rubric. Reviewed August 24, 2026.