Why affiliate ads need their own rubric
▾
A brand-direct ad is trying to build awareness and purchase intent. An affiliate ad is trying to do one specific thing: get a viewer to click a link and buy something in the next 20 minutes. The failure modes are different, the algorithm signals are different, and the compliance requirements are different.
The five most common ways affiliate ads leak revenue — buried product, code mentioned but not shown, missing FTC disclosure, studio-polished execution, and vague CTA — are all detectable before you post. The Ad Bench runs these checks automatically when it detects affiliate signals (discount codes, #ad tags, link-in-bio language, TikTok Shop product tags).
The eight-check rubric
▾
Product-first hook (3s)
Highest weightDid the product appear in frame within the first 3 seconds, or did you spend that time on a creator-first intro? Affiliate buyers need the visual answer fast.
FTC disclosure
Compliance checkWas #ad / Paid Partnership / 'I get a small commission' visible in the first 3 seconds — not just buried in caption hashtags? We flag missing or hidden disclosure.
Code / link clarity
Conversion-criticalIf you mentioned a code, was it also on screen long enough to read? Was 'link in bio' specific about what the link is for? Vague CTAs leak conversion.
Native execution
Algorithm fitStudio polish reads as 'ad' to TikTok viewers and to the algorithm. Handheld, ambient sound, no-makeup energy. We score against actual native-looking top performers.
Save-bait potential
Algo distributionTikTok's algorithm weights saves heavily for shopping content. We check whether your ad gives viewers a reason to save — gift idea, restock alert, comparison reference.
Comment-bait CTA
Compounding distributionDM-me-the-code, ask-in-comments, or 'name a brand that' CTAs drive comment volume — and comment volume drives algo distribution to other shopping-intent viewers.
Pacing for retention
Hold rateCuts every 1–2 seconds in the first 6, then 2–3 second beats for the demo, then a 2-second hold on the result frame. We score your actual rhythm against this.
Vertical fit
Category-specificBeauty wins on close-ups + sound-on. Tech wins on hands + functional demo. Fashion wins on GRWM / OOTD. We score against the category your product is in.
Per-platform differences
▾
The eight checks apply to every short-form affiliate ad. The mechanics underneath — which CTA primitive wins, which algorithm signal dominates, what kind of polish reads native — shift per platform.
TikTok Shop
- Hook rate benchmark: ≥30% strong, ≥40% breakout
- Top algorithm signals: Saves + comments. Comment-bait CTAs ("DM me CODE") drive comment volume — the #2 TikTok signal after saves.
- Winning CTA: On-screen code + comment-bait. The combination hits the algorithm twice: comment volume drives distribution, code recall drives conversion.
- Compliance: FTC disclosure in voiceover + on-screen text by frame 1. #ad in caption — buried in hashtag list doesn't count.
Instagram Reels (LTK / Amazon Live)
- Hook rate benchmark: ≥35% strong, ≥45% breakout
- Top algorithm signals: Saves are the #1 IG signal — heavier than likes or comments. Story-shares are #2.
- Winning CTA: Save-bait. "Save this for next time you're shopping for X." Layer on bio-link routing through LTK or Amazon Live.
- Policy note: Organic IG music library is off-limits for ads — use commercial-cleared Sound Collection or licensed audio.
- Compliance: FTC first-3-seconds rule + Branded Content tag in addition to on-screen and verbal disclosure.
YouTube Shorts (Partner Program / Amazon Influencer)
- Hook rate benchmark: ≥40% strong, ≥55% breakout
- Top algorithm signals: Repeat-views. Shorts that loop seamlessly outperform. Subscribe-lift carries ~10x algorithmic weight (Shorts → long-form funnel).
- Winning CTA: Subscribe-prompt + pinned-comment link. The pinned top comment is the YouTube affiliate convention — viewers tap it to copy the link.
- Policy note: Brand-safety bar is higher than long-form. Fake comment/like overlays = ad rejection. Non-original Shorts are demonetized.
- Compliance: YouTube's paid-promotion disclosure toggle should also be enabled alongside FTC verbal + on-screen disclosure.
FTC disclosure — what actually counts
▾
The FTC's standard is "clear and conspicuous." In practice for short-form video, that means:
- On-screen text or branded-content tag visible in the first 3 seconds — not 8 seconds in after the intro.
- Verbal disclosure in the voiceover ("I get a small commission") in the first 3 seconds.
- Caption #ad does not count as the sole disclosure — it must be clear and conspicuous within the video itself.
FTC enforcement was up ~340% in 2025 vs 2021. The agency issued over 150 warning letters to creators. The Ad Bench flags missing, late, or buried disclosure before you post.
GMV projection — how it works
▾
After scoring, The Ad Bench projects affiliate revenue by multiplying predicted hook rate × realistic CTR × conversion rate × AOV × commission rate — at 10K, 100K, and 1M views. It then shows what that number becomes if you implement the top fixes.
The projector is calibrated for TikTok Shop economics by default (5–30% commission, $10–$80 impulse-buy AOV). Per-platform commerce math for Reels (LTK / Amazon Live) and Shorts (Amazon Influencer / YouTube Partner) uses conservative multi-platform averages.
Common questions
▾
What counts as an affiliate ad in The Ad Bench?
Any short-form video ad where you earn commission on attributed sales — TikTok Shop product tags, Instagram Reels with LTK / Amazon Live / Shopify links, YouTube Shorts with pinned-comment affiliate links, plus brand-specific discount codes and direct DTC affiliate links on any platform.
Does this work for Reels and Shorts, not just TikTok?
Yes. Paste a TikTok / Instagram Reels / YouTube Shorts URL, drop a video file, or upload an image. Detection runs automatically — you can override with the platform chip. The same rubric scores all three, with per-platform calibration for hook-rate thresholds, completion-rate bars, audio-on prevalence, and algorithm signals.
Does the FTC disclosure check actually catch problems?
Yes. We check both the visual (on-screen text or branded-content tag visible in the first 3 seconds) and the audio (verbal disclosure). Caption-only disclosure does not count — the FTC requires disclosure to be 'clear and conspicuous'. We flag missing, late, or hidden disclosure.
Will my hook score predict actual conversion?
Hook rate predicts the percent of viewers who watch past 3 seconds — the population that has a chance to click your link. The GMV projection multiplies hook rate × CTR × conversion rate × AOV × commission to project affiliate revenue. We do that math in the report.
Is it free?
Free plan: 3 Quick Checks + 1 Deep Dive a day with a free account (1 Quick Check/day without an account). Pro ($49/mo) and Agency ($199/mo) add 10+ reports/day and workbench tools.
Read to the end to earn a star.