// TIKTOK_ANALYZER
TikTok ad analyzer
Drop a TikTok ad in any format — video, image, Spark-eligible URL, or script — and get a senior TikTok strategist's read in under 90 seconds. Hook-rate prediction, six-category rubric, ten predicted performance metrics, Spark Ads readiness, GMV projection for TikTok Shop, FTC + commercial-content disclosure check, and three rewritten hooks calibrated to 2026 FYP benchmarks.
// BENCHMARKS_2026
2026 TikTok ad benchmarks
Median ranges from the Ad Bench dataset, calibrated against TikTok paid-creative performance in early 2026. Every Deep Dive report grades the submitted ad against these benchmarks.
| Metric | Median range | Note |
|---|---|---|
| Hook rate (3s) | 25–40% | TikTok ads average 25–30%. Above 40% is strong; below 20% the FYP throttles distribution within the first hour. |
| Save rate | 0.4–1.2% | TikTok weights saves heavily for TikTok Shop and affiliate content. >1.2% suggests strong rewatch intent. |
| Comment rate | 0.5–1.8% | TikTok's #1 engagement signal for early distribution. Code-bait CTAs ("DM me 'lip'") drive this hardest. |
| Watch-completion | 30–55% | Below 30% is the cold-start kill window. The FYP runs a reactive bucket-test on every post in the first 200 views. |
| Cost-per-click range | $0.30–$2.50 | Spark Ads typically run 30–50% cheaper than native In-Feed because the creative arrives with organic signal baked in. |
// WHAT_WE_SCORE
What the TikTok analyzer scores
The six universal rubric axes (hook, clarity, CTA, brand fit, native feel, pacing) calibrated for TikTok— plus per-platform sub-axes that don't exist on the cross-platform rubric.
Hook
First-frame thumb-stopper + 3-second contract. Calibrated to TikTok's FYP cold-start window where the first 200 views decide the next 200K.
Native feel
Polished = ad-coded = scroll. TikTok rewards selfie-cam authenticity harder than any other platform; the polish penalty is real and measured.
CTA architecture
Comment-bait, code-clarity, link-in-bio framing, pinned-comment timing. The mid-roll + closer pattern that converts on TikTok Shop creative.
Spark suitability
Four readiness gates: comment volume, save/share signal, watch-completion, and creative-control consent. Spark Ads inherit organic signal — strongest organic clip isn't always strongest Spark candidate.
Affiliate / Shop rubric tilt
When TikTok Shop or affiliate signals fire, the rubric reweights: CTA + native-feel + brand-fit up, hook slightly down, FTC + commercial-content disclosure becomes a hard requirement.
// FAQ
Frequently asked questions
What does the TikTok ad analyzer actually score?
Six rubric categories (hook, clarity, CTA, brand fit, native feel, pacing) plus ten predicted performance metrics (hook rate, hold rate, watch-completion, save rate, comment rate, share rate, follow rate, CTA CTR, average watch time, engagement rate) — all calibrated to 2026 TikTok benchmarks. Deep Dive mode adds Spark Ads readiness, GMV projection for TikTok Shop creative, business projections (CPM/CPC/CVR brackets, daily-budget range, creative lifespan), risk flags, ranked top issues, and three rewritten hooks plus caption + CTA variants.
What inputs does it accept?
Four formats: (1) MP4 video upload — runs locally in the browser, only JPEG keyframes upload; (2) static image — Spark Ad still or planned thumbnail; (3) public TikTok URL — pulls metadata via oEmbed + processes the cover frame; (4) script text for pre-production review. All four route through the same six-category rubric.
Is it tuned for TikTok specifically?
Yes. The TikTok mode uses a TikTok-specific system prompt that calibrates hook against FYP cold-start, native-feel against the polish penalty, CTA against TikTok's comment-led ranking, and includes a Spark Ads readiness sub-axis that doesn't exist on the cross-platform rubric. URL submissions auto-detect TikTok and route to this mode; the platform chip lets you override.
Does it score TikTok Shop / affiliate creative differently?
Yes. When affiliate or TikTok Shop signals are detected — keywords like 'code', 'link in bio', 'shop my', visible cart UI, FTC-disclosure phrases — the rubric reweights: CTA + native feel + brand fit get heavier weight, hook gets slightly lighter, FTC + commercial-content disclosure becomes a hard requirement, and the analyzer surfaces a GMV projection.
How long does an analysis take?
Quick Check mode (5-category scorecard with 3-5 fixes) takes about 60 seconds. Deep Dive mode (full rubric, predicted metrics, business projections, rewrites) takes about 90 seconds for a 30-second clip. Longer videos take proportionally longer; we cap frame extraction at 10 to keep tail latency reasonable.
Is there a free tier?
Yes. A free account gets 3 Quick Checks + 1 Deep Dive per day (1 Quick Check/day without an account). Pro is $49/mo (10 Quick + 10 Deep per day, full Deep report, saved history). Agency is $199/mo (100 analyses per day, brand kit, drafting workbenches, API access, team seats).
Does my video stay private?
Yes. Original MP4 bytes never leave the browser — only extracted JPEG keyframes upload to the analyzer. Frames are sent to Anthropic's API for analysis; per Anthropic's commercial API terms, they're not used for training. Saved reports live in your account only.
// RELATED_GUIDES
TikTok ad strategy guides
// READY_TO_SCORE
Drop your TikTok ad below and run the scoring.