How to analyze competitor video ads without guessing what worked
A public ad can show you the message, footage, offer, and advertiser. It rarely shows the full campaign. Use this workflow to separate what you can observe from what needs your own data.
Start with a boundary: creative analysis describes the video. Performance analysis measures what happened in a campaign. A visible hook or product demo is an observation. “This ad is profitable” is a claim that needs reliable spend and outcome data.
1. Find examples in the platform's own library
Choose the country, advertiser, product category, or platform that fits your research question. Keep the platform and observation date in your notes because ad libraries have different coverage and change over time.
- Meta Ad Library: search by advertiser or keyword to review ads currently running across Meta technologies. The library is useful for seeing the creative and basic advertiser context. Do not treat ordinary commercial ads as a complete archive or a source of their sales results.
- TikTok Creative Center Top Ads: browse the selected, advertiser-authorized examples and filters. TikTok may display engagement or performance indicators for a listed ad. Those indicators apply to that ad and the library's selected set, not your campaign or every ad in the market.
- Google Ads Transparency Center: search for an advertiser or website and inspect ads published through Google. Its transparency listing helps identify advertiser and creative examples; it does not provide a complete competitor performance report.
Save the library entry or ad identifier, advertiser name, platform, country or region filter, and the date you looked. If an ad disappears later, your dated note will still show what you reviewed.
2. Build a small evidence log
Pick three to five relevant examples instead of collecting a large unfiltered swipe file. Record only details you can point to in the ad or library. The final column is where you put a question or interpretation, not a fact.
| Record | Observable evidence | Keep separate |
|---|---|---|
| Opening | First spoken line, visible action, on-screen text, or opening image | Whether a particular audience found it persuasive |
| Demonstration | What the creator shows, in what order, and whether the result is visible | Whether the demonstration caused a sale |
| Proof | Specific comparison, testimonial, interface, claim, or supporting detail | Whether the claim is independently substantiated |
| Offer and next step | Price or offer shown, spoken ask, on-screen ask, and destination when visible | Conversion rate, customer value, or return on spend |
| Context | Advertiser, library ID, region, active date, format, and source URL | Audience targeting or budget when the library does not show it |
Use two labels in your notes: “Observed” for something visible, audible, or explicitly listed by the source. “Hypothesis” for what you want to test in your own creative. This stops a plausible explanation from turning into a false fact.
3. Review one video in sequence
Watch the whole ad once without pausing. Write one sentence describing the promise and the action requested. Then replay the opening and the key demonstration. Note approximate playback positions only as navigation aids; a sampled report timestamp may not mark an exact edit boundary.
- Opening: What does the viewer see and hear first? Do the words and image support the same promise?
- Product proof: When does the product appear, and what can a viewer actually see it do?
- Progression: Does each scene add a new reason to care, demonstrate use, answer a concern, or repeat an earlier point?
- Offer and CTA: What should the viewer do next? If a library page shows the destination, does it match the ask?
- Missing evidence: What would you need campaign reporting, a source replay, or another document to verify?
For a human review of the first three seconds, write the four checks separately: opening promise, visible proof, transitions, and audible beats. If a field is absent from the generated report, label it as your own review. If you assign a number, state the rubric and call it an editorial score.
4. Turn the pattern into an original test
Do not copy the competitor's footage, wording, creator, or distinctive treatment. Reduce the observation to a broad strategy, then express it with your own product, audience, evidence, and permission-cleared assets.
Illustrative example, not a real ad or performance result: You observe several authorized software ads showing the interface before the end card.
- Observed: the screen recording appears before the offer in the examples reviewed.
- Hypothesis: showing the workflow earlier could help our intended audience understand the product.
- Original test: keep our claim and CTA fixed, but move our own permission-cleared interface demonstration earlier in one variant.
- Measure: compare the chosen campaign outcome under comparable audience, placement, offer, and budget conditions.
The observation suggests a test. It does not show why the competitor ran the ad or predict the outcome of your version.
Change one important creative element at a time when practical. Keep the audience, placement, offer, landing page, and measurement window consistent if the goal is to learn what the edit changed. Use your own campaign dashboard for results. A public library is not a substitute for your experiment.
Where AdScanVideo fits
AdScanVideo's Ad intelligence mode can create a first-pass report from an uploaded video file you have the right to submit and analyze. It can organize observations about the hook, audience, promise, proof, pacing, offer, and CTA. It does not read a competitor's private campaign metrics, and video links are not accepted. If you do not have permission to submit a clip, review it in the public library and make your own notes without uploading or republishing it.
Check the original source before relying on a quote, claim, timestamp, or visual interpretation. The Ad intelligence overview explains the report's scope. The CTA and landing-page audit covers the click path, and the public sample report shows the difference between generated observations and a separately labeled human hypothesis.
Copyable competitor ad review template
Prefer a spreadsheet? Download the blank competitor video ad review sheet. It records the library source and separates what you observed from your hypothesis and the result you will measure in your own campaign.
Frequently asked questions
Can a public ad library tell me whether a competitor's ad is profitable?
Usually no. Public libraries can show creative and advertiser details, and some selected programs expose specific performance indicators. They do not generally reveal a competitor's complete spend, audience, margin, or return. Use your own campaign data to measure your results.
Can AdScanVideo analyze an ad from its public URL?
No. AdScanVideo currently accepts uploaded video files, not video links. You must have the right to submit and analyze the file. If you do not, use the public library to review the ad in place and record your own observations.
What should I record when reviewing a competitor video ad?
Record the advertiser, platform, library or ad identifier, observation date, visible opening, product demonstration, proof, offer, CTA, and any missing evidence. Label your interpretation separately from what is visible or audible.
Should I copy a competitor's ad that appears successful?
No. Extract a broad strategy such as demonstrating a product early, then create an original script, footage, and visual treatment for your own brand. Verify permissions before using any source material.
Official research sources
- Meta Ad Library and Meta's explanation of the library's coverage.
- TikTok Creative Center Top Ads and TikTok's Top Ads description and authorization note.
- Google Ads Transparency Center.
- Google Search Console performance reports for checking your own site's queries, clicks, and impressions.
Turn an authorized video into a reviewable brief.
Upload a video you have the right to analyze, choose Ad intelligence, and check the generated observations against the source. One on-screen analysis is free each day during early access.
Analyze an authorized video