Tracking video ad performance on TikTok comes down to three things: setting up the right reporting columns in TikTok Ads Manager, watching a small set of metrics in the right order, and reviewing creative on a schedule instead of only when a campaign underperforms. Most advertisers either track too little (just spend and ROAS) or too much (every column TikTok offers), and both approaches make it hard to tell whether a video ad is actually working or just getting lucky with placement.
Why tracking video ad performance on TikTok is different from tracking a static ad
A static image ad either gets clicked or it doesn't. A video ad has a lifecycle: it gets seen, it gets watched for a few seconds, it gets watched further, and somewhere in that journey a person decides to tap through or scroll past. That means a video ad can fail for different reasons at different points, and the fix is different depending on where it fails.
If people aren't watching past the first couple of seconds, the problem is the hook. If they watch most of the video but don't click, the problem is the offer or the call to action. If they click but don't convert, the problem is the landing page, not the ad. Tracking video ad performance on TikTok properly means you can tell these three failure points apart instead of just seeing "ROAS is low" and guessing.
Set up TikTok Ads Manager reporting the right way
Before you can analyze anything, your reporting view needs to show the right columns side by side. Most accounts default to a view built for a media buyer checking spend, not a creative person checking why an ad works.
- Open the Reporting or Campaign view and use the Customize Columns option rather than the default preset.
- Add video-specific metrics: 2-second video views, video views at 25%/50%/75%/100%, and average watch time.
- Add performance metrics next to them: CTR (destination), CPC, cost per result, and ROAS if you're running purchase campaigns.
- Save this as a custom view/template so every report you pull going forward uses the same columns in the same order.
- Set your date range to compare like with like, for example this week versus last week, not an arbitrary 30-day window that blends old and new creative.
Once this view is saved, checking performance becomes a two-minute habit instead of a research project. This is the single biggest time saver for anyone tracking more than a handful of ads.
The core metrics that matter (and the ones to ignore for now)
TikTok Ads Manager offers dozens of columns. Most of them are noise for creative decisions. The table below shows what to actually look at, what it tells you, and what to do when a number is off.
| Metric | What it tells you | Rule of thumb to test | If it's weak |
|---|---|---|---|
| 2-second video views / impressions | Whether the hook stops the scroll | Treat this as your hook benchmark and compare new ads against your own best performer, not an industry number | Rewrite the first 2-3 seconds; test a new opening visual or line |
| Watch time at 25%/50%/75% | Where attention drops off inside the video | Look for the steepest single drop between two checkpoints | Trim or re-cut the section just before the biggest drop |
| Destination CTR | Whether the message and call to action are convincing enough to click | Compare across your own ads over time, since it varies a lot by niche | Tighten the offer or make the CTA clearer and earlier |
| Cost per result / CPA | Whether clicks are actually turning into the result you're paying for | Compare to your break-even CPA for the product, calculated from margin | Check the landing page before blaming the ad |
| Frequency | Whether the same audience is seeing the ad too often | Watch for a rising trend over time rather than one fixed number | Refresh creative or expand the audience |
Notice that ROAS isn't the first thing on this list. ROAS is the scoreboard, but it doesn't tell you what to change. The metrics above tell you exactly where in the video people lose interest, which is what you actually need to fix the ad.
Build a weekly video ad performance review
Checking performance daily during a launch makes sense. But most accounts waste time staring at numbers that haven't had enough spend or time to be meaningful. A weekly review, done on the same day each week, is usually enough to catch what matters.
- Pull your saved custom report for the last 7 days and the previous 7 days side by side.
- Sort by spend first, so you review the ads that matter most, not the ones with the most interesting-looking charts.
- Flag any ad where hook rate (2-second views) dropped compared to the prior week, even if ROAS still looks fine, since this is usually an early warning sign of creative fatigue.
- Flag any ad where CTR is strong but CPA is weak, which usually points to the landing page or offer rather than the video.
- Pause or refresh anything with rising frequency and falling hook rate together, which is the clearest sign an audience has seen the ad too many times.
- Note what's working in a simple running doc: hook style, video length, whether it used a founder-style delivery, whether it had on-screen captions, and so on.
That last step matters more than most advertisers realize. If you only track performance at the ad level, you'll never notice that, say, your demo-first hooks consistently outperform your problem-first hooks across ten different products. For a deeper breakdown of this, see how to track video ad performance by hook.
Track performance by creative element, not just by ad
An individual TikTok ad is a mix of several decisions: the opening hook, the length, the pacing, the voiceover or captions, and the call to action. If you only log a win or loss per ad, you lose the ability to see which of those decisions actually drove the result. Tag every ad with a few attributes before you launch it, so your weekly review can group by them later.
- Hook type: question, bold claim, before/after, demo-first, founder-to-camera.
- Length: under 15 seconds, 15-30 seconds, over 30 seconds.
- Format: standard vertical ad, split-screen, UGC-style, motion poster/offer announcement.
- CTA placement: early (first half), late (final seconds), repeated.
Once you have even eight or ten ads tagged this way, patterns show up fast: maybe demo-first hooks hold attention better but founder-to-camera hooks convert better once people click through. That's a much more useful finding than "ad #14 did well." If you're also watching how attention fades across the video itself, pair this with how to read video ad retention graphs, since hold rate and retention shape tell you different things about the same video.
Common mistakes that skew your tracking
A few habits quietly make video ad tracking on TikTok less reliable than it should be.
- Judging an ad before it has enough delivery. A video that's only had a small amount of spend hasn't had a fair chance to show its real hook rate or CPA yet. Wait until it clears a reasonable spend threshold for your budget before drawing conclusions.
- Comparing new creative to zero instead of to your best ad. The real question isn't "is this ad profitable," it's "is this ad better or worse than what's currently running."
- Blending attribution windows across a report. If you change your attribution setting mid-campaign, older and newer numbers in the same report aren't really comparable.
- Ignoring video length when comparing hook rate. A 30-second ad and a 60-second ad will naturally show different percentage drop-offs at the same checkpoints; always check whether you're comparing videos of similar length. See video ad length for TikTok Spark Ads if you're testing different durations.
- Tracking performance but never acting on it. A dashboard is only useful if a weak hook rate actually triggers a new video, not just a note in a spreadsheet.
Where FrameNotion fits into this process
FrameNotion doesn't track or report on ad performance inside TikTok Ads Manager, so your tracking setup above still runs entirely through TikTok's own reporting. What FrameNotion speeds up is the part that comes after you spot a weak hook or a tired creative: making the next version. You paste a product or landing page link, and FrameNotion AI writes and renders a 30-second 9:16 video ad from scratch, built for TikTok, Reels and Shorts, usually in about 10-20 minutes.
Because every ad is generated fresh rather than pulled from a template, you can brief a new hook idea, a shorter cut, or a different proof point and get back a distinct video to test against your current best performer. Once an ad is live and you've logged how it performed, you can request text or color changes without starting over, or generate A/B hook variants on the Pro and Agency plans to test the exact variable your weekly review flagged. Browse a few finished examples on the examples page, or see how the process starts from a URL in how to create a video ad from a landing page URL.
The first ad is free on sign-up with no card required, which is enough to test this loop once end to end: pull your weekly report, identify the weak metric, generate a new video aimed at fixing it, and compare it against the old one next week.
Frequently asked questions
What's the minimum spend before I should trust a video ad's performance data on TikTok?+
There's no universal number, since it depends on your daily budget and cost per result, but treat any ad with very little delivery as inconclusive. A practical rule of thumb to test is waiting until an ad has spent roughly enough to generate a meaningful sample of clicks or results for your product before deciding to pause or scale it.
Should I track video ad performance by campaign or by individual creative?+
Both, but for different reasons. Campaign-level numbers tell you whether your overall spend is profitable. Creative-level tracking (hook rate, watch time by checkpoint, CTR) tells you why a specific ad is winning or losing, which is what actually helps you make the next ad better.
Why does my video ad have a good hook rate but low ROAS on TikTok?+
A strong hook rate with weak ROAS usually means the problem sits after the hook: the offer isn't convincing enough to make people click, or people are clicking but the landing page isn't converting them. Check CTR and the landing page before assuming the video itself is the issue.
How often should I check video ad performance on TikTok?+
A light daily glance during a launch is fine, but most meaningful decisions should come from a weekly review once an ad has had time to accumulate enough delivery. Checking too frequently often leads to reacting to normal day-to-day noise rather than real trends.
Does FrameNotion track how my ads perform on TikTok?+
No. FrameNotion creates the 30-second vertical video ad itself; it doesn't connect to TikTok Ads Manager or report on performance. You track performance inside TikTok's own reporting, then use what you learn to brief and generate the next version of the ad.