How TikTok counts results
A TikTok recap shows the results TikTok credits to its own ads. Like every ad platform, TikTok uses its own attribution rule and only sees what its own tracking can see. It doesn’t know about your email, your Meta ads or the Reddit thread someone read the night before.
The numbers on top are usually views, watch time and engagement. Those tell you how the ads landed with the people who saw them. Results come lower down, often as one number that blends people who tapped with people who only watched. A view-through conversion counts a result for someone who saw the ad and never tapped. Find the window in the report: how many days after a view or tap a result still counts.
TikTok recaps also love engagement rate. It’s a useful number for creative, and it isn’t a result. Ask what the rate is divided by, because views, reach and followers give very different answers.
Where results get counted twice
- Boosted creator posts. When paid spend goes behind a creator’s video, its results can show up in your creator report and your paid-media report. Two reports can count the same post. This is the most common overlap on TikTok.
- Creator codes. A creator code from a TikTok video can spread to coupon sites, where people who were already buying find it.
- First touch, last credit. TikTok is often where someone first hears of you. They may buy a week later through search or email, and TikTok’s view rule can still give TikTok the credit.
- Views at scale. TikTok serves a lot of views. Even a small share of viewers who buy for other reasons adds up to a big view-through number.
The fix starts with one list. Take your own numbers for the month, then mark each result with every report that counted it. Results with two or more marks are the overlap. TikTok usually shares a lot of them with creators and Meta, because the same people see all three in the same week. Once everyone can see the overlap, the budget conversation gets shorter and friendlier.
A checklist for your next TikTok recap
| Look at | What to look for | Why it matters |
|---|---|---|
| Window | Tap and view windows behind each result | View rules reach far on a platform this big |
| Taps vs. views | Results split by tap and by view | Shows how much rests on views alone |
| Boosted posts | Creator videos with paid spend | Two reports can count the same post |
| Codes | Code uses by source | Coupon sites inflate code results |
| Engagement rate | What it’s divided by | Views, reach and followers give different rates |
| Comments | What people said under the ads | Tells you how the audience took it |
Questions worth asking
Ask them of whoever runs the account, in-house or a partner.
- Can we see results split by tap and by view, with the window for each?
- Which results came from boosted creator posts, and do they also show up in our creator reports?
- How many code uses came from coupon sites?
- Is engagement rate divided by views, reach or followers?
- What did people say in the comments, and did it change the creative?
The comments are worth reading on their own. TikTok audiences say what they think, and they say it fast. Our guide to listening that counts shows how to turn comments into numbers you can use.
What a regions-on, regions-off test looks like on TikTok
Lining up the reports removes double counting. It can’t tell you whether TikTok caused results. For that, choose regions that look alike, run TikTok ads in some and pause them in others for a few weeks, and compare sales in your own numbers. That’s a regions-on, regions-off test, also called a holdout test.
Paid ads can be paused by region. Organic creator posts are harder, because they travel wherever the algorithm sends them. If creators are a big part of your TikTok spend, test the paid boost on their posts by region, and judge the creators themselves on codes, links and timing against your own numbers. Our guide to creator campaigns covers that side.
Decide before you start what you’ll do with each outcome. A test that ends in “interesting” is a test nobody acts on.
Want to see what worked for your brand? Try it free: pick a question, and we answer it with your own numbers and what people say online. No card.