Est. reading time: 6 minutes
TikTok’s Ads Manager and Google Analytics will tell you two different stories about the same campaign, and brands that only read one of them make predictable mistakes. Reading TikTok alone means trusting a platform reporting on itself, with view-through conversions and modeled data padding the numbers. Reading GA4 alone means undercounting a channel whose contribution often shows up two clicks later. The integration between them isn’t a product you install. It’s mostly UTM discipline plus a clear understanding of what each system is structurally unable to see.
What each system sees, and what it can’t
TikTok’s reporting runs on the TikTok Pixel and Events API, which observe what happens after people interact with your ads, including impressions that never produce a click. GA4 runs on your website’s tag, which only knows about sessions that actually arrive, carrying whatever labels the URL brought with them.
Three gaps follow from that design, and none of them are bugs. GA4 cannot see view-through conversions at all, so TikTok will always claim purchases GA4 credits elsewhere. TikTok’s in-app browser and iOS privacy restrictions strip tracking from a meaningful share of clicks, so some TikTok sessions land in GA4 looking like direct traffic unless the URL itself carries identification. And the two systems run different attribution logic and windows, so even the conversions both can see get credited differently. Expect a permanent gap, watch whether the ratio between the two reports stays stable, and stop trying to make the numbers match.
The UTM setup that gets TikTok ads into Google Analytics
UTM parameters are the entire bridge, because they identify traffic in the URL itself, which survives everything the in-app browser strips. Three rules make the difference between clean data and a mess.
First, choose values GA4 actually recognizes. Set utm_source to tiktok, lowercase, always, since GA4 treats tiktok and TikTok as different sources. Set utm_medium to a value GA4’s default channel grouping maps to a paid channel, such as paid_social or cpc, because medium is what decides the channel bucket, and an unrecognized value drops your entire TikTok spend into Unassigned where standard reports effectively hide it. If TikTok traffic lands in Unassigned anyway, the fix is a custom channel group (Admin, then Data display, then Channel groups) with a rule matching source contains tiktok and your chosen medium, applied going forward.
Second, let macros do the labeling. TikTok supports dynamic URL parameters like __CAMPAIGN_NAME__ and __AID_NAME__ that populate the real campaign, ad group, and ad names at click time, so one URL template covers the whole account with no hand-typing and no typos. Build it once with utm_campaign on the campaign macro, utm_term on the ad group, and utm_content on the creative, and every breakdown you’ll ever want in GA4 labels itself. Two cautions from TikTok’s own documentation. Don’t combine a manual UTM string with the Auto-attach feature, since duplicated parameters make GA4 read one and discard the other unpredictably, and check the supported macro list before templating Smart+ campaigns, where the available parameters differ.
Third, tag at launch, not retroactively. Editing a live ad’s destination URL is an edit to a running ad, with all the delivery consequences that implies, so UTMs get added when creative ships or at a natural refresh point, never as a Tuesday afternoon cleanup across active campaigns.
Set the conversion layer on both sides
On the TikTok side, run the Pixel with the Events API and map the full event ladder, since that’s what powers both TikTok’s reporting and its optimization. If your conversions happen off-site, phone calls being the classic case for service businesses, the tracking has to be extended to capture them, the same problem we solved on the Meta side in maximizing ROI with Facebook ad call tracking.
On the GA4 side, the old Universal Analytics concept of goals is gone. GA4 tracks events, and you promote the ones that matter (purchase, generate_lead, sign_up) to key events, which is what makes them countable as conversions in acquisition reports and usable in attribution. Ecommerce stores should have the standard purchase event with value flowing, because revenue per source is the number that ends most internal arguments about whether TikTok is working.
Where to read it in GA4
The daily view is Reports, then Acquisition, then Traffic acquisition, with the dimension switched to session source/medium and tiktok on the row. Add session campaign as a secondary dimension and your macro-populated campaign names appear, with sessions, engaged sessions, key events, and revenue beside them. For creative-level reads, build an Exploration with session source, session campaign, and session manual ad content, which carries your utm_content values, meaning your ad names sit next to their revenue.
Two reads earn their place in a weekly review. The ratio of engaged sessions to sessions per creative is a traffic-quality detector, since an ad producing a thousand clicks and forty engaged sessions is winning the auction with clickbait and losing everywhere that costs money. And the Advertising workspace’s attribution path reports show where TikTok sits in multi-touch journeys, which matters because TikTok skews toward discovery, and judging it purely on last click systematically undervalues it right before someone cuts the budget. The GA4 side of this analysis goes deeper than one post, and we covered the segment-level work in using GA4 for better audience targeting and ad optimization, with the in-platform complement in our TikTok analytics deep dive.
If GA4 is showing almost no TikTok traffic at all, debug in order. Confirm the ads are actually delivering, which is its own diagnosis we covered in why your TikTok ads aren’t getting views, then click your own tagged URL and watch GA4’s Realtime report for a session arriving with tiktok as the source. A missing session there means the URL is malformed, and one detail catches people regularly, which is that TikTok drops everything after a # character in destination URLs.
Assign each report a job
Once the plumbing works, resist the instinct to crown one number. TikTok’s report, fed by the pixel, is the tool for in-platform decisions, meaning creative comparisons and audience calls, and it’s the data the delivery system learns from. GA4 is the tool for TikTok’s place in the channel mix, for post-click quality, and for the assisted-conversion story last click misses. Backend orders remain the ceiling neither report may exceed. Read trends rather than absolutes, investigate when the gap between systems moves rather than when it exists, and route what you learn into the testing cadence we laid out in strategies for optimizing TikTok ad performance. The integration isn’t finished when the data flows. It’s finished when each number has a decision it owns.
