How to Actually Measure TikTok Ad Effectiveness (Hint: Not In-Platform)

TikTok has become a legitimate ad platform. Brands are spending $50K-$500K/month on it. But almost no one can tell you with confidence whether that spend is working, because TikTok's own measurement is deeply flawed.

The core issue: TikTok is a video-first, awareness-first platform, but it reports conversions like it is a direct response channel. That mismatch creates measurement chaos.

TikTok's View-Through Problem

TikTok's default attribution window includes 7-day click-through and 1-day view-through. The view-through piece is where it gets messy. If someone watches 2 seconds of your ad (which counts as a "view" on TikTok) and then buys your product within 24 hours, TikTok claims credit.

The problem is that TikTok shows ads to a lot of people. Millions of impressions per day for mid-sized campaigns. Statistically, many of those people were going to buy anyway. They saw a 2-second flash of your ad while scrolling, did not engage, and converted through a completely different channel. But TikTok counts it.

In our experience, TikTok over-reports conversions by 2-5x compared to what MMM shows as incremental. One DTC brand we analyzed was seeing 5.8x ROAS in TikTok Ads Manager. MMM measured it at 1.4x. Still positive, but a completely different investment thesis.

Why Standard Attribution Fails for TikTok

TikTok operates differently from Google or Meta. Most of its value is in brand building and demand generation. Someone watches your product review ad on TikTok, does not click, and three days later searches your brand on Google and buys. Google Search gets the credit in last-click. TikTok gets zero.

So last-click undervalues TikTok. But TikTok's own reporting overvalues it. You are stuck between two wrong numbers. This is exactly the situation where MMM shines: it measures the relationship between TikTok spend changes and business outcome changes, regardless of who clicked what.

What MMM Actually Shows for TikTok

Across the accounts we have modeled, TikTok ROAS from MMM typically lands between 1.0x and 3.0x, with a median around 1.8x. That is significantly lower than what TikTok reports (usually 3-8x) but often higher than what last-click shows (usually 0.3-1.0x).

The other finding: TikTok has a strong adstock (carryover) effect. The impact of TikTok ads decays slowly, over 2-4 weeks. This makes sense because TikTok is a brand awareness channel at its core. You see an ad today, it influences your purchase next week or the week after.

This means weekly spend fluctuations on TikTok are less effective than on search. Consistent spend over time outperforms the approach of cranking budget up and down based on daily platform ROAS.

How to Set Up TikTok Measurement Right

Step 1:Implement TikTok Events API (server-side tracking). This improves signal quality for TikTok's algorithm, which helps campaign optimization even if you do not trust its attribution.

Step 2:Set click-only attribution in TikTok Ads Manager for your internal reporting. Remove view-through from your dashboards. This won't change how TikTok optimizes (it still uses all signals), but it gives you a less inflated view.

Step 3:Run MMM with TikTok as a separate channel. Do not lump it with "social" or "paid social." TikTok's adstock decay is different from Meta's, so the model needs to estimate it independently.

Step 4: Watch the branded search lift. One indirect way to validate TikTok impact: when TikTok spend goes up, does branded search volume go up 1-2 weeks later? If yes, TikTok is generating demand that converts through search.

With Spendmix, TikTok is modeled as a separate channel with its own adstock and saturation curves. See how it shows up in a real MMM report.