Report finds traditional ad measurement falling short, calls for MMM to reveal true ROI

A new report from TikTok, Kochava, and Kantar argues that the advertising industry’s long reliance on Last Touch Attribution (LTA) is no longer sufficient for measuring true campaign effectiveness, especially in an era of privacy regulations and fragmented media consumption. The study, titled “Beyond Last-Touch: How Media Mix Modeling (MMM) Fills the Gaps of LTA to Unlock Predictive ROAS Measurement,” explores how Media Mix Modeling can offer a more comprehensive view of return on ad spend (ROAS) across the entire consumer journey.

According to the report, traditional attribution models often overemphasize short-term conversions while overlooking earlier brand-building interactions that influence final purchase decisions. “LTA is outdated — it only credits the last click,” the report notes, citing findings that many consumers take days or weeks to make buying decisions. Kantar’s data shows that only 5%–25% of consumers convert immediately after seeing an ad, depending on the industry, with the majority requiring additional time and information before acting.

TikTok emerged as one of the top-performing media platforms in the study, ranking #1 globally in ad equity in Kantar’s Global Media Reactions 2024 report. The platform also played a leading role in brand awareness and purchase consideration, with seven in ten TikTok users saying they had purchased a new product or brand after seeing content on the platform, according to a joint study by TikTok and WARC.

The report attributes TikTok’s effectiveness to its ability to deliver “high-quality attention” — attention that is immersive, emotional, and socially reinforced. This type of engagement, the study says, enhances brand perception and recall more effectively than traditional ad channels.

To better capture such effects, the report advocates for wider adoption of next-generation Media Mix Modeling. Unlike traditional MMM, modern solutions integrate real-time data and machine learning to provide dynamic insights. In one example, research from TikTok’s partner Kochava found that MMM attributed 106 day-seven revenue events per day to TikTok, compared with 74 detected via LTA, demonstrating how traditional methods may undervalue higher-funnel activity.

The study concludes that combining LTA and MMM offers the most reliable measurement strategy — using LTA for short-term, user-level insights and MMM for holistic, cross-channel performance analysis. This dual approach, TikTok and Kantar argue, helps marketers optimize budgets, reveal hidden ROI, and make smarter, privacy-compliant investment decisions.