Why media mix modeling is becoming essential for predictive ROAS measurement

The challenge isn’t that last-touch attribution is wrong; it’s that it’s incomplete. And incomplete measurement inevitably leads to incomplete investment decisions. That doesn’t mean that today’s most effective marketers are abandoning last-touch attribution; they’re pairing it with next-generation media mix modeling to create a more comprehensive, privacy-safe and predictive approach for measuring return on ad spend. MMM doesn’t replace the tactical value of last-touch attribution. It fills in the gaps by capturing the full-funnel impact that the latter routinely overlooks. As mobile marketing accelerates into a privacy-centric future, this combined approach is quickly becoming the new baseline for comprehensively measuring media effectiveness.

Why last-touch attribution can’t keep up with today’s marketing reality

Advertisers are realizing the limitations of last-touch attribution. It doesn’t reflect how consumers actually move through the full funnel, how creative drives brand lift or how upper-funnel investments compound downstream lower-funnel performance.

The metric assigns all conversion credit to the last interaction a user has with a brand. In the earliest days of mobile app advertising, this linear model aligned closely enough with typical user behavior to be serviceable. However, the modern customer journey is anything but linear.

Users are influenced by dozens of moments: a creator’s video, a social trend, a brand exposure that lodges subconsciously or a series of impressions across multiple channels that prime behavior long before a final click. Last-touch attribution ignores all of this.

Modern MMM is no longer a retroactive model

Historically, MMM had a reputation for being slow, something marketers commissioned annually rather than something used for daily media decisions. Today’s MMM is fundamentally different.

For app marketers in particular, MMM is especially powerful because the privacy-first nature of only using aggregated market-level data allows it to better include nontrackable media into contribution calculations. By holistically covering media investments in models, enabling return analysis on longer timeframes without the bias of signal loss and building a more market-reflective response curve, MMM can more effectively calculate ROAS.

MMM is capturing the real value of TikTok in the decision-making journey

Across the globe, TikTok is reshaping how people discover brands, products and apps. It inspires first-time purchases, fuels deep brand affinity and influences every stage of the funnel. But these strengths don’t always show up in last-touch attribution, leading to an underreporting of TikTok’s incrementality.

Why? Because TikTok often initiates the journey rather than ending it. MMM, however, reveals the full story: TikTok’s real ROI is consistently and significantly higher than what last-touch attribution suggests. When early-stage engagement is accounted for, TikTok emerges as one of the most efficient and influential platforms for app growth.
When measurement captures the entire journey, and not just the last click, TikTok’s value becomes undeniable. According to a new report from Kochava, TikTok and Kantar, when measured via MMM, TikTok is about 35% more incremental on both iOS and Android than last-touch reporting. TikTok is also 11% to 16% more likely to be underreported than top competitors, despite being 17% to 20% more efficient at driving acquisition than those same competitors. The takeaway: As a full-funnel channel, TikTok’s contribution and broader incremental influence on driving conversions is most effectively captured through MMM.

How high-performing marketers combine last-touch attribution and MMM

Last-touch attribution and MMM aren’t competing philosophies. They’re the new power pair of modern measurement. Together, they deliver the speed marketers need and the accuracy the future demands. Brands already leaning into this dual-model approach are unlocking higher ROAS, more predictable growth and a far clearer view of how platforms like TikTok drive value far earlier in the conversion path.

The marketers winning today do three things differently. They blend last-touch and MMM, using short-term attribution and always-on incrementality in the same decision loop. They measure influence, not just last-touch, understanding how brand exposure affects conversion probability. Finally, they simulate budget shifts before spending, running scenario-based MMM to project real ROAS before campaigns even launch.

The future of measurement is holistic, privacy-safe and predictive

As signal loss accelerates, relying solely on user-level attribution is no longer viable. Measurement strategies must evolve to keep pace with consumer behavior and emerging privacy norms that favor aggregated, model-based approaches.

MMM offers that evolution. TikTok suggests proof of how much marketers have been missing. Paired together, MMM and last-touch attribution can reshape app marketers’ measurement strategies in 2026 and beyond by creating an accurate and actionable framework. The result is a future where ROAS is more predictable, budgets are more efficient and marketers finally have visibility into the true value of their investments.

Download the full report for free and explore how MMM and last-touch attribution can reshape your measurement strategy in 2026 and beyond.

Partner insights from Kochava