Azzam Bilal Chamdy
For years, the direct-to-consumer (DTC) ecosystem has been locked in a recursive, often frustrating debate: which marketing platform deserves credit for a sale? This "attribution anxiety" has become the default state for growth-stage brands. Because Meta, Google, and TikTok all utilize proprietary models that prioritize their own influence, advertisers are routinely faced with conflicting data. A single transaction is frequently claimed by three different channels, leading to a fragmented view of reality.
However, the industry is reaching a critical inflection point. As privacy regulations tighten and signal loss becomes the new normal, the search for the "perfect" attribution model is being replaced by a more sophisticated approach: the integration of incrementality testing within a holistic, business-centric measurement framework.
The Flawed Logic of Isolated Incrementality
Incrementality testing—the practice of running lift studies to determine the causal impact of a specific channel—has been touted as the "silver bullet" for budget allocation. The logic is deceptively simple: run a lift study, identify which channels generate demand versus those that merely harvest it, and shift budget accordingly.
Yet, this framing is dangerously incomplete. Many growth-stage brands have fallen into the trap of using these tests as a blunt instrument. They observe a low incremental lift for upper-funnel channels—like Meta awareness campaigns—and decide to slash the budget. The predictable, often painful, result is a subsequent drop in total business revenue.
The missing link is the understanding of the "customer journey ecosystem." A customer might view a Meta ad on Monday, bypass the click, perform a direct search on Wednesday, and convert via a branded Google search. Meta reports a view-through; Google claims a last-click conversion. A lift study on either channel in isolation fails to capture the synergy between them. The Meta impression primed the consumer; the branded search closed the sale. Cutting either one risks collapsing the entire funnel.
The Three-Layered Measurement Stack
To resolve this, sophisticated marketing organizations are adopting a three-layer measurement stack, where each layer serves a distinct diagnostic purpose:
- The Business Anchor (Marketing Efficiency Ratio – MER): At the foundation lies MER, calculated as total revenue divided by total ad spend. This is the only metric that ignores channel-level ego and views marketing as a unified investment. It provides the "truth" for CFOs and founders.
- The Diagnostic Layer (Incrementality Testing): This layer provides the "why." If MER fluctuates, incrementality tests (lift studies, geo holdouts) explain which specific levers caused the change.
- The Operational Layer (Platform Attribution/ROAS): While often inaccurate in isolation, this layer remains useful for day-to-day tactical optimizations—provided it is kept subservient to the higher-level signals of MER and incrementality.
Chronology of the Measurement Evolution
The transition toward this multi-layered approach did not happen overnight.
- 2020–2022 (The Attribution Era): Brands were obsessed with Multi-Touch Attribution (MTA). With the arrival of Apple’s iOS 14.5 update, the industry lost the granular tracking that fueled these models, leading to widespread panic.
- 2023–2024 (The Rise of Incrementality): Advertisers pivoted to lift studies and conversion experiments as the primary solution to signal loss. Many, however, made the mistake of treating these tests as absolute verdicts rather than contextual signals.
- 2025 (The Holistic Maturity): The current landscape is defined by the democratization of causal inference tools. With the release of open-source frameworks like Google’s Meridian and the refinement of Bayesian statistical methodologies, the barrier to entry for high-level measurement has dropped significantly.
Supporting Data and Methodologies
The cost of running incrementality tests has fallen, making them accessible to a wider range of brands. In 2025, four primary methods have emerged as the standard for data-driven allocation:
1. Platform-Native Lift Studies
Meta and Google now offer integrated lift testing. Google’s transition to Bayesian statistical modeling has allowed for lower-budget studies (effective with as little as $5,000 and 1,000 conversions). However, these tests remain limited to the platform’s "walled garden," meaning they cannot capture the cross-platform synergies that drive total MER.
2. Geo Holdout Testing
For brands with a large geographic footprint, geo-matched markets remain the gold standard. By pausing spend in specific regions while maintaining it in others, brands can observe the revenue gap. This method is statistically cleaner than user-level tracking but requires rigorous matching of baseline performance and seasonality.
3. Spend-Down Testing
This is perhaps the most practical method for measuring MER sensitivity. By cutting a channel’s budget by 50–75% for a defined window, the business can observe the ripple effect on total revenue. If revenue drops proportionally, the channel is likely driving core demand. If revenue holds, the channel may be redundant.
4. Full Causal Inference Models
At the pinnacle are Marketing Mix Models (MMM). Google’s Meridian has brought these complex data science models to the open-source community. While powerful, they require significant data science overhead and are often overkill for mid-market brands that can achieve 90% of the value through simpler testing cadences.
Official Perspectives: The Shift in Strategy
Major platform documentation has evolved to reflect this shift. Google’s 2025 Ads Highlights emphasizes that conversion lift now works at lower spend levels, acknowledging that smaller brands are being forced to take measurement into their own hands.
Conversely, Meta has raised the barrier for some of its Brand Lift products, requiring a minimum of $120,000 in spend for certain studies in the U.S. This signals a strategic split: platforms are bifurcating their measurement offerings, pushing larger enterprises toward deep brand-equity studies, while providing lighter conversion-based testing for smaller, growth-oriented advertisers.
The Implications for Marketing Strategy
The primary implication of this new measurement landscape is a move away from "binary" marketing decisions. Growth-stage brands must stop asking, "Should we cut this channel?" and start asking, "How does this channel influence our blended efficiency?"
The Quarterly Testing Rhythm
For brands spending between $100,000 and $1 million monthly, a quarterly rhythm is essential:
- Month 1: Execute a geo-holdout test on your largest channel.
- Month 2: Run a spend-down test on a secondary channel to validate its incremental contribution.
- Month 3: Synthesize the findings against the quarterly MER trend to reallocate the following quarter’s budget.
Conclusion: Closing the Measurement Gap
The debate over which platform provides the most "accurate" reporting is a distraction. Attribution describes the journey; incrementality measures the sensitivity; and MER dictates the financial reality.
The brands that win in 2025 and beyond will be those that integrate these three elements into a unified system. When a CMO looks at a dashboard, they should no longer be looking for the "right" ROAS number. Instead, they should be looking at their MER as the heartbeat of the business, using incrementality studies as the diagnostic tool to keep that heart healthy.
By accepting that no single test provides the full picture, brands can finally move past the attribution trap. They can stop guessing where to allocate their next dollar and start investing with the confidence that they understand the true impact of their marketing on the bottom line. The tools are available, the methodologies are proven, and the era of relying on platform-reported "vanity metrics" is effectively over. The future of growth is, and always has been, in the business-level outcome.
Lina Irawan
Laily UPN
