What emerged from our conversation, and our joint research, reveals a fundamental reimagining of how brands should measure and optimise campaign performance.
The Problem with Traditional Attribution
“We developed the Incremental Attribution setting to address the limitations of traditional (often click-based) attribution models, which often fail to capture the true incremental impact of advertising,” explains Skokan. This limitation has profound implications: sophisticated advertisers have long relied on Conversion Lift experiments as their source of truth, but optimisation systems haven’t been aligned with this incrementality-focused approach.
The challenge is particularly acute for businesses engaging with immersive formats like video, where consumers don’t necessarily click through immediately. Traditional click-based measurement tools simply can’t capture the full customer journey in today’s complex media landscape.
A Machine Learning Breakthrough
This solution represents a fundamental shift in how campaign optimisation works. The Incremental Attribution setting uses machine learning models trained on global, cross-advertiser Conversion Lift data to optimise ad delivery for truly incremental conversions – conversions that wouldn’t have occurred without the ad.
“It’s generally a pretty hard problem to figure out which conversions are more or less incremental than others,” Skokan acknowledges. The key breakthrough came from leveraging AI and data from past Conversion Lift studies to bridge the gaps in pixel-based attribution, while continuously refreshing training data and updating feature sets to account for advertiser-specific attributes.
As of May 19, 2025, this setting is available to all advertisers through Ads Manager, the Meta Marketing API, and third-party platforms, making it remarkably accessible for brands ready to embrace incrementality-focused optimisation.
Real Results, Real Impact
Our joint research with Pom Pom London provides compelling validation of this technology’s potential. The project demonstrated significant improvements in both ROAS and incremental ROAS, aligning with broader Meta testing that shows advertisers using Incremental Attribution achieve an average 46% lift in performance.
“The project actually revealed that the traditional reporting methods actually significantly underestimate the improvements that were achieved by incremental attribution, highlighting the setting’s ability to uncover this hidden value,” notes Skokan.
Perhaps most importantly, this “hidden value” represents missed opportunities that many brands are leaving on the table by relying on outdated measurement approaches.
The Multi-Modal Measurement Advantage
What made our research particularly valuable was the multi-modal measurement approach we employed. By calibrating Conversion Lift studies, marketing mix modeling (using Robyn), and Ads Manager results, we created a comprehensive validation framework that demonstrates the true impact of incremental attribution.
“The results validation in MMM was a pleasant surprise in that it was net new data for our Product team,” comments Skokan on BARK’s insights. This alignment between optimisation systems and measurement methodologies provides a roadmap for how these systems can work together more effectively.
The Hidden Cost of Last-Click Attribution
The implications extend far beyond individual campaign performance. According to data from Meta, approximately 31% of incremental Meta conversions are misallocated to other channels based on non-incremental attribution models. This means businesses relying on last-click attribution for investment decisions may be systematically under-investing in channels that drive real value.
“Businesses who make investment decisions based on last-click may not be investing appropriately, as they don’t have insight into the full range of channels that are driving value,” Skokan explains. The solution involves moving to attribution models that incorporate additional touchpoints and exploring statistical methods that better reflect true incrementality.
Actionable Insights for Ecommerce Brands
Through our research and partnership with Meta, we’ve identified several immediate recommendations for ecommerce brands:
Optimise for incrementality, not just volume. Focus on incremental conversions rather than total conversions to achieve more meaningful business outcomes and better budget allocation.
Embrace automation with intention. In today’s era of AI-enabled platforms, ensure your optimisation inputs align with business objectives by leveraging insights from your measurement framework.
Test and validate. Implement Incremental Attribution as a standalone test, then explore combinations with value optimisation for outcomes that are potentially greater than a sum of their parts.
Calibrate your measurement stack. Use multiple methodologies, Conversion Lift, MMM, and traditional reporting, to validate and understand the full impact of your optimisation strategies.
The Future of Campaign Management
Looking ahead, Skokan envisions a fundamental shift in how the industry approaches campaign measurement: “We hope to shift the industry’s focus from traditional metrics to incrementality, encouraging marketers to prioritise strategies that drive true business growth.”
This transformation is part of a broader evolution toward more automated and optimised campaign management, where machine learning and AI enable unprecedented precision and personalisation at scale. The brands that embrace this shift early and partner with agencies equipped to navigate this complexity will gain significant competitive advantages.
Why This Partnership Matters
Meta rarely produces external-facing content with agency partners, making this collaboration particularly significant. It reflects BARK’s position as one of just a handful of global agencies that Meta has certified as best-in-class across all three essential performance capabilities: campaign strategy, data science, and creative strategy.
Our ability to conduct rigorous, multi-modal measurement research and to work directly with Meta’s product teams to validate new technologies, is representative of the kind of advanced capabilities that leading ecommerce brands need in today’s complex marketing landscape.
The shift toward incremental attribution isn’t just a technical upgrade; it’s a fundamental reimagining of how brands should think about marketing effectiveness. Brands that continue to rely on outdated measurement approaches risk not just suboptimal performance, but systematic misallocation of their marketing investments. In a world where costs are rising, the stakes are high.
The future of performance marketing is incremental. The real question is how quickly brands can adapt to take advantage of it.
Watch our full interview with Igor: Hear directly from Meta’s Marketing Science Director about the development and real-world impact of Incremental Attribution: Video Link
Want to dive deeper into the research? Read our detailed case study and white paper on the Pom Pom London project: Meta Incremental Attribution Case Study | BARK Whitepaper
