The biggest challenge facing scaling ecommerce brands today isn’t budget constraints or creative fatigue, it’s attribution clarity. Without understanding what’s truly driving incremental revenue across channels, even sophisticated brands find themselves making budget decisions based on incomplete data, leading to wasted spend and suboptimal growth outcomes.
At BARK, we regularly see this challenge firsthand when working with global ecommerce brands. But whilst many think the complexity is purely technical, it’s also highly strategic, requiring a fundamental shift in how we think about measurement and growth.
The Attribution Gap
Modern ecommerce measurement faces a crisis across four critical areas that directly impact your ability to scale profitably:
Platform-Centric Reporting Creates Tunnel Vision
Each platform’s native reporting exists in isolation, showing only fragments of the customer journey. For most of our clients, overlapping attribution means that if you add up all the conversions across channels, the total exceeds actual sales by 50-150%. Comparing in-platform KPIs to inform the media mix is, therefore, highly problematic.
Many tout Multi-Touch Attribution (MTA) as the solution. While MTA can track user journeys across interactions with different channels, and even measure some cross-device pathways, there are inherent issues that are inescapable. Mainly: not everyone clicks. So even if you measure every click, over a lengthy attribution window, for all users, you still miss a huge chunk of the picture. But what about calibrated MTA, where you use real-world experimentation and calibration to improve accuracy?
Calibrated MTA Is Only Useful When all Other Conditions are Equal
We encourage calibrating MTA with experiment results, MMM insights and post-purchase survey data. With post purchase surveys adding the most unique extra value. The big caveat however, is that the bias in MTA numbers for different channels changes massively under different conditions. For instance, during promotional periods like Black Friday or major sales events, attribution patterns become heavily distorted. Customer behaviour changes, conversion volumes spike, and the typical customer journey compresses – all of which skew how credit gets distributed across channels.
Calibrated MTA is most trustworthy when comparing performance across similar, stable periods. When conditions shift significantly, the calibration loses its effectiveness, and you’ll need alternative measurement approaches to make sound budget allocation decisions.
Misattribution Leads to Misinvestment
Without unified measurement systems, and an appreciation of the limitations of all data sources, even data-driven brands are often chasing a red-herring. During high demand periods, attribution models suddenly tell a different story. Demand driving channels receive too little spend, an opportunity to grow is missed, and ROAS is much higher than it needs to be to fully capitalise on the available demand. We’ve also seen brands inadvertently cut their highest-performing activities simply because they look bad on last-click reporting. The negative effect can often be delayed, digging a hole that can be slow to recover from.
Is Incrementality the Solution?
A customer clicking on a Facebook ad before purchasing doesn’t definitively mean the ad caused the purchase. They might have converted through organic channels regardless. This is the issue of ‘incrementality’. The critical question every ecommerce brand must answer: Would this conversion have happened anyway?
Measuring that is no mean feat, as the measurement bias changes whenever you target different audiences (e.g. new versus existing customers), use different optimisation goals (e.g. clicks, conversions, awareness) or promote different products. But without measuring true incrementality, you’re likely to over-invest in channels that look effective but aren’t actually driving additional sales- and under-invest in channels that are genuinely growing your business.
For instance, in our work with global accessories retailer Pom Pom London to implement holdout group testing – the gold standard for measuring true incremental impact – by comparing results from customers who saw ads versus those who didn’t, we discovered a 94% increase in incremental ROAS that was completely invisible in both MTA systems and platform reporting. That’s a colossal efficiency and significant unfair advantage for the brand that was only discoverable with advanced, ‘ground truth’ measurement techniques.
Advanced Attribution Solutions
Multi-Touch Attribution: The Partial Solution
Multi-Touch Attribution (MTA) systems attempt to distribute conversion credit across multiple touchpoints, providing a more nuanced view than last-click attribution, like we often see in ad platforms or traditional analytics tools. MTA offers valuable insights into customer journey patterns and can improve budget allocation compared to basic models.
However, MTA faces inherent limitations. It can only track users who remain identifiable across touchpoints, missing the substantial portion of customers who browse privately or across devices. Additionally, MTA shows interaction sequences but doesn’t definitively prove causation, leaving the incrementality question unanswered.
Marketing Mix Modeling: The Strategic Foundation
Marketing Mix Modeling (MMM) takes a fundamentally different approach, using statistical analysis to isolate the true impact of each marketing channel while accounting for external factors like seasonality, promotions, and competitive activity.
Modern MMM solutions like Meta’s Robyn can process real-time data and provide actionable insights for budget optimisation. The key advantage lies in MMM’s ability to measure true incrementality by comparing actual performance against modeled baselines that account for what would have happened without specific marketing activities.
This approach reveals important insights that traditional attribution misses. For example, when brands experience significant drops in “organic” conversions during paid social pauses, MMM can quantify how much of that organic traffic was actually influenced by paid social touchpoints that conventional tracking couldn’t capture.
Crucially, whilst availability to MMM tools is on the rise, be cautious with any out of the box solution. MMM and other attribution tools don’t tell you what should be working, they just show what is. Without expert guidance, and proper understanding of the tactical and strategic inputs in your advertising that drive the results, they can quietly steer your spend in the wrong direction and cost you dearly. Be wary! More on this here.
Practical Implementation Framework
Build a Robust Measurement Infrastructure
Deploy MMM solutions alongside platform-native tracking to create multiple measurement perspectives. Implement geo-based tests that compare performance in matched markets with different media approaches. Regular conversion lift studies validate incremental impact across channels, giving gold-standard experimental results to anchor and calibrate other systems around.
Establish Cross-Channel Performance Integration
Create systems where insights flow seamlessly between channels. High-performing social creative concepts should inform search ad copy and landing page development, while emerging search query data can unlock new social targeting opportunities.
Optimise Budget Allocation Strategically Across the Full Funnel
Use MMM insights to reallocate investment toward truly incremental activities. Breaking free from last-click reporting often reveals that search campaigns scale more effectively when supported by strategic, upper-funnel brand awareness investment, while lower-funnel activities efficiently convert this expanded awareness into revenue.
Account for External Variables
Consider factors beyond digital attribution that influence conversion rates: seasonality, product stock levels, sales, search trends, competitor activity, PR campaigns, or retail presence. These elements significantly impact performance and cross-channel measurement but often remain invisible in digital-only measurement approaches.
Real-World Impact Example
Our exclusive co-branded research with Meta reveals how proper attribution measurement transforms business outcomes. Working directly with Meta and Pom Pom London, we rigorously tested Meta’s new Incremental Attribution feature using gold-standard Conversion Lift Studies and modern MMM.
The results challenge conventional campaign measurement, uncovering a 57% lift in ROAS that was completely missed by traditional attribution methods. Download the report to read the detailed methodology and results, and how you can implement it for yourself.
Building Your Attribution Advantage
The brands achieving sustainable, profitable growth aren’t necessarily those with the largest budgets, they’re the ones with the clearest measurement frameworks and deepest understanding of true incremental impact across their growth ecosystem. This is why we’ve made advanced data science an integral part of our integrated growth teams. It’s the essential component for your agency partner to have in their offering.
Your competitive advantage lies not just in running better campaigns, but in measuring what truly matters for sustainable business growth.
