The Attribution Crisis: Why Traditional Measurement Fails Modern Marketing
By James Kirwan
Marketers are feeling the pinch of an increasingly turbulent economic climate. Earlier this year, Gartner revealed that the average marketing budget was 7.7% of total business revenue, compared to 9.1% in 2023, representing a 15% decrease. But fewer resources haven’t led to lower expectations. CMOs are under more pressure than ever to justify every pound spent—and many are struggling to do so.
A recent study found that less than a third of marketers can effectively measure ROI across both traditional and digital media, despite 85% believing they can. This discrepancy suggests that confidence may be masking a growing gap in measurement accuracy. And with more channels to contend with than ever before, both online and offline, the cracks are widening.
So, why is traditional measurement falling by the wayside, and how can marketers get back on track?
Granular Insights > Surface-Level Details
For years, last-click attribution has been the default method for measuring performance, with eMarketer reporting that more than three-quarters of marketers used it in 2024. Its appeal lies in its simplicity: by assigning credit to the final touchpoint before conversion, it should reveal which channel is most effective.
As a metric, last-click barely scratches the surface of the various factors that influence the whole customer journey. The brand-building activity that drives brand awareness, mid-funnel touchpoints that nurture intent and the multiple devices and channels that shape decision-making along the way—all cast aside in favour of the final point of conversion.
A potential drawback of this method is the misrepresentation of success rates, with channels that provide easily measurable metrics, such as pay-per-click (PPC), prioritised ahead of those that drive long-term growth. The traditional structure of agencies can exacerbate this issue further, when brand and performance are kept in separate siloes, championing their own channels—whether TV, search or social—in the hope of being allocated more budget and resources. This level of fragmentation isn’t conducive to efficiency.
From Observed to Actual Impact
Enter causal-based modelling: a system that emphasises incrementality.
Unlike traditional attribution models that simply show what worked, causal-based modelling goes one step further and illustrates why it worked. By looking beyond the conversion and determining the actual value of each channel, marketers can unlock more accurate forecasting—and, more importantly, funnel investment into the channels and strategies that drive real business growth.
In other words, uniting your marketing team behind causal-based modelling ends the constant clamour for credit, ensuring everyone is aligned and working towards common business goals.
While in-housing can often lead to the channel mix being dictated by internal politics, working with an independent agency amplifies these benefits. A third-party, therefore, can ensure that data and outcomes act as the north star for future activity.
Translating Theory Into Reality
At Tug, we have developed Real Impact, an advanced causal-based modelling product that can link previously disconnected datasets in near real-time. This AI model can measure any metric that matters – from revenue and purchases to visits and engagement.
We used Real Impact when working with Zipcar, one of the UK’s leading car-sharing services. Having previously over-indexed on Meta, diminishing returns had prompted Zipcar to look at diversifying its media mix. Our task was to demonstrate that Connected TV (CTV), a channel the business had no prior exposure to, could be a driver of measurable revenue impact.
Real Impact enabled us to connect two previously disconnected data sets: the CTV brand awareness campaign and Zipcar’s backend revenue. Not only were these on separate systems, but they had different metrics (such as impressions, reach and frequency), making direct attribution challenging, especially given the limitations of last-click or traditional attribution models.
By using Real Impact and causal modelling, we were able to isolate the impact of CTV alone, removing the influence of other marketing activities. This approach allowed us to measure the true incremental value and ROI of the campaign. Not only did this secure Zipcar’s ongoing commitment to CTV, it also provided confidence for the business to experiment with new channels, using causal-based modelling as the foundation.
Key Results:
- 3.4% incremental revenue growth
- 6.84 : 1 ROAS
Is your marketing department still relying on last-click attribution? Want to evolve beyond observed metrics and identify proven results? Get in touch with us—we’d love to demonstrate how Real Impact can help you prove and improve your ROI.
Related Articles:
