Blog
Conquer, Scale and Thrive in Retail Media
Learn more
Can any marketing leader today afford to let media budget go unaccounted for? Not likely. Yet with the extended impact of signal loss, consumers’ cross-device consumption habits and continually increasing channel fragmentation, it’s harder than ever to prove media performance.
Digital media measurement is at an inflection point.
Many marketers are still relying on misleading “legacy” metrics that don’t reflect how media actually works today. To make the most out of ad spend, marketers need a new measurement playbook.
These are the top three to master to set your brand up for success.
TLDR; iROAS measures how many conversions happened because of the advertising versus what would have happened anyway.
Return on ad spend (ROAS) has been a gold standard in media measurement, but it isn’t telling you the full story. ROAS measures return from all conversions attributed to media—but not all conversions are caused by media.
This metric therefore inflates your performance when:
ROAS assumes all attributed revenue is caused by your advertising. But what about customers who would have purchased anyway? What about the impact of organic search, word-of-mouth or other marketing activations? Traditional ROAS often gives advertising credit for sales that would have happened regardless.
Incremental ROAS (iROAS) is a conversion lift metric that isolates the lift generated specifically by your media—the conversions that would not have happened without the ad exposure. This metric separates correlation from causation by measuring all marketing efforts against your total revenue. iROAS offers a more accurate picture of whether media is actually driving performance by linking impression costs to downstream actions (like purchases).
Conversion lift measures the incremental impact of advertising efforts on conversions, or in other words, the increase in number of conversions that can be attributed to media within a campaign or channel in your control versus factors outside of your control. Several methodologies exist, but the most accurate way is compare outcomes between a test group that was exposed to a brand’s ads and a control group that was not exposed (or even more rigorous: shown unbranded ads). By isolating the effect of media exposure, conversion lift helps marketers understand whether advertising actually changed behaviour or simply captured demand that already existed.
So what does this mean in practice for marketers?
While many walled gardens offer conversion lift reporting (which may seem compelling), those results often come with limited transparency and little insight into how lift was calculated and what actually performed well. On the other end of the spectrum, many other DSPs frequently rely on third-party lift vendors and cookie-based signals, which can introduce gaps in coverage and confidence.
Epsilon is one of the only non-walled gardens that offers intensive conversion lift measurement conducted natively within our platform to deliver brands with transparency and objectivity.
TLDR; eCPM allows marketers to evaluate effectiveness across channels on an equal playing field.
In the current digital media environment, CPM is also a misleading metric. It measures the cost per thousand impressions but not the quality of those impressions. As a result, CPM can reward cheap inventory that doesn’t reach real people or isn’t actually in view on the consumers’ screen.
Effective CPM (eCPM), on the other hand, adjusts CPM to be more accurate using performance factors like:
eCPM is therefore a more accurate reflection of what it really costs to reach a true, qualified consumer, which is becoming increasingly important today. With more media running across fragmented channels (CTV, mobile, in-app, retail media networks), cost surfaces are harder to compare.
TLDR; Unique reach is a true measurement of how many distinct people an ad reached.
The problem with standard reach is that the metric counts every impression, even if they come from the same person. This leads to inflated scale and misleading “coverage” metrics.
Unique reach is the new best practice for digital media measurement. It’s a true measurement of how many distinct people an ad reached. Relying on unique reach over traditional reach helps fight frequency waste, oversaturation and ad fatigue.
However, measuring unique reach can be harder than it sounds. It requires deterministic identity that can accurately find consumers and measure their ad impressions across devices and channels.
But many platforms rely on probabilistic modelling rather than deterministic identity. This results in both significant gaps (e.g., logged-out impressions that can’t be confidently tied back to a known consumer) and duplicate counts (e.g., failing to match multiple devices to the same consumer and counting them multiple times).
Epsilon COREid’s deterministic identity helps brands reach and accurately measure more unique people rather than the same people repeatedly—improving efficiency and scale.
It’s no surprise that media budgets will continue to be under intense pressure. To meet targets and secure precious dollars, marketers need to measure performance with metrics that reflect real impact, not superficial activity.
iROAS, eCPM and unique reach help your team shift toward causality, effectiveness, efficiency and identity-driven measurement. These metrics aren’t just “nice to have”—they’re table stakes for smarter planning and reporting moving forward.
