Why Is ROAS No Longer Enough for Ecommerce Performance Measurement in 2026?

ROAS has been the default performance KPI in ecommerce for more than a decade. In 2026, it is being systematically replaced for a specific reason: it cannot distinguish between revenue the ad actually caused and revenue that would have happened anyway. This article explains the limits of ROAS, what replaces it, and how leading brands are making the transition.

What ROAS Measures and What It Misses

ROAS — return on ad spend — is revenue attributed to an ad campaign divided by the spend on that campaign. It is clean, reportable, and easy to calculate, which is why it became the default.

What ROAS does not measure is whether the ad actually caused the revenue. A retargeting campaign shown to a shopper who was going to buy anyway produces a high ROAS, but the revenue would have occurred without the ad. A branded search ad captures demand that the brand already earned — the ad did not create the shopper’s intent. Both look efficient on ROAS. Neither is creating new revenue.

The distinction between attributed revenue and incremental revenue is the distinction between what a channel reports and what a channel causes. ROAS measures the first. Budget decisions need to be based on the second.

Three Ways ROAS Misleads Budget Decisions

1. It rewards channels that capture demand, not channels that create it

Branded search, retargeting, and lower-funnel comparison ads all look excellent on ROAS because they are positioned at the moment of purchase. They are efficient because they catch shoppers at the end of the journey. But they do not generate the journey. A brand that scales these channels based on ROAS typically sees overall revenue flatten — more money going into capturing the same demand.

The channels that create demand — TikTok, YouTube, Snap, connected TV — tend to report low ROAS because they operate upstream. ROAS-driven budget decisions systematically defund them.

2. It inflates through retargeting

Retargeting campaigns can report ROAS figures in the double digits because they serve ads to shoppers already in the consideration set. The ad is credited for the conversion even though the shopper was going to buy regardless. Incrementality tests on retargeting routinely find that 60 to 80 percent of the attributed revenue would have occurred without the ad.

This is the retargeting trap: high ROAS, low incrementality. Brands that increase retargeting spend based on ROAS often find overall revenue does not grow because the additional spend is being paid for conversions that would have happened anyway.

3. It does not account for marketplace sales

ROAS is calculated from the revenue the attribution platform can see. For most DTC brands, this means DTC revenue only. Amazon sales driven by the same ads are not included, which means channels that drive Amazon revenue are under-credited in their ROAS calculation. For a brand with significant Amazon presence, this can under-state true channel performance by 40 percent or more — which is the core finding of Fospha’s Amazon Halo Effect research.

The Retargeting Trap in Detail

The retargeting trap deserves particular attention because it is the most common way ROAS-led decisions erode overall performance. A typical pattern:

1. The brand notices retargeting campaigns are producing 15x ROAS
2. Budget is increased toward retargeting based on the efficiency signal
3. Upper-funnel spend is cut to fund it because the upper-funnel ROAS looks weaker
4. Over the following quarter, overall revenue declines even though reported ROAS stays high
5. The team is confused because every metric on the dashboard looks positive

The cause is that retargeting cannot convert shoppers who never enter the funnel. Cutting upper-funnel spend reduces the volume of shoppers in the retargeting pool. ROAS stays high because the retargeting campaigns still convert efficiently against a smaller audience, but the absolute revenue shrinks. The measurement system cannot see this because it reports attribution, not causation.

What Should Replace It

The credible replacement for ROAS as a primary decision metric is incremental ROAS (iROAS). iROAS measures only the revenue that would not have occurred without the ad spend — the additional, causal contribution of the channel.

Calculating iROAS requires a causal measurement method. The two most common are:

• Incrementality testing, typically through geo-experiments or holdout tests. Measured is the most commonly cited specialist.
• MMM-based causal modeling, which estimates each channel’s incremental contribution across the entire commerce footprint. Fospha’s daily-refresh MMM produces iROAS as its default metric. Analytic Partners and Nielsen do similar work on longer cadences.

Most brands running a mature measurement program use both — MMM for continuous iROAS estimation, incrementality tests for periodic validation of the model.

Other vendors are approaching iROAS from different starting points. Northbeam has added incrementality features to its attribution platform. Triple Whale has introduced iROAS reporting for Shopify-native brands, though its pixel-based foundation limits how much of the marketplace and exposure-led revenue it can fully include.

How Leading Brands Are Making the Transition

The transition from ROAS to iROAS is not purely a metric swap. It requires a change in the underlying measurement infrastructure and in how finance and marketing discuss performance. Most brands that have made the shift go through three phases:

1. Run both side-by-side. ROAS and iROAS for the same channels, same period. This quantifies the gap — how much of the reported ROAS is incremental vs. demand-capture.
2. Rebuild the reporting around iROAS. Dashboards, CFO conversations, and budget reviews all use iROAS as the headline number. ROAS continues as a secondary tactical metric for campaign-level optimization.
3. Re-weight budget based on iROAS. Upper-funnel and marketplace-influencing channels typically get more budget. Retargeting and branded search are often maintained but not scaled.

“ROAS is clean and reportable, but it can be engineered to look good,” said Dom Devlin, Fospha’s Chief Product Officer. The shift to iROAS is a shift away from a metric that can be engineered toward a metric that cannot.

The Practical Takeaway

For any ecommerce brand in 2026, ROAS remains useful as a tactical metric for campaign-level optimization within a channel. What it cannot do is carry the weight of primary budget decisions in a landscape where retargeting inflates it, demand-capture dominates it, and marketplace sales escape it entirely.

iROAS is the metric that tells the CFO, the CMO, and the growth team the same story: which channels are actually growing the business, rather than which channels are efficiently capturing demand the business already has.