E-commerce & DTC Brands · Scenario 34

Home & Furniture Online

Every platform claims credit for the same sales, so reported ROAS adds up to more revenue than the brand makes. We run incrementality tests to measure what each channel truly causes — and stop paying for sales we’d have made anyway.

Method · Incrementality testing

The situation

A furniture brand’s ad platforms each report a glowing ROAS, but they are all claiming the same buyers — and much of that revenue would have happened without the ad. Branded search and retargeting look spectacular precisely because they catch people already intending to buy. The brand scales winning channels that are really just taking credit.

Last-click and platform-reported numbers cannot separate correlation from causation, so budget pours into channels with high reported ROAS but low real, incremental impact.

Double-counted
reported ROAS
Claiming credit
for existing demand
Last-click
misleads budget
No causal test
of true lift

Where we dig for the truth

We run controlled incrementality experiments — geo holdouts and lift tests — to measure the sales each channel actually causes, not just the ones it claims.

Spend & reported ROASGeo-level sales dataHoldout test resultsBranded vs non-brand splitConversion pathsMargin per order
Reported ROAS vs true incremental ROASWhat each channel claims vs what it actually causes0x3x5x8x11x9x1.2xBranded search7.5x2.1xRetargeting3.8x2.9xPaid social2.6x2.4xNon-brand search1.4x1.9xConnected TVReportedIncremental

Branded search and retargeting report huge ROAS but cause little incremental revenue — they harvest demand that already existed. The real drivers look modest but are genuinely additive.

Our approach — Incrementality (Geo-Lift) Testing

We design geo holdout and lift tests: comparable regions or audiences are withheld from a channel, and the difference in sales reveals its true incremental effect. Each channel earns an incremental ROAS, and budget moves from demand-harvesting channels to the ones that actually create new customers.

The brand stops paying a premium to take credit for sales it would have made anyway, and reinvests in genuinely additive demand generation — measured, not assumed.

From reported ROAS to causal lift1Design the testSet up geo holdouts andaudience liftexperiments.2Measure the liftCompare exposed vswithheld to isolatetrue causation.3Score incrementalityGive each channel areal incremental ROAS.4Re-allocateFund what createsdemand; trim what justharvests it.
Profit after re-allocating to incremental channelsSame budget, aimed at true lift$0k$40k$81k$121k$161kM1M2M3M4M5M6

Once budget follows incremental rather than reported ROAS, the same spend produces measurably more real, profitable sales.

What changes

Same budget, validated by experiment. Representative for an online home & furniture brand.

Representative 90-day movementWasted on harvesting~35%~12%▼ -23 ptsIncremental ROAS1.9x2.8x▲ +47%New-customer share46%63%▲ +17 ptsContribution profit$100k/mo$140k/mo▲ +40%
Where the gain comes from+$40kmonthly profitCut demand-harvesting spend42%Reinvest in true demand gen36%Higher new-customer rate22%
Why this is not "social media management"
We didn't trust the platforms grading their own homework. We ran controlled experiments to measure what each channel actually causes, and stopped paying premium prices for sales the brand would have made anyway. Incrementality is the difference between reported and real ROAS.

Frequently asked questions

How do you measure the true ROAS of advertising?
We run incrementality experiments — geo holdouts and lift tests — where comparable regions or audiences are withheld from a channel. The difference in sales reveals what that channel actually caused, not just the sales it claimed credit for.
What is incrementality testing?
Incrementality testing is a controlled experiment that isolates the causal effect of marketing: it measures the extra sales that would not have happened without the ad. It separates demand-harvesting channels from genuine demand creators.
Why are platform-reported ROAS numbers misleading?
Platforms grade their own homework and over-claim, especially branded search and retargeting, which catch buyers already intending to purchase. Incrementality cuts through it. Book a marketing audit.

Want this run on your numbers?

Send your channel and sales data and we’ll design the lift tests that reveal your true ROAS.

Book a marketing audit
Back to E-commerce & DTC Brands