You’re Paying for Customers Who Were Already Going to Buy

An incremental customer is one your ad actually made — someone who wouldn’t have bought without seeing it. The customers you can’t claim that for were already on their way: they’d visited your site, searched your brand, or were going to find you regardless. Most DTC attribution stacks treat both groups identically, which is why ROAS can look healthy while customer acquisition is flat. On Vibe.co, holdout-based incrementality testing is built into the platform across all campaign goals — not a third-party requirement, but part of standard campaign reporting.

What is an incremental customer?

An incremental customer is one whose purchase was caused by the ad — not just preceded by it. Attribution tools track touchpoints; incrementality isolates causation. The question attribution answers is “which channels did customers interact with before buying?” The question incrementality answers is “would those customers have bought anyway?”

The gap between those two answers is significant. When brands run holdout tests — suppressing ads to a random sample of their audience and comparing purchase rates — many find that a substantial share of attributed revenue would have been generated organically. The customers in that group weren’t created by the campaign. They were claimed by it.

Incremental ROAS (iROAS) is the metric that captures this: incremental revenue generated divided by ad spend. It’s almost always lower than reported ROAS. It’s also the number that correlates with actual growth, because it only counts the revenue your advertising actually caused.

Why does ROAS reward the least incremental spend?

ROAS is highest on the audiences closest to buying — and those audiences are the least incremental by definition.

Retargeting your own site visitors: these people already found your product. Many will convert regardless of whether you serve them another ad. Last-touch attribution credits the retargeting campaign; holdout testing often shows the incremental lift is modest. Branded search works the same way — you’re capturing people who already know your name, already decided to look for you. Both channels look excellent in attribution dashboards because you’re optimizing for credit capture on audiences already in motion.

This creates a trap. As DTC brands mature, they naturally shift budget toward what “works” in ROAS terms — retargeting, branded search, loyalty audiences. Reported ROAS climbs. New customer acquisition slows. The attribution stack isn’t lying; it’s answering the wrong question.

Which channels generate the most incremental customers?

The most incremental channels are those reaching people who haven’t found you yet. That’s a simple definition with meaningful implications for budget allocation.

Retargeting (lowest incrementality). Reaching your own site visitors and lapsed customers converts warm demand efficiently — but it’s mostly converting demand you already created. The incremental population in a retargeting audience is smaller than reported ROAS implies.

Branded search (lowest incrementality). If someone searches your brand name, they already know you exist. You’re capturing demand, not generating it.

Social prospecting (moderate incrementality). Prospecting campaigns reach new audiences, but as brands mature and lookalike pools saturate, the cost of finding genuinely new buyers rises and incremental returns diminish.

Streaming TV (highest incrementality for most DTC stacks). CTV delivers to households that haven’t found you yet — there’s no retargeting bias built into the channel the way there is on platforms where your pixel has been tracking users. Household-level delivery means you’re reaching people in a shared living-room context, including the large share who have shifted screen time away from social. Audience targeting on Vibe includes lookalike modeling, intent signals, and CRM integrations — the same precision as social, applied to audiences social hasn’t saturated. Blindster tracked a $45 CPA from CTV compared to $89 on Meta, measuring both channels with the same attribution tools. See the Blindster case study.

This isn’t an argument to eliminate retargeting. It’s an argument for understanding which channels are creating customers versus which ones are converting them.

How does Vibe’s incrementality testing work?

Vibe’s holdout-based incrementality testing is built into the platform across all campaign goals — not a third-party tool or a managed-service request.

When you enable incrementality testing on a campaign, Vibe designates a randomized holdout group from your target audience — typically 10–20% — that doesn’t receive ad delivery. Both groups (exposed and holdout) are tracked throughout the campaign window. At the end of the test, Vibe compares the purchase rate between the two groups. The difference is the incremental lift your campaign actually drove.

From there, iROAS is calculated against your ad spend: how much incremental revenue did the campaign generate above what the holdout group produced on their own? This number is lower than reported ROAS — that’s not a failure, it’s accurate measurement.

Knix, a DTC apparel brand, runs Vibe’s incrementality testing specifically to report iROAS to their CMO — scaling month-over-month spend based on incremental data rather than attributed ROAS. For brands already using Haus for incrementality measurement, Vibe integrates directly. Measurement and reporting on Vibe also connects to Triple Whale and Northbeam, so CTV performance appears in the same attribution dashboard as your paid social and search spend.

What does this mean for your DTC ad budget?

If most of your spend is concentrated in retargeting and branded search, you’re heavily indexed on channels that capture existing demand — and lightly indexed on channels that create new demand. A holdout test will almost certainly show that your reported ROAS overstates true incremental impact. That’s normal. The goal isn’t to make iROAS match reported ROAS — it’s to understand which channels and audiences are genuinely generating new buyers.

Channels that look weak in last-touch attribution are often more incremental than they appear. CTV shows up late in attribution windows, or not at all on last-touch models. But a holdout test isolates the lift it drove — independent of where it lands in the attribution chain.

If holdout testing shows a meaningful share of your retargeting revenue is non-incremental, you have budget to redeploy toward channels reaching genuinely new audiences. CTV best practices covers how to structure a streaming campaign for measurable incremental outcomes. Review the full CTV measurement guide for a breakdown of how attribution and incrementality work together in practice.

Pricing on Vibe starts at $50/day — a low enough entry point to run a meaningful incrementality test before committing significant budget.

FAQ

What is an incremental customer in marketing?

An incremental customer is one whose purchase was caused by seeing an ad — someone who wouldn’t have bought without that exposure. The contrast is a non-incremental attributed customer: someone the ad appeared before but didn’t cause, because they were already going to purchase. Most attribution models treat both groups identically, which means reported ROAS often overstates the revenue advertising actually generated. Incrementality testing uses holdout groups to isolate the true causal lift.

What is the difference between incrementality and attribution?

Attribution tracks which touchpoints preceded a purchase. Incrementality measures whether those touchpoints caused it. A customer who clicked a retargeting ad and then bought is attributed to that campaign. Whether they would have bought anyway — without seeing the ad — is an incrementality question. The gap is significant: brands running holdout tests frequently find that a meaningful share of attributed revenue would have occurred organically.

How do you measure incremental customers?

The standard method is a holdout test: suppress ads to a randomized 10–20% of your target audience and compare their purchase rate to the exposed group. The difference is incremental lift. iROAS = incremental revenue / ad spend. This runs at the campaign, channel, or audience level — each holdout test tells you how much lift that specific spend actually drove. Vibe runs holdout-based incrementality testing natively, built into the platform across all campaign goals.

Which advertising channels generate the most incremental customers?

The most incremental channels are those reaching audiences that haven’t previously engaged with your brand. Retargeting and branded search are the least incremental — those audiences are already in motion. Social prospecting is more incremental but diminishes as audiences saturate. Streaming TV is typically the most incremental channel in a DTC stack because it reaches households that haven’t found you yet, with no built-in retargeting bias. Channels that look weak in last-touch attribution often show meaningful incremental lift in holdout tests, because they’re doing the work of creating customers earlier in the funnel.