Your ROAS looks great — but is it actually driving growth?

An ecommerce company hires your PPC agency to explore paid search. A solid plan follows, and after approval, the campaigns go live. Soon, you’re seeing stellar results: high conversion volumes and a healthy ROAS.

On the surface, the strategy is a resounding success.

But look closer.

Some of these conversions might have occurred anyway via direct or organic search traffic — meaning the campaigns may not be driving real growth. Too often, this goes unmeasured.

To truly understand performance, you need to look at incremental lift and marginal ROAS.

The truth about ROAS

Perhaps you’ve heard about eBay’s paid search experiment? They were spending heavily on brand PPC ads. Then they ran a controlled test, turning those ads off for a portion of users to measure impact.

Organic traffic picked up most of those conversions, with minimal impact on revenue. But guess what? Despite the clear results, eBay turned the branded ads back on. Fear, or smart? You tell me.

With search becoming increasingly automated, and the customer journey spreading across more surfaces than ever, attributing conversions to the right channels is harder than ever. Advertising platforms are quick to claim credit for these conversions, but be skeptical.

What most platforms report is attributed return, not causal lift. In other words, ROAS tells you how much revenue the platform says it influenced; it doesn’t tell you how much of that revenue would have happened without the ads.

When it comes to black-box automation like Performance Max and Advantage+, platforms have become exceptionally good at one thing: finding the path of least resistance to a conversion. They aren’t necessarily finding new customers. They’re often just becoming the most expensive touchpoint in a journey that was already destined to convert.

Without measuring incrementality, automation simply amplifies non-incremental signals, such as:

• Brand search campaigns capturing existing demand.
• Retargeting campaigns hitting users who were seconds away from purchasing.
• Reporting that makes “safe” channels appear more valuable than they truly are.

Dig deeper: Paid media efficiency: How to cut waste and improve ROAS

Incrementality tells you whether marketing created something extra

Incrementality is causal lift — what changed because the campaign existed, typically measured by comparing exposed groups with holdout or control groups. So what did this campaign actually drive that wouldn’t have happened otherwise?

Even though you may not want to admit it, this is a much more useful lens for budget allocation than platform attribution alone.

A channel can have a fantastic in-platform ROAS and still generate a weak incremental impact. Why? Because it might be harvesting demand rather than creating it.

If you want to know whether a campaign genuinely drove growth, the better question is incrementality.

But it’s still not the full answer.

To decide what to do next, you also need marginal ROAS.

Dig deeper: Why incrementality is the only metric that proves marketing’s real impact

---

Marginal ROAS tells you what to do next

A channel may be incremental. But that still doesn’t tell you where the next $10,000 should go. That’s a marginal ROAS question.

Marginal ROAS measures the return on the next unit of spend, not the average return across all spend. Here’s how it works: the first tranche of budget often performs well, then the next performs worse.

Keep going, and the final dollars become dramatically less efficient than the average suggests. The same applies to CPA metrics: a blended CPA may look acceptable, while the last dollars spent were far less efficient, leaving many advertisers bidding beyond where they should.

Imagine you spend $10,000 and generate $50,000 in revenue (500% ROAS). You decide to scale and spend an additional $5,000. This extra spend generates only $5,000 in additional revenue.

• Your new average ROAS: 366% 
• Your marginal ROAS: 100% (You essentially traded $1 for $1.)

In this scenario, the last $5,000 you spent was entirely wasted, even though the total “average” performance still looks decent on your dashboard.

This is the trap of average ROAS. It makes a channel look scalable when it may only be efficient at lower spend levels, and it hides the difference between profitable core demand capture and weak incremental expansion.

To make better decisions, you need to look further. Platform ROAS helps with in-platform optimization, incrementality shows whether campaigns actually created value, and marginal ROAS tells you whether more budget should go there.

A strong ROAS can signal true efficiency, or it can mean the platform is capturing demand that would have converted anyway. That’s why you should focus more on incrementality tests.

Don’t ask whether the channel has been efficient. Ask whether the next dollar is efficient enough — that’s what determines smart scaling.

Dig deeper: The marketing measurement flywheel: A 4-step framework for proving impact

Options for incrementality testing

You don’t need a perfect measurement lab before you start. Geo tests, holdouts, audience exclusions, and controlled spend reductions can all teach you more than another month of attribution debates.

• Geo-split testing: Divide your markets into two comparable geographic groups, keep your ads running in the “test” group, and turn them off in the “control” group. The difference in total revenue between the two regions reveals the true incremental lift of your ads.
• Search lift tests (holdouts): Use platform tools to create holdout groups, a small percentage of users who are intentionally not shown your ads. By comparing their behavior to the exposed group, you can see the direct impact of your (for example) Search or YouTube campaigns.

Beyond these, you can also test the impact of remarketing, branding, awareness campaigns, or additional social channels.

The real shift: From reporting performance to allocating capital

Too many marketing teams still use measurement to explain what happened. The better use of measurement is to decide what should happen next.

Incrementality helps you understand whether a channel created value. Marginal ROAS helps you understand whether more investment is justified. Together, they move marketing measurement out of the reporting function and into capital allocation.

ROAS tells you who gets credit. Incrementality tells you what actually moved. Marginal ROAS tells you where the next budget should go. But be aware: incrementality is not the same as attribution. Attribution tells you who, or which channel, should get the credit, while incrementality shows you whether or not it was worth it.

Dig deeper: How to take your marketing measurement from crawl to sprint

View the full article