Cross-channel · Media Planning

What Is Marginal ROAS (and Why It Changes How You Allocate Budget)

ROAS — return on ad spend — is the metric most marketing teams live by. Total revenue divided by total spend. A campaign returning $5 for every $1 spent sounds like a success story. Often it is. But average ROAS can mask a deeply inefficient allocation. A channel with a 5x average ROAS can simultaneously be destroying value at the margin — if the last dollars you're pouring into it are returning less than $1.

Marginal ROAS is the metric that catches this. It measures the return on your next dollar of spend, not your average dollar. Understanding the difference changes how you think about scaling, cutting, and moving budget across channels.

Average ROAS vs. Marginal ROAS

The distinction is straightforward in theory but easy to overlook in practice.

Marginal ROAS = Incremental Revenue from Additional Spend ÷ Incremental Spend

In a simple example: you spend $10,000 on Google Search and generate $50,000 in revenue (5x average ROAS). You increase spend to $12,000 and revenue rises to $51,500. The marginal ROAS on that $2,000 increment is $1,500 ÷ $2,000 = 0.75x.

The overall campaign looks healthy. The incremental spend is actively losing money. This gap — between a strong average and a weak margin — is one of the most common misread signals in performance marketing.

Why This Happens: Diminishing Returns

Marginal ROAS declines because of diminishing returns — a structural feature of every advertising channel, not an anomaly. The first dollars you spend on a channel reach your most receptive audience at the most efficient auction prices. As you spend more, you exhaust that segment and move to progressively harder-to-convert, more expensive audiences.

This produces an S-shaped or concave response curve: revenue rises steeply with early spend, then the slope flattens as spend grows, then eventually becomes nearly horizontal — you're spending more for almost no additional return.

The Saturation Trap

A channel in the flat part of its response curve is saturated. It may still report a strong average ROAS because the early efficient spend is baked into the average. But the marginal return on new spend is close to zero. Continuing to scale into a saturated channel at the same rate is the most common form of wasted ad budget — and it's almost invisible in standard reporting because the blended ROAS stays high.

The counterintuitive implication: a channel with a lower average ROAS may deserve more budget than one with a higher average ROAS, if it's earlier on its response curve and its marginal return is still strong. Budget decisions made on average ROAS alone systematically over-invest in saturated channels and under-invest in channels with remaining headroom.

The Correct Allocation Rule

If your goal is to maximize total revenue from a fixed budget, budget is optimally allocated when marginal ROAS is equalized across all channels. If Channel A has a marginal ROAS of 3x and Channel B has a marginal ROAS of 1.5x, moving a dollar from B to A increases total revenue — and you should keep moving dollars until the marginal returns converge.

Optimal allocation: equalize marginal ROAS across all channels at your total budget.

This is the theoretically correct rule. In practice, it requires knowing the marginal ROAS of each channel at its current spend level — which means measuring the response curve, not just reporting on the average. That measurement problem is the hard part.

How to Measure Marginal ROAS

There is no single clean way to measure marginal ROAS — every method involves a tradeoff between precision, cost, and time. The three main approaches:

1. Incrementality Testing (Geo Holdouts)

Split your market into test and control regions. Hold spend constant or reduce it in the control group, increase it in the test group, and measure the revenue difference. Incrementality testing gives you a causal estimate of what additional spend actually drove — not what attribution models credit it with. A 4–8 week geo holdout test is the most reliable method for measuring true marginal return at a specific spend level, but it requires careful market segmentation to avoid confounds.

2. Marketing Mix Modeling (MMM)

MMM uses regression analysis on historical spend and revenue data to model the response curve for each channel — including saturation effects. Once the model is fit, you can read off the marginal ROAS at any spend level. The advantage is that it covers all channels simultaneously without requiring holdout experiments. Saturation curves are a direct output of MMM and are the standard input for data-driven budget allocation at scale. The tradeoff: MMM requires 1–2 years of clean historical data and meaningful spend variation across channels to produce reliable estimates.

3. Bid-Level Analysis (Channel-Specific)

For channels with granular bid or spend controls — Google Ads, Meta, programmatic — you can approximate the marginal return curve by analyzing performance at different spend or bid levels over time. In Google Ads, the Bid Simulator provides a direct view of projected clicks and conversions at different Max CPC levels, which is effectively a point estimate of marginal return per keyword. Across channels, this requires more manual data work but is accessible without a full MMM build.

Practical Signals That Marginal ROAS Is Falling

You don't always need a formal model to detect diminishing returns. Several operational signals point to a channel approaching saturation:

Marginal ROAS and Multi-Channel Budget Allocation

The budget allocation problem across channels is precisely a marginal ROAS problem. At any given total budget, you want to find the split across channels such that moving a dollar from any channel to any other channel doesn't increase total revenue — meaning marginal ROAS is equal across all active channels.

In practice, this requires:

  1. A response curve (or at minimum, a marginal return estimate) for each channel at its current spend level
  2. A solver that finds the budget split where marginal returns converge
  3. An assumption about which channels have remaining headroom vs. which are saturated

For a deeper look at how this framework applies to cross-channel planning, see How to Allocate Marketing Budget Across Channels. If you want to run the allocation math against your own historical data, Media Budget Allocator fits a saturation curve per channel from your spend and revenue history, then finds the budget split that maximizes total blended ROAS at your specified total spend. Free, no login required.

FAQ

Is marginal ROAS the same as incremental ROAS?

They're closely related but not identical. Incremental ROAS measures the revenue driven by a campaign that wouldn't have happened without it — separating organic from paid. Marginal ROAS measures the return on the next unit of spend, assuming the campaign exists. In practice, a well-run incrementality test gives you a marginal ROAS estimate at the spend level tested, so the two concepts converge when the test is designed correctly.

Can I calculate marginal ROAS from standard reporting data?

Not directly — standard reports show average ROAS, not marginal. You can approximate it by comparing performance across periods where spend changed significantly, but this conflates spend changes with seasonality, creative changes, and competitive shifts. A cleaner estimate requires either a holdout test, MMM, or platform-specific tools like the Google Ads Bid Simulator for keyword-level estimates.

What's a good marginal ROAS threshold for scaling?

The right threshold depends on your contribution margin. If your product has a 40% gross margin, you need a marginal ROAS above 2.5x to cover cost of goods — anything above that generates incremental profit. Below your breakeven marginal ROAS, additional spend destroys value even if the average looks fine. Calculate your breakeven point (1 ÷ contribution margin) and use that as the floor for marginal ROAS when deciding whether to scale.

Should I use marginal ROAS or target ROAS for Google Ads smart bidding?

These operate at different levels. Your Target ROAS smart bidding setting tells Google's algorithm what average ROAS to optimize toward at the campaign level. Marginal ROAS is a planning metric you use to decide how much total budget to put into that campaign in the first place. Set your Target ROAS based on your performance goals; use marginal ROAS to decide whether to increase or decrease the campaign's daily budget cap.

References

Find the right split for your budget

Upload your historical spend and revenue data across channels. Set your total budget. Get the allocation that maximizes blended ROAS — accounting for where each channel's returns start to flatten.

Open Media Budget Allocator →