What Is Marginal ROAS? How to Find the Real Ceiling on Your Ad Spend
Marginal ROAS measures the return from your next dollar of ad spend, not the average. Learn how to calculate it, why it matters more than standard ROAS for budget decisions, and how media mix modeling makes it actionable.
Marginal ROAS measures the revenue generated by the next dollar you spend on advertising, not the average return across all dollars already spent. It answers a question that standard ROAS can't: "If I spend one more dollar on this channel, what will I actually get back?"
This distinction sounds minor but it changes how you allocate budget. A channel with a 5x average ROAS might only return $1.50 on the next dollar you put in. If your breakeven is $2, scaling that channel is losing you money, even though the average looks healthy. Marginal ROAS reveals where that tipping point is.
How Is Marginal ROAS Different from Standard ROAS?
Standard ROAS divides total revenue by total ad spend. If you spent $10,000 on Meta ads and generated $50,000 in revenue, your ROAS is 5.0. That number tells you how the channel performed overall, but it hides a critical pattern: diminishing returns.
Most advertising channels follow a curve. The first dollars you spend reach the most receptive audiences and drive the easiest conversions. As you spend more, you push into less responsive segments. Each additional dollar produces a little less revenue than the one before it.
Here is a simplified example:
| Spend Level | Total Revenue | Average ROAS | Marginal ROAS (on last $1K) |
|---|---|---|---|
| $5,000 | $30,000 | 6.0 | 6.0 |
| $10,000 | $50,000 | 5.0 | 4.0 |
| $15,000 | $62,000 | 4.1 | 2.4 |
| $20,000 | $68,000 | 3.4 | 1.2 |
Average ROAS at $20,000 in spend still looks like 3.4x. That seems fine. But the last $5,000 you spent only returned $6,000 in revenue, a marginal ROAS of 1.2. If your contribution margin is 30%, you need a ROAS of at least 3.3 to break even. That last chunk of spend is losing money.
How to Calculate Marginal ROAS
The formula is:
Marginal ROAS = Change in Revenue / Change in Ad Spend
Or more precisely: take two spend levels, measure the revenue at each, and divide the difference in revenue by the difference in spend.
If increasing your Google Ads budget from $8,000 to $10,000 per month drives revenue from $40,000 to $44,000:
Marginal ROAS = ($44,000 - $40,000) / ($10,000 - $8,000) = 2.0
That $2,000 increment returned $4,000. Whether that is profitable depends on your margins.
The Challenge: You Can't Just Eyeball It
The formula is simple but the inputs are hard to isolate. Revenue doesn't change solely because of ad spend. Seasonality, promotions, competitor activity, organic growth, and dozens of other factors all move the needle at the same time. If you bump Meta spend by $5,000 and revenue goes up by $15,000 that same week, you can't confidently attribute all $15,000 to the extra spend.
This is where media mix modeling comes in. MMM uses statistical models to isolate the incremental contribution of each channel while controlling for external factors. A well-built media mix model produces a response curve for each channel, showing the relationship between spend and incremental revenue. The slope of that curve at your current spend level is your marginal ROAS.
Why Marginal ROAS Matters for Budget Allocation
Most budget conversations happen in terms of average performance. "Meta is doing 4x, Google is doing 3x, so let's put more into Meta." That logic breaks down once you account for diminishing returns.
Consider a scenario with two channels:
- Channel A: Average ROAS of 5.0, but marginal ROAS at current spend is 1.5
- Channel B: Average ROAS of 3.0, but marginal ROAS at current spend is 3.5
If you're deciding where to put the next $10,000, Channel B is the better bet, even though its average performance looks worse. The optimal budget allocation is the one where marginal ROAS is equalized across channels. When every channel's next dollar produces roughly the same return, you've maxed out your efficiency.
This principle comes from basic economics (equalize returns at the margin), but applying it to marketing requires data that most teams don't have access to without proper modeling. A Nielsen study on marketing effectiveness found that about 25% of media channel investments were too high, meaning the marginal returns had already dropped below breakeven. Reallocating that overspend to underfunded channels could have improved total ROI by as much as 50%.
How to Measure Marginal ROAS in Practice
There are three main approaches, each with different tradeoffs:
1. Media Mix Modeling (MMM)
Media mix modeling is the most comprehensive approach. It fits a statistical model to your historical spend and revenue data across all channels simultaneously, producing response curves that show how revenue changes as you increase or decrease spend in each channel.
The output includes marginal ROAS at any spend level, which makes it directly useful for budget optimization. Modern MMM tools (including Formula) update these curves continuously as new data comes in, so your marginal ROAS estimates stay current rather than going stale.
Best for: Ongoing budget allocation across all channels, scenario planning, and understanding where you are on each channel's response curve.
2. Incrementality Testing
Incrementality tests measure marginal ROAS experimentally. You increase or decrease spend in a channel (or pause it entirely in certain markets) and compare results between test and control groups. The difference in revenue, divided by the difference in spend, gives you an empirical marginal ROAS.
Research from the Marketing Science Institute has shown that experimentally-measured incremental returns are often 30-50% lower than what platform-reported metrics suggest.
Best for: Validating a specific channel, calibrating your MMM, and proving causation rather than correlation.
3. Spend Variation Analysis
If you've naturally varied spend levels over time (seasonal ramps, budget cuts, tests), you can compare periods of high and low spend to estimate marginal returns. This is less rigorous than MMM or formal incrementality testing, but it can provide directional insight when you don't have access to either.
Best for: Quick directional estimates when you have clear spend variation in your historical data.
Common Mistakes When Using Marginal ROAS
Confusing average and marginal. This is the most frequent error. Teams see a 5x average ROAS and assume every dollar in that channel is producing 5x. The whole point of marginal ROAS is that it almost always diverges from the average, and the gap widens as you spend more.
Ignoring time lag. Some channels (particularly brand-building channels like TV, podcasts, and out-of-home) have delayed effects. The marginal return on this week's spend might not show up for weeks or months. Short measurement windows undercount these channels. This is one area where MMM has an advantage, since it can model lagged effects and carryover (sometimes called "adstock") explicitly.
Setting a single breakeven threshold. Your breakeven marginal ROAS depends on your contribution margin, which varies by product, customer segment, and lifetime value. A channel that looks unprofitable on a first-purchase basis might be highly profitable when you factor in customer lifetime value and acquisition cost. Calculating your true marketing ROI requires looking beyond the initial transaction.
Treating the response curve as static. The diminishing returns curve shifts. Creative refreshes, audience expansion, new ad formats, and competitive changes all alter the relationship between spend and revenue. A channel that was saturated six months ago might have more room today. Your marginal ROAS estimates need regular updates.
Setting Your Marginal ROAS Floor
Your marginal ROAS floor is the minimum return you need from the next dollar to justify spending it. It is directly tied to your unit economics:
Marginal ROAS Floor = 1 / Contribution Margin
If your contribution margin is 40%, your floor is 2.5x. Every dollar of ad spend needs to produce at least $2.50 in revenue to cover costs and contribute to profit. If your margin is 25%, the floor jumps to 4.0x.
Any channel where marginal ROAS has dropped below this floor is overfunded. Every channel where marginal ROAS is above this floor still has room to scale.
This is the framework that makes budget optimization quantitative instead of political. Instead of arguing about which channel "feels" like it's working, you're comparing marginal returns against a clear financial threshold.
How Formula Helps You Measure Marginal ROAS
Formula's media mix modeling platform produces channel-level response curves and marginal ROAS estimates from your actual spend and revenue data. Instead of guessing where you sit on the diminishing returns curve, you can see it, and run scenarios to understand what happens when you shift budget between channels.
The result is a data-driven marketing approach where budget decisions are grounded in how your next dollar performs, not how your last thousand averaged out.
Try Formula free and see your marginal ROAS across channels.