A channel can appear strong on average ROAS while becoming weak at the margin. Saturation curves help identify that before spend is committed - which is why they sit at the heart of forecast-led media planning rather than after-the-fact reporting.
Most budget mistakes are not made on the channels that obviously fail. They are made on the channels that look healthy in a dashboard while quietly burning the last slice of budget at a loss. The blended average hides it; the curve exposes it. Below we walk through what a saturation curve is, why paid media refuses to scale in a straight line, how to read a channel’s budget ceiling from its curve, and how ElenIQ models all of this so you can forecast the impact of a budget change before committing the spend.
What is a saturation curve?
A saturation curve is the relationship between how much you spend on a channel and the result that spend produces. Plotted out, it is not a straight diagonal line - it is a curve that rises steeply at first and then bends towards the horizontal. The early, cheap spend reaches the most responsive audience and converts efficiently. As you keep adding budget, the curve flattens, because each additional pound is reaching people who are less likely to act.
The shape matters more than any single point on it. Two channels can deliver an identical result at today’s spend yet have completely different curves: one still climbing steeply with plenty of headroom, the other already flat and close to its limit. A reporting average treats them as equal. The curve treats them as opposites - and the second is the one that will disappoint the moment you scale it. This is closely related to the Hill function, the mathematical form most commonly used to describe diminishing media returns, and to adstock, which captures how advertising effect carries over from one period to the next.
Why media does not scale linearly
The intuitive assumption - double the budget, double the result - is almost never true in paid media, and the reason is structural rather than accidental. Every auction-based channel reaches its most valuable, lowest-cost audience first. Those are the in-market buyers, the warm retargeting pools, the high-intent search queries. Once that finite pool is exhausted, the platform has only one way to spend more of your money: bid into broader, colder, more expensive inventory.
So as spend rises, frequency on the same users climbs, audiences widen to people with weaker intent, and the price of each incremental impression goes up. The result that each new pound buys shrinks even as total volume keeps growing. That is diminishing returns in action, and it is why the spend-to-result relationship bends. Understanding this is the difference between marginal ROAS and average ROAS: the average looks backwards across all your spend, while the margin tells you what the very next pound is actually worth.
What is a budget ceiling?
A budget ceiling is the spend level beyond which a channel stops producing worthwhile incremental return. It is not a hard cap set by the platform - you can always spend more. It is the point on the saturation curve where the next pound returns less than your efficiency target, so that pound is better deployed on another channel that still has headroom.
Crucially, the ceiling is defined by the margin, not the average. A channel can sit comfortably above its blended ROAS target while its next increment is already underwater, because the average is propped up by all the efficient spend that came before. Treating the headline average as permission to scale is exactly how teams pour budget past the ceiling without noticing. The ceiling is also why the same total budget, allocated differently across channels, can produce materially different results - the art is keeping every channel below its own ceiling rather than maximising any single one. You can test that trade-off directly with the budget allocation simulator.
How marginal returns flatten near saturation
Reading a saturation curve from left to right, the slope at any point is the marginal return - the result you get from one more pound at that level of spend. Near the origin the slope is steep, so marginal return is high. As you approach saturation, the slope flattens towards horizontal, and marginal return collapses towards zero even though cumulative volume is still rising.
Put numbers on it. Suppose a Meta campaign spends £50,000 at a 4.5 ROAS. Push it to £80,000 and total revenue still rises - but that extra £30,000 might only return at 2.0. The channel is still profitable on average, yet the marginal efficiency of the new spend has more than halved. The gap between the headline 4.5 and the marginal 2.0 is the saturation problem in miniature - and it is invisible to anyone watching only the blended average.
This is why scaling a winning channel so often disappoints. The headline efficiency that justified the increase was an average earned at lower spend; the increase itself buys you the flat part of the curve. The honest question is never “what did this channel return last quarter?” but “if I add the next £10,000 here, how much genuinely new result will it create?” That marginal, incremental framing - the same logic behind incremental ROAS - is what a saturation curve makes visible. The curve does not just warn you that a ceiling exists; it tells you roughly where it is before you hit it.
How ElenIQ models saturation
ElenIQ fits a saturation curve to each channel from your historic spend and response data, using a Hill-style diminishing-returns function together with adstock to account for the carryover of effect between periods. Rather than reporting a single backward-looking average, it estimates the full curve - and therefore the marginal return at any spend level you might be considering.
That turns budget planning into a forecast rather than a guess. You can see where each channel’s budget ceiling sits, identify which channels still have efficient headroom, and reallocate towards the pounds that work hardest before committing a single one. It is the practical expression of ElenIQ’s positioning - forecast before you spend - and it is built into the workflow when you create a media plan rather than bolted on as a separate report.
Frequently asked questions
What is a saturation curve in marketing?
A saturation curve describes how the response from a paid media channel grows quickly at low spend, then flattens as the channel runs out of efficient audience to reach. It maps each level of spend to the result it is realistically expected to produce, capturing diminishing returns rather than assuming a constant rate of return.
What is a budget ceiling in paid media?
A budget ceiling is the spend level beyond which a channel stops producing worthwhile incremental return. It is not a hard cap imposed by the platform but the point on the saturation curve where each additional pound returns less than your efficiency target, so further investment is better placed elsewhere.
Why does paid media not scale linearly?
Paid media does not scale linearly because the most responsive audiences are reached first. As spend rises, platforms widen targeting and bid into more expensive, less interested impressions, so each additional pound buys a smaller incremental result. The relationship between spend and outcome bends rather than following a straight line.
Can a channel have strong average ROAS but be weak at the margin?
Yes. A channel can appear strong on average ROAS while becoming weak at the margin. The average blends together the highly efficient early spend with the inefficient latest spend, so it can stay healthy even after the next pound has stopped paying its way. Saturation curves help identify that before spend is committed.
How does ElenIQ model saturation?
ElenIQ fits a saturation curve to each channel using historic spend and response, applying a Hill-style diminishing-returns function alongside adstock to account for carryover. It then forecasts marginal return at different spend levels, so you can see where each channel’s budget ceiling sits before committing the spend.