The calculation for overall ROI is the total revenue you’ve achieved, divided by the total spend.

In contrast marginal ROI models focus on the predicted increase in revenue from an increase in spending, so what your next unit of budget will deliver.

For example, if spending increases from £80,000 to £100,000, and forecasting tools predict that revenue would increase from £200,000 to £230,000, the marginal ROI of that £20,000 is 1.5 (£30,000 divided by £20,000)

Marginal ROI can deliver insights to guide your spend across your digital programs. Let’s look at how it could help inform spending decisions within paid search…

Marginal ROI and paid search

By breaking your strategy down into separate components, such as advertising on different search engines, you can predict what spending more on particular components would deliver.

Essentially, there will be a point where spending more, on even the most successful search engine, will see a drop in marginal ROI compared to rivals, simply because the return curve flattens out.

You can see this more clearly in the graph below.

This is based on forecasting results based on spend for two different publishers. It shows the predicted curve of returns, and where your current spend positions you on this curve.

From here, it is simply a matter of which curve has the higher marginal ROI at that point.

In this example, you can see that Publisher B delivers a higher overall ROI. However given where you are on the curve, investing your next pound of marketing spend with Publisher A would give you greater incremental return for that particular pound, at least until spending on Publisher A moves further up the x-axis.

Essentially the curve for Publisher A is steeper as spending is on the earlier part of the curve.

This is typical of a scenario when a marketer underspends on a publisher – when the spending catches up with the other publisher then the curve will flatten and marginal ROI lessens.

To test this theory, Kenshoo analysed a series of paid search campaigns that spanned the two primary search publishers. It identified nine campaigns that had the necessary campaign organisation that meant it could see which publisher yielded the best return on the next pound of investment.


The research reveals that in more than half (five out of nine) of the campaigns we studied, Bing had a higher marginal ROI, showing that it would be a better place to spend that next £ of budget.

In three cases the greater marginal ROI was elsewhere, and one campaign showed no clear winner. This obviously doesn’t mean marketers should transfer all their budgets to Bing – simply that at certain points it offers a better incremental return on your investment.

Essentially in a two-publisher model a historically less-developed one can have greater upside for incremental investment.

Forecasting and understanding marginal ROI can help marketers make smarter decisions about their budgets. Clearly, we don’t operate in a binary, two-publisher world, but inside a complex ecosystem with a multiplicity of publishers and channels, each competing for spend and offering valid opportunities to reach customers.

By understanding the marginal ROI forecast for each channel and publisher, marketers can make sense of these opportunities and take informed, data-driven decisions.

In fact, they can finally get closer to answering the age old question: where do I best spend my next pound of marketing budget?