Multiply the profit of your PPC campaigns by using marginal ROI.

Why using marginal ROI to optimise your PPC campaigns is the only genuinely effective method — and why 95% of companies do not run their campaigns in a truly optimal way. This article shows and explains why every marketer should use marginal ROI* to optimise the bids of their paid campaigns.

This article shows and explains why every marketer should use marginal ROI to optimise the bids of their paid campaigns.

Take an example: here is the data for three keywords on which we spent three different budgets, according to those keywords’ positions, and which each generated more or less profit.

Question: how can the spend be reallocated among these three keywords to maximise overall profit, while respecting the total budget of CHF 375?

What should we do?

The best way to optimise the bids is to calculate the marginal ROI of each of these keywords.

That indicator tells you the revenue the next franc spent on each of them will generate.

To calculate marginal ROI, you have to analyse the impact of micro variations in the bid — up and down — on the keyword’s position and on its CPC. And you have to do that for every keyword.

The aim is to estimate each keyword’s future performance against very precise variations in the bid amount.

Spoiler alert: it is not humanly possible.

Such calculations would take an enormous amount of time for our three keywords alone — so they are unthinkable for a campaign holding thousands, or millions — given the multitude of analyses needed on the smallest effects of raising or lowering our bids.

To that are added other parameters we cannot ignore either, such as our competitors’ behaviour, since the changes they make on their side also affect our marginal ROI.

What is the solution?

It is impossible for a human being to calculate the marginal ROI of every keyword — but it is not so complex for a machine. Machine learning algorithms can today calculate the marginal ROI of several thousand keywords on a daily basis, while taking into account the impact of competition, of seasonality, and the rest.

In 2019 we use KPO (Kenshoo Portfolio Optimiser**) to optimise our campaigns’ bids automatically. We combine KPO’s remarkable capabilities with our methods for calculating customer lifetime value, so as to make sure every unit invested generates added value for our clients.

Update, September 2026: bid optimisation now runs largely through Google’s own algorithms, from Smart Bidding to AI Max, which takes over the whole Search chain; see Google AI Max: opening the black box and adapting your structures.

Why can the machine not (yet) replace the human?

Machines can indeed process a colossal amount of data in a minimum of time — BUT they are not yet intelligent enough to turn business objectives into a PPC strategy, nor to carry out all the steps needed to configure that kind of calculation, nor to check that the information processed is true.

We marketers are therefore (still) irreplaceable — a relief. We create and integrate the tracking system, plan the marketing work and so feed the algorithm with the most efficient data, so as to adapt our strategy and reach our business objectives.

Sources

* Marginal ROI is a term used in finance to identify where the next unit of value should be invested. The calculation of marginal ROI rests on the law of diminishing returns, by which the marginal return — or marginal productivity — obtained from using one further factor of production (land, capital, labour or other) falls, all other things being equal.

** KPO is a proprietary solution from Kenshoo for optimising keyword portfolios. The tool runs predictive models and groups keywords automatically by several attributes, so as to get the most out of the campaigns.