Skip to main content
Measuring improvement in marketing
Written by Konsta Mustasilta
Updated over a week ago

In this article, we discuss a few alternative methods for measuring improvement in marketing over time.

1. Tracking ROI of marketing

The most obvious metric to track over time is the ROI of marketing. A requirement for seeing improvement in ROI, is that investments have been shifted from low-ROI media and campaigns to ones where additional investments bring higher ROI. When shifting budgets, it is of course important to take into account the diminishing returns effect.

When conducting comparisons, it is good practice to choose timeperiods, which are highly comparable. For example, it is not a good idea to make strong judgements if comparing December's ROI to July's ROI, given that there are likely strong seasonal effects at play, at least during December.

Below is a screenshot from the Sellforte live demo ( where to find ROI.

2. Tracking incremental sales from marketing

Another closely related performance metric is the incremental sales from marketing. While the ROI metric gives the average performance for each Euro or Dollar in the marketing budget, incremental sales provides the absolute value how much your marketing has driven sales. By tracking the development of incremental sales, you can evaluate how much value in absolute terms your changes in marketing activities have delivered.

Below is a screenshot from the Sellforte live demo ( where to find incremental sales.

3. Controlled tests and experiments

An additional way to test whether certain activities have positive impact on marketing performance, is to conduct controlled tests.

In these tests, one chooses a control group and a test group. The activity-to-be-tested, for example budget re-allocations, is conducted on the test group, but not on the control group. After the test is done, the results are compared.

The design of the test depends on the setting. For example, tests for digital media are much faster to execute compared to offline media.

Example of a potential test/control could be for example one country vs. another fairly similar country.

Did this answer your question?