Optimising future media budget

At its core, an MMM analysis is about understanding past performance to help optimise future marketing decisions. So, measuring how different levels of spend for media channels impact the levels of returns we would expect is important.

Measuring “diminishing returns”

In an MMM model, we not only calculate the adstock we also calculate the diminishing returns. This measures that as spend increases, we generally expect not to receive the same level of returns at higher spends. In practice, this captures that as we spend more on TV or Social media, we are: Hitting the same people multiple times (frequency) or to increase our reach we need to be placed to a less relevant audience.

The two charts bellow illustrate how coefficients and diminishing returns work together to achieve the right model fit.

Budget optimisation

Once we have the impact of each channel and how it changes at different spend levels, we can optimise the budget. At Linea, we typically do this by using a hill climb algorithm. This at each level of budget allocates the next €/£/$ to the most efficient media channel.

See how this works with the interactive chart below.