Channel effectiveness: knowing what works and by how much

At its core, MMM quantifies how each media channel contributes to incremental sales or conversions. Using historical data on spend, impressions, sales and external factors such as competitor activity; MMMs estimate the ROI of a channel. In Bayesian, something called a posterior distribution is generated, which represents a range for the ROI of that channel, not just a single point estimates (e.g. the ROI for TV is likely between 2 -2.4).

The key metrics that emerge include:

• ROI (Return on Investment): incremental revenue per pound spent on a channel
• Elasticity: how responsive sales are to percentage changes in channel spend
• Credible intervals: the probability range for each channel’s true ROI (e.g. 50% probability that the true channel ROI lies between 2- 2.2)

For decision-makers, this means more than just knowing ‘TV performs better than social’. The outputs allow clients to ask their media agencies, “how confident are you that TV performs better than social?”

Response curves: finding the point of diminishing returns

One of the most valuable outputs of an MMM is the spend-response curve. These show how each channel’s performance scales with increased spend, capturing aspects known as ad stock (how long ads keep working after airing) and saturation effects (when returns diminish at higher channel spends).

These are estimated through nonlinear transformations made to the data using specialist functions which describe how a KPI approaches a plateau as media spend increases. Put simply, it tells you how much is too much (over-saturation) and how much is too little (under-saturation).

These curves help media agencies and their clients:

• Identify optimal spend levels per channel before diminishing returns kick in
• Understand marginal ROI at each channel spend level
• Recognise under- and over-invested channels which helps optimise spend
• Build intuitive ‘what-if’ scenarios showing the incremental revenue for different media budget channel splits or scenarios

MMM takes marketers from intuition to evidence-based decision making

The re-emergence of MMMs helps marketers use data to move from intuition-based to evidence-based media investment decisions. When MMMs are built robustly and are validated by channel experts, they undoubtably enable smarter media planning. MMMs help marketers answer key business questions such as where to spend more or where to cut and how confident we are in the return on our investment.

Modern MMMs, using techniques such as Bayesian, provide outputs alongside uncertainty. This is powerful because levels of risk can be communicated alongside more typical MMM outputs such as channel saturation, budget optimisation and scenario analysis.

At MI Media, MMM outputs are communicated to clients in MIDAS, our automated reporting dashboard which adds interactivity, allowing our clients to play with ROI tables, response curves, KPI contribution charts, budget optimisation scenarios and more. While this adds a useful interactive element over traditional presentations, it’s still important that we work closely with our clients to discuss the key business findings and how they will affect future media decisions.