+9.0 ROI from YouTube brand investment, measured through MMM over an 18-month cycle

Case studies / Measurement and MMM

The situation

A London-listed betting and gaming group with several brands and more than £50m of annual media investment. Platform attribution disagreed with itself and with finance. Brand channels, YouTube in particular, were cut at every budget round because nobody could show a return.

The problem

Last-click attribution gave almost everything to brand search and the affiliates. Prospecting channels looked expensive. YouTube looked like it delivered nothing. The board’s question was simple: what would we lose if we stopped this? Nothing in the reporting could answer it.

What Joseph did

• Built the data. Two years of weekly sales, media spend by channel and market, price, promotions, sporting calendar and seasonality, cleaned and assembled into one dataset.
• Built the marketing mix model. A statistical model explaining weekly acquisition and revenue by channel, with adstock and saturation curves, validated against the geo tests already running.
• Ran it as a cycle, not a project. The model was refreshed quarterly over an 18-month cycle, so each budget round was planned from the latest read.
• Used it to plan YouTube. With MMM as the measurement, YouTube was planned as a brand and demand channel and scaled from £0 to £4m a year.
• Reported it to the board. One page per quarter: what each channel returned, where the next pound should go, and the confidence range around each number.

The result

+9.0 ROI from YouTube brand investment, measured through MMM over an 18-month cycle. The model also informed budget allocation across the other channels as part of a wider performance programme that delivered a 3.1 ROI.

What this means for you

If you spend across several channels and cannot answer the board’s “what would we lose” question, MMM is the method. Joseph has built it, run it and used it to make budget decisions at scale.