Our Mission
we're building the measurement layer that brands can actually trust.
marginal exists to give marketing teams an honest answer to the question that matters most: what did our spend actually cause? we do this with rigorous methodology, open-source tooling, and no commercial incentive to tell you anything other than the truth.
why we built this
the measurement problem in marketing isn't new. last-click attribution has been the dominant methodology for two decades, and everyone in the industry privately knows it's wrong. it records who was in the room when the conversion happened — not who influenced the decision. it makes brand investment invisible, treats channels as independent when they're not, and optimises for the measurable rather than the impactful.
for a long time, the alternatives were worse. enterprise marketing mix modelling cost tens of thousands of pounds per run, required months of setup, and delivered outputs that arrived too late to influence the next quarter's planning. incrementality testing was the preserve of teams with dedicated data scientists and the appetite to run controlled experiments at scale. the cost of the alternative never cleared the bar.
what changed is that the tooling caught up. open-source bayesian mmm engines like google meridian are now academically rigorous, publicly auditable, and accessible without a seven-figure measurement budget. the barrier to running a proper model collapsed. marginal was built to make that accessible — to take the methodology that was previously reserved for the largest advertisers and build a service around it that any brand running meaningful media spend can use.