Articles & posts.

MMM, MTA & Decision Econometrics: answered.

What is Market Mix Modelling (MMM)?

Market Mix Modelling is a statistical, econometric technique that decomposes sales or another KPI into the contribution of every commercial driver - paid media by channel, price, promotions, distribution, seasonality, macro factors and a base level. It produces ROI, response curves and saturation points per channel, and is the standard top-down approach for measuring marketing ROI across both online and offline activity.

What is multi-touch attribution vs MMM?

Multi-touch attribution (MTA) follows individual user journeys across digital touchpoints and assigns fractional credit per click or impression - granular, short-term and digital-only. MMM is aggregate and top-down: it models all sales drivers (online and offline media, price, promo, distribution, brand, seasonality, macro) to produce defensible ROI and saturation curves per channel. MTA optimises in-flight digital bidding; MMM answers full-business marketing ROI and budget allocation. Modern programmes run both, calibrated against geo-experiments and platform lift tests.

How do you measure marketing ROI?

Pick the commercial KPI (revenue, gross profit or contribution), pull weekly spend and exposure by channel together with price, distribution and macro data, then fit a Bayesian MMM that decomposes the KPI into a base plus each driver's incremental contribution. Calibrate with geo-experiments and platform lift studies, divide each channel's incremental contribution by its cost for ROI, derive saturation curves to find the optimal mix, and refresh monthly so the numbers steer live planning, not retrospective reporting.

What are the best marketing mix modeling tools?

Three families. Open-source (Meta Robyn, Google Meridian, LightweightMMM) - flexible and transparent, but require an in-house data science team. Enterprise SaaS (Nielsen, Analytic Partners, Mass Analytics, Recast) - turnkey but slow and opaque. Modern decision-econometrics platforms like twenty10 - Bayesian MMM with always-on data pipes, scenario simulators, budget optimisation and direct decision recommendations: faster than enterprise, more defensible than DIY. The right tool depends on data maturity, required refresh cadence and whether you need a model or a decision system.

How is MMM different from Multi-Touch Attribution (MTA)?

MTA assigns credit across user-level digital touchpoints and is most useful for short-term optimisation of trackable channels. MMM is aggregate and top-down, covers all channels including offline and brand, and is the credible answer for full-business marketing ROI and budget allocation. Modern measurement programmes typically use both, plus incrementality experiments to calibrate them.

How does Decision Econometrics measure marketing ROI?

It models the KPI as a function of every relevant driver (media, price, distribution, seasonality, macro) and isolates the incremental contribution and return of each. Modern builds use Bayesian priors informed by experiments and platform lift studies, are validated against out-of-sample holdouts, and are refreshed monthly so the numbers drive planning rather than retrospective reporting.

Who is twenty10 for?

twenty10 is built for CMOs, CFOs and leadership teams who need one trusted, defensible answer to what drives growth, what marketing ROI looks like by channel, and what happens to revenue and profit if spend changes. Typical engagements unlock 10-30% marketing efficiency through reallocation and £1m-£20m of profit per programme.

What outcomes should I expect from an MMM engagement?

A well-scoped modern MMM should deliver: defensible channel-level ROI and saturation curves, a reallocation that finds 10-20% more value in the same spend, scenario forecasts for the planning cycle, and a refresh cadence (monthly or quarterly) so the model informs in-year decisions, not just the annual budget.