The tip of the iceberg issue with MMM

Most MMM projects give marketing a view on media investment, then leave price, promotions, distribution and new products somewhere else in the organisation.

The cost comes later: finance has its own price and promotion analysis, demand planning builds a sales forecast, product teams assess launches and marketing runs an MMM, often using much of the same data to answer versions of the same question.

What is driving revenue, and what should we do about it?

That is why I think econometrics is underused - businesses put serious money and internal time into these projects, then use the result mainly to decide where the media budget should go next.

Media matters, but it is only a part of the commercial picture.

A media-only MMM leaves too much inside the base

A media mix model that includes media and a single “base” line can look neat, but it makes a big assumption: it treats everything outside media as a lump of unexplained sales.

That base can include price changes, promotional activity, distribution gains or losses, new product launches, seasonality, competitor activity and wider economic factors.  You cannot get a clean read of media without dealing with those things properly.

A strong promotion can make paid social look better than it is, a distribution loss can make brand activity look weaker than it is and a price rise can reduce volume while improving profit.

All of those movements can happen while media is in play, which means the model has to separate them before it can tell you what media actually contributed.  There are no shortcuts here.

This is why I prefer the term econometrics to MMM. MMM can mean media mix modelling (vs the more correct and robust market mix modelling), and media mix modelling is really media analysis with a base line underneath it, whereas proper econometrics decomposes the full set of commercial drivers behind sales.

That distinction matters to a CMO, but it matters just as much to a CFO. You need to know whether revenue moved because of marketing, price, promotions, distribution or a combination of all of them, then understand what each change means for profit.  Putting marketing in context with other drivers really matters.

Similar models often sit in different departments

In many CPG and most businesses, finance is assessing promotions while marketing is measuring media uplift and demand planning (or finance) are forecasting sales.

Each team has a valid reason for doing its work, but the models often overlap.

They use similar sales data, they need the same promotional calendar, rely on pricing, distribution and market information and they all aim to explain the movement in revenue.

The business then pays several times to source data, feed models and interpret the answers.

It also gets several versions of the truth.

Marketing may see a campaign as the main driver of growth, while finance sees promotion and demand planning has another explanation based on its forecast. None of those teams are necessarily wrong, but separate models with different assumptions create a lot of unnecessary reconciliation.

That is where trust gets lost.

The board should not need a translation exercise every time it asks why sales moved.

One model should help each team make its own decision

A proper econometric model can support several teams because it measures the commercial system as a whole.

For marketing, it should measure media, CRM, direct response and the relationship between brand and performance activity.

Recommended by LinkedIn

For revenue management, it should assess promotions properly, including brought-forward sales and cannibalisation. A promotion that increases this week’s sales may pull demand from a later week, or simply take volume from another product in the range.

For finance and pricing teams, it should show price elasticity and the likely effect of a base price change on revenue and profit.

For product teams, it should help distinguish genuinely incremental sales from sales that have moved between products.

For demand planning, it should create a forecast based on what is happening across the business, rather than treating media, promotions and pricing as separate manual adjustments.

For the CFO, it should support scenario planning through the P&L: what happens if we change price, reduce promotional intensity, increase brand spend or launch a new product into a different distribution position?

That is a much more useful discussion than which media channel had the best return last quarter.

The biggest gain comes from one commercial view

Consolidating analytics can reduce model-build cost by multiple £100ks and save substantial internal time, but the bigger gain comes from getting marketing, sales, finance and demand planning behind the same explanation of revenue.

Marketing can see the effect of price, promotions and distribution on the sales movement; Finance can see media in the context of revenue and profit; the Board gets one view of what happened, rather than several analyses that need reconciling.

This makes decisions quicker because teams spend less time debating whose number is right.

It also changes the conversation - rather than asking whether marketing or promotions caused a sales movement, the business can look at the contribution of each driver and decide which lever offers the best commercial outcome.

That is how an econometric model becomes part of planning and budgeting, rather than something marketing presents after the main decisions have already been made.

Start by finding the gaps and the duplication

You may not need to replace your existing MMM, but you do need to understand what it is actually measuring.

Does it isolate price, promotions, distribution, new product activity, seasonality, competitor effects and wider economic factors, or does it leave them inside an unexplained base?

Then map the other models across the business.

What question does each one answer? What data goes into it? Who owns it? How often is it updated? Which decision does it support?

• What question does each one answer?

• What data goes into it?

• Who owns it?

• How often is it updated?

• Which decision does it support?

That audit shows where work is genuinely separate and where the organisation is paying several times to answer the same question.

The test is simple: can your current model explain the main reasons revenue moved, then help each team make a decision from that answer?

If it cannot, you are only seeing the tip of the iceberg.

Summary

MMM and econometrics should help with much more than media decisions.

Make sure your model decomposes the main drivers of sales, then look for separate teams building overlapping models from the same data. One single commercial view can reduce cost, cut internal effort and give marketing, finance and the board a clearer route to action.