top of page

The basics

MMM - what does it stand for?

MMM is Market Mix Modelling.


Market Mix Modelling is called many things, namely:
MMM, Media Mix modelling, Marketing Mix Modelling, Econometrics, Triple M (mainly APAC), MMX (some CPG / FMCG brands)
​
Whilst they by and large mean the same thing, there are some discrepancies.  
​
Market mix modeling is far more descriptive of what a comprehensive model should be, as it considers the whole market, not just media.  This includes other drivers such as promotions, pricing, COVID, seasonality, etc. If we don’t include them, we don’t get an accurate read on media and risk overstating its impact and making terrible investment decisions based on the “insight” generated.
 

There are many outfits out there who say they execute "media mix modeling". However, some of these are just building a simple regression with media variables, or even just using multi touch pathway techniques (which is digital media uplifts only and not an incremental analysis).
 

MMM should always be Market Mix Modelling - it's the more accurate model for brands.

Okay, but what does it do?

Market Mix (Econometric) Models is one of the most accurate way of isolating and separating marketing impacts.  It uses statistical analysis mixed with economic theory applied to companies sales to quantify all drivers of sales.  Generally models are built on KPI such as Sales or Revenue and then used to relate marketing impacts back to profit.


This is achieved by separating out all factors affecting Sales or Revenue over time to build up a picture of the key drivers, such as isolating the impacts of price, promotions, distribution, marketing, seasonality, economic effects, etc.  
 

This allows the quantification of how marketing spend affects performance, and the subsequent marketing efficiency can be calculated by overlaying spend.  In turn, this is then used to optimise future marketing investment decisions.

So does it just to understand what drives Sales?

​No it doesn't.  It can be used to model any Key Performance Indicator (KPI).  For larger clients, you generally want to model the impact of media down the customer funnel, so then you derive a good understanding of the role your differenet media are playing.


These start from brand health metrics such as awareness, consideration, intent;
To research metrics suck as google searches, web visits;
To traffic metrics like web visits, store footfall, phone call volumes;
To harder financial metrics like volume sold, quotes, sales, acquisitions, revenue;
And finally loyalty metrics such as retention, churn and onto lifetime value.

Why would you use this technique?

​MMM delivers an incremental quantification of all things that drive your sales.  This makes it incredeibly useful for media, as media is notoriously difficult to measure consitently across different media channels for the following reasons:

​

  • Carry over effects (i.e. you see an ad this week, you bought next week) are very hard to measure without disentangling all drivers of sales
     

  • Its difficult to to get measures of offline vs online which are consistent and therefore comparable
     

  • Often media is planned with other events, such as seasonal peaks (Christmas) or other media, making it hard to pinpoint exactly what is driving what

 

MMM solves all of this by measuring the carryover effect of media, consistently measures across offline and online so you can optimise between the channels, and  can isolate all other drivers such as seasonality, events, promos, etc.

Where is it used in business?

There are quite a few applications of MMM, although the majority of the use is with quantifying and optimising media.  Here are some applications of MMM ranked by use:
 

  • Media uplifts: quantifying the incremental impact of media across offline and online channels, allowing the optimisation of media plans in the future to maximise effectiveness
     

  • Forecasting: MMM gives an informed view of the future based on historic relationships. MMM is generally pretty good at forecasting up to 6 to 9 months in the future, anything longer than that and varying consumer trends and unobserved impacts (i.e. covid) start creeping in and affecting accuracy.
     

  • Promotions / Trade: MMM can not not only measure the incremental impact of promotions for CPG brands (multibuys, money offs, etc), it can also measure the cannibalisation impact of promotions between packs as well as any brought forward sales impact,  This makes it a very useful tool in identifying the true profitability of trade promotions.
     

  • Portfolio Optimisation: one again for CPG, but this looks at which products and markets to prioritise to maximise returns, whether that is with trade spend, media spend or assortment (i.e. which products to list/delist).  MMM provides a way of measuring all of these incremental impacts as well as identifying base sales, which help prioritise brands.
     

  • Price optimisation: One for most brands, MMMs ability to measure the impact of price changes, and therefore read price elastiticty, therefore makes it a great tool for optimising profit from price changes.  You can do this at a product / regional level, or even more granular such as by store, depending on how you set up the model.  The end results is to use the measured elasticity to simulate impacts through the P&L and see what price will deliver the most profit.
     

  • Competitor analysis: MMM can not only be used on your brand, but also on your competitors.  If you have there sales (from syndicated studies such as GfK or data companies like Nielsen), then you can model your competitors to see what they are doing to gain or lose share, and build that in to your own strategy.  Even if you dont have competitor sales, you can use proxies such as google trends to model digital consumer interested in your competitors, and see if that new campaign of theirs is working or not.

Is it the right thing for me to implement?

​MMM is not for all brands - it can be expensive or time consuming to set up.  However if you fulfil any of the following criteria then it could be for you:

  • Spend over $1m on media per annum, with a mix of channels used

  • Have a large % of retail / phone / offline sales

  • Spend on offline media channel and want to get a read

  • Spend on digital channels only, but want to get an accurate view on how social drives paid search, and measure all digital channels on an incremental, level playing field

​

Generally MMM is not for you if:

  • You have very little media spend (i.e. under $1m per annum)

  • Only use one media channel

  • You don't believe seasonality, price, promotions, economy will impact your brand (you can then see media uplifts directly)

bottom of page