The What

Media Mix Model or Triple-M(MMM) as it is fondly called by marketers of the day is quickly turning out to be the first thing on the priority to-do list, compared to its relative ‘future-stage-utopian-solution’ moniker it had earned in the not-so-distant past.

The very fact that MMM has climbed up the priority list of marketers is a testament to the widespread availability of well-defined good-quality data and the recognised need for optimal spending versus traditional opportunistic spending.

For those not in the know, MMM is a virtual mathematical accountant that offers a statistical well-rounded way of finding the optimal investment in each marketing channel, along with a host of recommendations on why the investment would work and what the expected Return on Marketing Investment(ROMI) would be for each channel.

The Why

If you are the marketing decision maker or someone who enables the decision making of a marketing decision maker, MMM is the first thing you need. And if you do it in-house or through a third party, it behooves a well-rounded decision maker to ask a few questions to appraise the quality of the MMM. Apart from the de-rigour questions of accuracy, frequency and latency, here are 7 other key questions you should be asking – along with the why for asking each of them.

1. Granularity – What granularity of ROMI will the MMM provide?

On the outside, this might seem like a silly naive question. But there’s more to it than just the word granularity or in common terms, level of data. Will the MMM be able to provide ROMI not only at a channel level, but also go down to a campaign level and even further at a specific campaign & product combination level? This is critical because the granularity lets you make better actionable decisions rather than just knowing what doesn’t work at a global scale.

E.g. Instead of just saying, TV works and Press doesn’t work – would your MMM be able to say TV works in Rural areas where press circulation is naturally lower, but only for Product X – and from an Acquisition strategy perspective?

2. Overlapping halo effects of campaigns – Will the MMM be able to measure and account for overlapping halo effects of campaigns flighted simultaneously?

Most MMM models offer a simple mathematical construct that take in spend across various marketing channels and measure them against an optimal revenue figure – but in the process, do they also bring in the subjective aspect of one campaign’s halo effect on another. Simply computing ad stock for a channel and time-decaying it won’t suffice because it doesn’t account for the decaying quantity of a channel’s campaign on other live campaigns.

E.g. Instead of just having an adstock value for TV, and time decaying it, does the MMM also take into consideration the impact of TV ad,which decays in 3 days, on a digital ad campaign 1 day after the TV ad? When does TV stop and when does digital begin? Or do they overlap each other for a couple of days?

3. Integrating impact of offline & online together – Does the MMM offer a combined view of both online & offline channels & how does it measure the synergy?

MMM models are virtual mathematical accountants that can transfer the unspent money in one channel to another – but can the accountant also take into consideration that one set of channels are Below The Line, while the other set of channels are Above The Line. How then will the MMM handle the duality of messaging and difference in need for marketing to both of them?

E.g. Instead of recommending to spend 50% of Radio ads on SEM ads, will the MMM be able to measure the traffic driven from a radio ad to search online via SEM. If so, how is that taken into consideration?

4. What-If Scenario Optimiser – Will the model be real-time, flexible and able to churn out multiple views instantaneously?

MMM comes with a notorious moniker for being late – 6-12 months late at times. If the MMM takes 6 months to refresh every time, wouldn’t the marketer lose the potential to make use of a very current opportunity that might have changed the marketing paradigm due to, say, competitor product launch, for example

E.g. If you need to remodel your spend and increase the boundary limits for Digital display ads by 2x because of a sudden surplus in budgeting, how do you real-time figure out the best way to spend them? Would you wait 6 months?

5. Ability to expand MMM to new products & strategies – Will the MMM be able to scale, expand and flexibly accommodate new products & strategies?

A What-if Optimiser is just that – plays what-if games with your marketing hypotheses. But if you have a new product launch or your company’s marketing strategy has suddenly changed from one quarter to another, what do you do with your MMM? Re-do it all over again? How scalable and flexible is the model to incorporate the changing flux of the organisation and the industry?

E.g. If you have a new product launch with 5 new products and pricing in a quarter and the marketing strategy has shifted from awareness to retargeting to enable cross-sell of these 5 products, what would your current MMM do?

Conclusion

Apart from all these good-to-know questions, it’d also be wise to enquire about the kind of reporting and scenario optimisation interface to view and review the MMM outputs. And last but not least, the ability to translate all of the mathematical jargon into a media-agency-usable media plan would be a definite good-to-have. 

When you have good answers to all these questions, go ahead and get into MMM. May the Marketing Force be with you!