Getting Ready: The Three Roles, The Data, And The First 90 Days

The practical questions matter as much as the strategic ones. What data do you need? Who on your team needs to be involved? And what does a successful first engagement look like? 

Practical Guide  -  7 min read  -  MASS Analytics 

By this point in the series, the case for MMM should be clear: it gives you an accurate, independent, forward-looking picture of what’s driving your business and where your marketing budget should go. What remains is the question most marketing directors ask immediately after: “Alright, but what does it actually take to get started?“ 

The answer is that the barriers are lower than most people expect, honestly. The data requirements are manageable. The internal resource demand is real but not overwhelming. And the process, when structured properly, is designed to fit around your existing operations rather than disrupt them. 

The Data Question  

The most common source of hesitation is the data. “We’ve got data everywhere, it’s a mess” and “I’m not sure we have three years of clean weekly sales” are things we hear regularly. Neither is the obstacle it might appear to be. 

What an MMM engagement needs, at its core, is: weekly sales data over approximately three years; spend data for each media channel over the same period; and a reasonable record of significant business events: price changes, promotions, store openings, anything that caused a discontinuity in normal trading. 

The rest: economic indicators, competitor spend proxies, weather data for relevant categories, is largely available from external sources and can be ingested automatically. Media agency data, for those working with agency partners, can be pulled directly through established connections. The picture of “someone needs to manually gather three years of data from fifteen different systems” is an outdated one. 

Imperfect data is not a reason to wait, it’s a reason to start because the model itself will tell you where the gaps are and what to prioritise fixing. 

Where data is genuinely inconsistent or partially missing, there are established techniques for handling it. A model built with some data gaps is still substantially more useful than no model at all, and the process of building it tends to improve data discipline across the business, because it surfaces exactly which data streams matter most and where the collection is weakest. 

The Three Roles Your Business Needs 

You don’t need a data science team or a dedicated analytics function to run a successful MMM programme. You need three people, not necessarily full-time, but clearly identified and genuinely engaged: 

The MMM Champion 

Most involved throughout (especially at the start and during results readouts) 

This is the person who owns the programme and drives its adoption internally. They don’t need to be some technical expert in modelling, they just need to understand the business questions being answered, be able to communicate the value of the work to senior stakeholders, and have the credibility to act on uncomfortable findings. This role is often the CMO, Head of Marketing, or a senior marketing strategy lead. 

The Data Owner 

Most involved during data collection and the model-building phase 

This is the person who knows where your data lives and can get to it. They’re the connection point between the MMM partner and your internal systems. Their role is to  make sure sales data gets extracted in the right format, that media spend records are accessible, and that business events are documented. This is often a head of data, a senior analyst, or someone in the BI or IT function. They don’t need to be dedicated to the project full-time, they need to be available and responsive when data questions arise. 

The Media Expert 

Most involved during optimisation and scenario planning phases 

This is the person who understands how your media works: what’s been running, when, at what spend levels, and what the context was. This role can sit entirely within your media agency rather than in-house; agency partnerships work well here. What matters is that someone with genuine media knowledge is involved in interpreting the model’s outputs and translating them into actionable changes to the media plan. 

What to Expect: A Realistic Timeline 

MMM engagements vary in scope, but a structured first programme tends to follow a recognisable shape: 

Weeks 1–3: Scoping and data collection 

Business questions are defined precisely, together. Data sources are identified and connected. The scope of the model: which KPIs, which channels, which level of geographic granularity, is agreed upon as partners. The Data Owner is most active at this stage. 

Weeks 4–8: Model building and validation 

The model is built, tested, and validated against historical data. The first iteration happens here, early versions are reviewed with the client team, questions are raised, assumptions are tested. The model is refined together until its fit with historical sales is strong and the outputs are commercially credible. 

Weeks 9–12: Results, scenarios, and recommendations 

The full output is presented: sales decomposition, ROI by channel, response curves, and synergy findings. Scenario planning runs which consists of what-if analyses on budget reallocation, seasonal shifts, and competitive responses. The Media Expert is most active here. 

Ongoing: Always-on refresh and optimisation 

The model becomes a live tool. New data feeds in automatically. The model refreshes as trading conditions change. Optimisation runs on a regular cycle aligned to planning cadence: quarterly budget reviews, seasonal campaign planning, annual strategy. The programme builds institutional intelligence over time. 

The Managed Versus in-house Question 

One practical decision to make early is how much of the ongoing programme you want to run yourselves versus have managed externally. There are two models, and a spectrum between them: 

Recommended starting point: 

Managed partnership 

Your partner handles model building, maintenance, and analysis. You access results through a platform and work with your partner to interpret and act on them. Lower internal resource requirement; faster time to insight. 

In-housed programme 

Your team runs the modelling using purpose-built tools, with support and training from your partner. Greater internal control and capability development, this requires more commitment from your team over time. 

For businesses approaching MMM for the first time, the managed route is almost always the right starting point. It gets you to insight faster, reduces the risk of early missteps, and lets your team learn what good looks like before taking on more of the technical responsibility. The goal, over time, is to move toward a position where your team understands the outputs deeply enough to use them confidently, whether or not the model-building itself is handled externally. 

The Stakeholder Conversation 

There’s one aspect of readiness that’s easy to overlook because it’s not about data or technology, and that’s the internal political preparation. 

MMM produces uncomfortable truths. A channel that’s been a cornerstone of your media plan for years may turn out to be saturated. A cost-saving that looked sensible may be revealed as a revenue risk. A budget allocation that a media agency has defended vigorously may not hold up under independent scrutiny. 

The value of that honesty is exactly the point. But it lands very differently in an organisation where everyone relevant has been brought into the process from the start, versus one where results are presented without prior alignment. Finance needs to understand what the model is trying to do. Sales leadership needs to be prepared for findings that might change how they think about promotional investment. The board needs to trust that the numbers are independently derived. 

Before you begin, you need an internal checklist: 

• Have you identified your MMM Champion, Data Owner, and Media Expert? 

• Do you have approximately three years of weekly sales data that can be accessed? 

• Is your media agency aligned on sharing spend data for the model? 

• Have key internal stakeholders ‘finance, sales, marketing leadership) been briefed on what the programme aims to produce? 

• Is there executive sponsorship for acting on the findings, even if they’re uncomfortable? 

That last point, executive sponsorship for acting on findings, matters more than any data question. The most technically excellent MMM programme delivers no value if its findings sit in a slide deck and nothing changes. The businesses that get the most from marketing measurement are the ones where the results are genuinely wired into planning and budget decisions.  

That’s a cultural commitment as much as a technical one. 

Previous article: Multi-Banner, Store-Level, Wholesome: Does MMM Work For My Business

Back to: Series Overview

View all articles