Marketing Mix Modelling: Getting It Right

July 7th, 2021 | Winston Li, Founder, Arima

Many marketers find it challenging to figure out the optimal budget they need to achieve their marketing objectives. Indeed, being able to correctly size your media budgets and determine the optimal way to allocate the money across a wide range of marketing channels is an essential exercise for every marketer. Successful businesses have always prioritized their marketing initiatives and having a robust Marketing Mix Model (MMM) is a key component. Having a well-designed MMM will give you the tools and information to avoid the risk of an under-sized marketing budget and, above all, will help you justify your resource needs.

What exactly is an MMM?

An MMM correlates key KPIs like sales, revenue, or site visits with drivers such as marketing spend, competition and economic indicators to forecast future sales and discover incremental business opportunities. In my view, a good MMM should give a comprehensive view of marketing activities and their impact on profits, make strong predictions to help size the budget and allocate media dollars, and connect the dots between brands, agencies, and publishers so that everyone works with the same benchmarks.

Arima has been a leading provider of MMM solutions in Canada, and in our framework, a well-designed MMM must be able to function as:

  1. An optimized budgeting tool, where it supports the budget planning process and determines the size of the media budget needed to achieve business goals. Typically, this includes budget calculation and what-if analysis.
  2. A forward-looking predictor, where it guides future business decisions. Typically, this includes sales forecasting and media plan evaluation.
  3. A backward-looking explainer, where it helps you review historical campaign performance, including ROAS calculation and sales attribution.
  4. A geo-centric framework, where it enables matched market tests, such as creative A/B tests, channel ratio tests, and reach/frequency variation tests. These sorts of tests help account for the differences in conditions between test and control markets to isolate the true effects of campaigns.

Challenges in Building an MMM

Building an MMM has always been a challenging task for marketers, both from a resources and technology perspective. In the past, the reliability of results from MMMs are challenged by three main reasons.

  1. Data Collection. Marketers have trouble aggregating the required datasets for a full robust model because data is siloed, and historical data may not be properly stored. Furthermore, data collection also limits the ability for marketers to work with the most up-to-date data and models.
  2. Actionability. Marketers are unsure how the results of MMMs could be integrated with their day-to-day marketing planning or business operations.
  3. Adapting to New Modelling Techniques. Data science is an evolving field and it can seem as though there’s a new modelling technique every month. Marketers need help on selecting the most appropriate statistical and machine learning models for their use cases.

A Marketing Mix Modelling Framework That Works

Arima’s marketing platform tackles the above stated challenges by providing marketers and their agencies planning analysis tools sourced from a single foundation – the Synthetic Society: a framework for generating individual-level data from aggregated and anonymized data sources. By taking a platform approach to marketing mix modelling, Arima has the ability to:

  1. Enable marketers to build and manage their own MMMs that automatically combine data from multiple sources and ensure continuity in their marketing measurement efforts.
  2. Allow marketers to immediately view actionable results of their MMMs through the use of a special media planning tool.
  3. Incorporate the latest modelling techniques to ensure that a marketer’s MMM delivers excellent results when it comes to measuring marketing ROI and forecasting future sales.

To summarize, Marketing Mix Models support your marketing decisions in protecting/defending marketing budgets, budget approval, choosing the right investments and maximizing return on marketing investments. The digital revolution has unleashed many new measurement and attribution capabilities, and Marketing Mix Modelling is quickly becoming one of the most exciting tools available to today’s marketers seeking to build a stronger brand and a more efficient marketing plan.

For more information, members can log in and view the recording and slide deck from Winston Li’s webinar.



Winston Li is the founder of Arima, a platform that enables marketers to build and manage their own media mix models combining data from multiple agencies, including any in-house services, to ensure continuity in their marketing measurement efforts. Prior to founding Arima, Winston was the Director of Data Science at PwC and Omnicom. Winston is also a part-time faculty member at Northeastern University Toronto and sits on the advisory board of the Master of Analytics program.