Beyond the Black Box: Revolutionizing Performance Measurement with Next-Gen Marketing Mix Modeling

When considering the evolution of Marketing Mix Modeling, picture, if you will, a weather forecast. If you were planning an outdoor event, this tool would be essential, as rain or snow could impact your logistics.

But what if this weather forecast was only available once a year and, fully static, left you unable to prepare for increasingly frequent “freak” weather events? All non-meteorologists can do is strategize around what they’re given and hope there’s no hurricane on the horizon. The comparison might be stark, yet traditional MMM isn't without similarities as a quintessential "black-box" tool. It produces crucial information but is only accessible now and then and reveals nothing about its internal workings.

Of course, MMM goes far deeper than any weather forecast. Nicolas Vaudran, Managing Director of Retail & Services at fifty-five France, defines MMM as "an analysis method to quantify the impact of marketing investments on business and brand goals." In short, using data science, MMM empowers companies to allocate their resources as efficiently as possible – a worthwhile yet imperfect approach in its traditional form that, after falling out of favor during the cookie era, has reinvented itself to suit modern demands.

In this article, we dive into how Next Generation MMM redefines marketing performance with key insights from our recent webinar - Beyond the Black Box: Revolutionizing Performance Measurement with Next-Gen Marketing Mix Modeling. The webinar featured expert presentations from Nicolas Vaudran, Managing Director of Retail & Services; Arnaud Parent, Director of Marketing Science; Nick Yang, Head of Media Tech; and Dustin Stewart, Data Science Lead.

PART 1: How to upgrade to Next-Gen MMM, and why?

I. The privacy-first approach: A drastic change in the MMM landscape

‍MMM itself is far from new. First developed in the 1950s, advertisers used this solution to measure the impact of their campaigns across television, radio, and print. In the 1980s, MMM expanded further as new data science fields developed, but the introduction of cookies and their browser-based Tracking capacities in the 2000s changed the game. From then on, marketers shifted from MMM to audience-level attribution en masse.

Recent developments have seen a genuine upheaval in this landscape, however. In recent years, political entities worldwide have started to crack down on data privacy with new legislation like GDPR (General Data Protection Regulation) and CCPA (California Consumer Privacy Act), whilst there is much greater awareness amongst the public around the capture and usage of their personal data. In response, tech giants like Apple and Google are giving back power over data to their users. Cookie consent banners are now ubiquitous, and customers are making their feelings clear – according to one study from Chile's Consumers National Service, if a website explicitly requests consent over data from its users, around 95% will refuse to share additional cookies.

A future where only 5% of the public can be tracked is becoming easier to imagine. Accordingly, all eyes turned once more to MMM and its inherently privacy-compliant nature. But to answer increasing expectations from marketing teams, Marketing Mix Modeling required something closer to a revolution: enter Next-Gen MMM.‍

II. Traditional MMM vs. Next Generation MMM

‍The best way to define Next-Gen MMM is by comparing it to its predecessor. While traditional MMM was useful in providing a big picture of marketing activities, it was also expensive, lacked granularity, and was difficult to activate. Typically, third-party agencies shared MMM findings once a year, delivered as an Excel sheet or a PowerPoint presentation, with no way to adjust mid-year once surprises arose. And as a perfect example of “black box” solutions, the methodology behind these models were rarely revealed.

By contrast, Next-Gen MMM takes advantage of recent advances in data science to propose a more complete, non-siloed view of marketing effectiveness by offering:

• A privacy-first approach
• A holistic process for decision-making
• Dynamic forecasting and scenario planning capabilities
• Granularity, speed, and deeper insights
• Pragmatic recommendations for strategic investments and activation
• Ease-of-access from a simple, comprehensive dashboard accessible to all stakeholders

As an added bonus, Next-Gen MMM is much more cost-effective than its predecessor – a true “north star metric” when considering investing in such a tool.

PART 2: The fifty-five Approach

I. Leveraging data, machine learning and expertise for optimal results: a method in 5-steps

‍As you might have guessed, a transformation of this scale is far from simple and requires an experienced guiding hand. Our team of experienced strategic consultants has identified the following central pillars for optimal results:

‍Data & Insights Artificial Intelligence Expertise

‍For leveraging the most relevant information from rich and trustable 1st and 3rd party sourcesFor combining techniques to drive deeper analysis, simulating the impact of future scenarios, automating processes, and guiding users in optimizing efficiency

For utilizing a cross-functional team's knowledge and expertise to ease adoption and providing an interactive solution to deliver actionable, pragmatic recommendationsfifty-five incorporated these key concepts into a five-step methodology for fully implementing Next-Gen MMM into your internal decision-making process:

1. Understand which core datasets are most relevant for your brand and ensure their reliability. Our data experts can harness ad spend, promotion, etc., depending on pertinence.
2. Once these datasets are established, use them to build a fully customized, machine learning enriched setup tailored to your specific business requirements.
3. Ensure that the model provided is fully transparent, and when possible, set it up entirely in-house to reinforce your ownership over your data.
4. Integrate your internal teams into the rollout process, technical and marketing alike, to ensure that they understand the model and can manage then maintain the solution themselves in the future.
5. And throughout the process, take a step-by-step approach instead of trying to solve everything all at once (and potentially omit details as you go).

At the end of this process, you should end up with a fully actionable dashboard designed specifically for you, easy to use and understand by your team, and leaving you free to run simulations and adapt strategies as often as needed.

‍II. fifty-five's use cases for a painless transition to Next-Gen MMM

‍As you transition to Next-Gen MMM, identifying and establishing your core use cases is critical. There is no such thing as a one-size-fits-all MMM model, and understanding which applications are most important to your business will shape the specific type of solution for you to onboard.

Remember that you can continuously adapt your initial setup with additional enhancements based on your evolving requirements. To ensure a seamless transition, fifty-five has created various pre-packaged modules that can be combined and tailored towards your specific needs:

• Performance monitoring - A fundamental starting point for most brands, and a core component of most marketing measurement solutions, including MMM. This use case provides a complete view of your ROI by channel, online and offline, and ongoing measurements of your core KPIs.
• In-depth analysis - Breaks down channel-level ROI results (ex: splitting paid search by region, social by campaign or audience, etc.) to enable more granular, tactical optimization decisions.
• Forecasting and scenario planning - Another core element within Next-Gen MMM is enabling advanced forecasting and scenario planning capabilities. By leveraging model outputs, this module creates a recommendation engine able to define optimal budget-planning scenarios alongside their forecasted performance.
• Experiments -While running various tests and experiments across your marketing activity, accurately measuring performance and verifying results is critical to the campaign lifecycle process. It can also be used to validate and further iterate the model itself as your requirements and data inputs evolve.
• Real-time monitoring - In addition to marketing ROI, several other metrics are vital for your business to target and monitor. Taking advantage of different modeling inputs can enable various views, including core business KPIs (brand index, customer lifetime value, new customer acquisition, etc.), operational measures (data governance adherence, active users), and broader macroeconomic factors (competitor activity, seasonality, inflation rates, etc.)

As no two solutions are the same, tailoring the right model to suit your unique needs is essential. At fifty-five, we strongly believe in the power of customization at every level, and our Next-Gen MMM approach is no different. Our experts can define which module best suits your situation and adapt as often as necessary.

Download our Ebook on fifty-five's In-House NextGen MMM Approach Before Getting Started on Your Transition

If you would like to know more about how fifty-five can help successfully guide you through your transition to Next-Gen MMM and check your marketing performance as easily as you check the daily weather forecast, please fill in our contact form for more details.