The Historic Evolution of Marketing Mix Modeling

The only thing that is constant is change. As business models, technologies, and consumer behaviors evolve, marketing has adapted and refined its methodologies, frameworks, and tactics. Today, markets are more dynamic, complex, and interdependent than ever before.

For nearly 40 years, Marketing Mix Modeling (MMM) has evolved alongside the swirling technological, economic, regulatory, and social changes that shape today’s business environment. MMM has consistently proven to be an effective solution for determining the impact of marketing activities on purchase behavior, guiding the decisions and resource allocations that enhance campaign performance.

A Brief History

In 1979 VisiCalc launched the first spreadsheet program for the Apple IIe. This new “visual calculator” inadvertently kicked off the “Go Go” era that built the Finance industry of the 1980’s, fueled by a heady combination of junk bonds and leveraged buyouts (made infamous by Milken and Kravis, respectively). At the same time, corporate America was grappling with one of the most severe recessions since WW2. For legacy brands, the LBO barbarians were certainly at the gate.

At the time, even behemoth companies relied almost entirely on the implicit knowledge of managers, and this was especially true for the Advertising and Sales: the annual marketing budgets were often a function of predetermined sales forecasts. So, when these large corporations unexpectedly started to buckle under new and increased pressures, top management had a sudden pressing need for greater top-down accountability across their organizations, especially in Marketing. But they had few options.

Besides VisiCalc, a suite of new technologies emerged that would revolutionize commerce. These are embodied by the supermarket scanner, connected to a new breed of smaller computers that could be made available at the retail store level. This enabled the roll-out of the Universal Product Code (the UPC) as an in-store inventory label. From its use in back room warehouses to the aisles, the UPC connected inventory with sales across retail outlets.

The system was rapidly instituted by none other than Sam Walton’s WalMart, which was in fierce competition to overtake Sears and KMart. For the first time in human history, suppliers could be linked electronically with retail chains. As common as it is in today’s world, this was a revolutionary upheaval.

The UPC POS system sparked a data explosion. Though marketing information systems were nascent, pressures on logistics and inventory management were met by demand forecasting that relied on econometric techniques first developed in the late ‘60’s.

These were ideas that leapt from academia and into the private sector, as econometricians were called upon to tie together immense streams (or rather, reams) of data at major American CPG firms. Suddenly, these corporate juggernauts were positioned at the cutting edge of MMM innovation.

During the 1980s, Marketing mix models were meticulous, time-consuming and resource-intensive projects. Constructing an accurate marketing model could take a year or longer, executed by the crème de la crème of econometrics and advanced statistics. The complexities and costs associated with MMM made it a high-barrier-to-entry system, accessible only to the corporate elite.

Its benefits were only available to firms that could meet three criteria:

• access to vast amounts of historical data at a national or market level;
• the financial muscle to acquire additional consumer data from syndicators like MAXI and Nielsen; and
• the ability to employ newly specialized, high-cost MMM consultancies qualified to guide the modeling process.

Even though the time required to process stacks of market-level data usually meant that the campaigns under study would be finished before the models themselves, the strategic guidance MMM provided was invaluable and it became the gold standard for marketing measurement. MMM became beacons for the Fortune 100 companies of the era.

• CPG companies maintained sprawling portfolios of products with complex distribution networks. Channel partners like Wal Mart relied on just-in-time supply chain management methods. MMM helped to identify the sales impact of marketing channels like television, OOH, and print, alongside promotional efforts like coupons or in-store deals. This granularity allowed these giants to tailor their marketing mix for maximum ROI for each brand, offering a new level of precision that changed the game.
• Automobile manufacturers, facing long product development cycles and capital-intensive manufacturing processes, often operated in seasonal markets influenced by macroeconomic factors like interest rate and oil price changes. MMM could enable them to allocate resources dynamically, channeling money into advertising or promotions, depending on what was most effective given the market conditions.

Eventually, MMM started to be applied to businesses with high consumer volatility, like technology and fashion, where trends shift more rapidly and without warning. MMM enabled businesses to elucidate their marketing budgets to stakeholders with a level of certainty that was previously impossible.

During periods of contraction, MMM can help mitigate the risks that can grind up marketing resources. But also, during times of plenty, MMM helps companies heavy-up on their spend to unlock revenue by optimizing their media mix, and dial in promotions, while factoring in media inventory discounts achieved by newfound financial leverage.

The exclusivity of access to MMM became a form of competitive advantage in itself. Companies that could afford to use marketing mix models were not just buying analytics; they were buying into a club of advanced decision-makers, setting themsleves leagues apart from their competitors still relying on managerial intuition or simple heuristics.

Recent Evolution of MMM

Throughout its history, MMM has gone through continuous refinement, and its relevance has been proven during both economic highs and lows. As marketing has evolved from customer-centricity to customer-driven personalization within consumer ecosystems, today’s MMM more accurately captures the marketing tactics, baseline and external factors, and purchase behaviors, using a top-down statistical approach that quantifies their impact on target KPIs. MMM is not restricted to any single channel; rather, it works with the entire media mix, encompassing both digital platforms and traditional media. By remaining channel-agnostic, MMM provides an impartial lens to evaluate online and offline channels alike, irrespective of their data collection or reporting mechanisms.

The evolution of MMM has brought it out of a black box. It is now a collaborative tool that provides transparency crucial for businesses to make clear-eyed decisions or take additional steps to establish causality: like comparing the results from an MMM model with incrementality tests and attribution data.

Ultimately, marketing mix models are at the service of better informed marketing decisions. Streamlined MMM has increased power, and can be executed at a fraction of the cost and time. As campaigns roll out, subsequent models can be available to check accuracy and support predictive insights.

Today, as the marketing ecosystem continues to endure continuous renewal, attribution methods for evaluating marketing efficacy have been losing ground to MMM as a source of truth. While the insights from attribution have arguably always been unreliable, these methods are further compromised by changes in privacy laws, and new technology shifts that safeguard user data (like iOS updates, and the deprecation out of third party cookies) — user-level signal loss compromises attribution. Inarguably, online advertising platforms are becoming increasingly opaque.

These are the exigencies that have pushed MMM to innovate significantly in months. Notable industry contributions, like Meta’s Robyn and Google’s Lightweight, along with the emergence of innovative automated platforms like the one we’ve developed at Momentum, will play a pivotal role in this evolution. These MMM platforms advance the field by seamlessly integrating machine learning with more nuanced statistical analysis, enabling MMM to leverage a broader spectrum of data sources.

These new MMM systems are particularly useful for emerging brands that rely on a mix of digital channels, like podcasts, or larger brands that have already branched out to marketing channels like connected or linear television, radio, and OOH. New MMM systems can provide a transparent, data-driven understanding for each channel’s impact on ROI, so that informed budget decisions become much more straightforward. Rather than depending on historical trends or intuitive guesswork, by using MMM you can allocate resources based on verifiable outcomes. This changes the budgeting process from an exercise in speculation to a strategy founded on empirical data, ensuring that every dollar spent is accountable and aligned with the company’s broader business goals.