A Tale of Two MMMs

Marketing Mix Modeling vs. Media Mix Modeling

Marketing Mix Modeling and Media Mix Modeling are used interchangeably but they’re not the same.

Marketing Mix Modeling derives its name from the “marketing mix,” a concept that was popularized mid-Century as The Four P’s: for Product, Price, Placement, and Promotion. The four P’s model was expanded to include more P’s, like People, Process, Psychology, and Physical evidence. And recently, the marketing mix is evolving into the 7 T’s (or, the seven tactics) of marketing: Product, Service, Brand, Incentives, Distribution, Communication, and Price. Regardless of these classifications, that’s the “marketing mix” in Marketing Mix Modeling. This is often referred to as “big M” marketing in order to differentiate it from advertising. It’s the conceptual basis of the engine that drives nearly all of commerce.

Commonalities in Methodology:

Both are tools that guide marketers to allocate resources in a way that maximizes a target KPI (like website visits, sales, or ROI). Here’s how they are similar:

Both Marketing Mix Modeling and Media Mix Modeling are regression-based techniques aiming to quantify the impact of marketing activities on target KPIs.

They are retrospective in nature, analyzing historical data to derive predictive insights.

Both methods assist marketers with decisions for budget allocation, planning, and forecasting.

Distinguishing Features:

While a Marketing Mix Model offers a 360-degree view of all influencing factors, a Media Mix Model hones in on optimizing advertising spend. A Marketing Mix Model integrates everything from economic conditions and competitor actions to customer experience, providing a well-rounded strategy for predictive scenario planning, making it a go-to approach for businesses seeking a holistic view of their market environment. Media Mix Modeling is a more streamlined subcomponent of the marketing mix: it uses data-driven insights from media channels only to measure the impact of advertising.

The definitions of each will help illuminate the distinction.

Defining the Models

Marketing Mix Modeling (MMM): This refers to the statistical analysis and quantification of past marketing activities (both traditional and digital) to determine their effect on sales. Today, Marketing Mix Modeling combines machine learning with advanced statistical analysis to identify relationships between all brand-related marketing activities and the influence of pertinent external variables to demonstrate their effects on target KPIs. The models measure past marketing performance and simulate future scenarios. MMM enables marketers to optimize the media mix and reallocate their budgets efficiently across channels, products and regions, while helping forecast the impact of future events on campaigns.

Media Mix Modeling: A subcomponent of MMM, Media Mix Modeling specifically quantifies the impact of different media channels on sales or other target KPIs. This means it focuses exclusively on the effects of media spends, such as TV, radio, digital, print, etc., on outcomes.

Marketing Mix Modeling not only applies to media channels, it also evaluates in a wide range of external variables that affect purchase behavior, which can include: weather, seasonality, economic trends, competitors’ activity, recall, brand equity, and experiential factors like ease of checkout. Marketing Mix Modeling incorporates the interactions between these different elements, providing a holistic analysis for performance evaluations and scenario planning.

A Use-Case for Marketing Mix Modeling

Let’s take the following use-cases to highlight the scope, application, and type of recommendations each method can offer.

FusionAudio is a (fictional) consumer electronics startup based in the US. Established in 2019, the company has quickly gained attention for its singular focus on creating an advanced gaming headset that aims to dominate a fiercely competitive marketplace. It offers two models of the headset, its Premium flagship headset, and the less-expensive model with less features (reduced AI enhancements like spatial audio and noise cancellation). With an upcoming round of financing tied to a product expansion into other gaming peripherals, the executive team understood the urgency to optimize their marketing operations, keeping a disciplined eye on Customer Acquisition Cost (CAC) and Lifetime Value (LTV). In addition, FusionAudio would need to rebrand their company because it planned to introduce gaming peripherals that went beyond audio headsets.

FusionAudio had previously relied on Media Mix Models only, but given the pressures of expanding during an economic downturn, the company decided to try Marketing Mix Modeling, to capture a variety of variables such as weather, seasonality, economic trends, and competitor activities. MMM revealed some counterintuitive findings.

• Seasonality and Economic Conditions
• How MMM Revealed Counterintuitive Insights for FusionAudio
• Seasonality had a significant impact on sales for their headsets, not only did this reveal itself during the holiday season but the summer promotions had a far stronger impact on ROI than previously anticipated. The company doubled down on both Q2 and Q4 marketing campaigns, aligning promotional activities and discounts with summer breaks for students that carried into the high-impact holiday season.

The economy does not always cooperate, and FusionAudio operates in a period of economic downturn, finding that consumers shied away from their premium headset. The company leaned into its value model, especially during the summer, encouraging a tactical pivot in advertising language to employ value-based marketing messages.

Market Expansion and Competition

Strategic Decisions Informed by MMM

Opening new online storefronts in international markets resulted in a marginal increase in sales and it increased CAC. That was a problem. FusionElectonics decided to strengthen existing markets before international expansion.

Instead, the company increased benefits within its successful loyalty program at home, by leaning into events within their gaming influencer network, aiming to improve LTV. This incentivized repeat purchases and referrals, which showed up in ongoing models as data was refreshed. This also paved the way for new offerings to the FusionAudio customer base.

Aggressive promotional activities from competitors led to a temporary dip in sales. Management had feared a larger impact, so FusionAudio continued to focus on differentiating the brand while dialing in the PPC and SEO budget to maintain visibility.

Media Channels

The Limitations of Media Mix Modeling in FusionAudio’s Strategy

FusionAudio discovered diminishing returns on social media ads, which it identified as Adstock carryover effects. They found that while social media and PPC advertising offer higher ROI, reallocating ad spend to TV advertising boosted brand recall, which was important for the rebrand, while contributing to an increase in overall sales and improving resilience to competition.

Choosing Between Marketing Mix Modeling and Media Mix Modeling: A Matter of Scope and Business Goals

Previously, the Media Mix Model was helpful for rebalancing ad spend across PPC, social media, and email marketing based on insights into channel efficiency. Because it only focused on the performance of the last category of the marketing mix, it didn’t take into account the other variables. Though it exposed the Adstock effects — the diminishing returns on social media — it didn’t provide insights into how to optimize spending based on broader market conditions, nor did it take into account the importance of the loyalty program for driving ROI and LTV.

Considering FusionAudio’s business goals, Marketing Mix Modeling is clearly the better choice. Media Mix Modeling is limited in scope, focusing solely on media spend efficiencies, a rather myopic approach given the multiple variables affecting an e-commerce business.

While both Marketing Mix Modeling and Media Mix Modeling are used to understand and optimize the impact of marketing activities, their scope is what differentiates them. While MMM offers a broader perspective encompassing various marketing activities, Media Mix Modeling narrows down to the effects of media channels alone. Depending on an organization’s business goals and the granularity of insights they seek, they may choose to deploy one or both methodologies.

MMM vs. MMM