For performance marketing teams, media mix modeling (MMM) is a relatively established practice. And for good reason: MMM can be a very useful tool for measuring media performance, budgeting, highlighting the impact of non-marketing variables, and, with available data, measuring just about anything.
However, in today’s dynamic environment, MMM is also not the only – or always the best – solution for measuring the impact of media investments. Leading marketers now see MMM as a piece of the puzzle: a complement to cutting-edge techniques like incrementality testing and the data provided by media platforms.
Pros and Cons of a Standalone MMM Approach
At the highest level, MMM estimates media’s impact on revenue by observing how variation in media exposure relates to variation in sales. Once a business has built up a solid amount of historical media data (typically two or more years), it can use MMM to bring together factors ranging from spend, impressions, and clicks to non-media factors like seasonality and pricing. The resulting outputs can be used to quantify the impact of media on sales at a high, broad level.
Quick side note: “MMM” can refer to “media mix modeling” and “marketing mix modeling.” While the two are often used interchangeably, there is a subtle difference. Media mix modeling typically focuses exclusively on paid advertising efforts, while marketing mix modeling tries to measure how all marketing activities relate to sales, including things like PR, coupons, in-store events, and more.
Standalone MMM solutions come with a number of advantages and challenges. In the “Pros” column, media mix models can:
- Create high-level forecasting and budgeting scenarios
- Run long-tail, multi-level, and multi-stage conversion models (as long as you have the data to input)
- Quantify a wide variety of marketing tactics with less restructuring model-based measurement
- Showcase how non-marketing variables impact sales alongside traditional marketing variables
These are all valuable insights. But MMM isn’t without its “Cons.” These include:
- Limited granularity, without the ability to do campaign-level analysis
- They can often be expensive and slow
- Focused on the past and unable to project outside of historical bounds, with limits that growing brands can surpass quickly
- Subject to the rule of “correlation does not equal causality.” For instance, branded search is always correlated to sales - but is not always causal
Let’s dig a bit deeper into this concept of correlation and causality.
Correlation vs. Causality
MMM measures correlation: how different variables relate to each other. This is important but can also be inconclusive. Over a given period of time, it’s probable that many of your media and marketing tactics are highly correlated. It’s common to ramp up multiple efforts around a holiday, for instance, or a product launch. That makes it difficult to tell which specific channel is driving which particular impact.
Causal measurement, like incrementality testing, on the other hand, is exactly what it sounds like - it identifies the true cause of an effect. By running experiments, incrementality solutions can determine the exact impact of each media channel on sales. It enables much more granular and in-depth analysis and removes the guessing games that MMM can require.
A Three-Pronged Approach for Unified Marketing Measurement
As with many things in life, the answer to successful media measurement isn’t one or the other; it’s both. In fact, it’s three legs of one stool - a triangulated approach that unites MMM, incrementality testing, and platform-reported data for a comprehensive, wide, and deep view into the impact of your marketing efforts.
Why Measured for MMM
Here’s what that looks like at Measured. Our Measured Incrementality Model (MIM) brings together three types of measurement:
- Incrementality experiments that reveal the causal impact of media on sales, providing granular and frequent insights into paid media contributions to single conversion events
- Media mix modeling to measure non-testable channels and stabilize reads across your complete mix, delivering a directional assessment of all paid and non-paid activities across more complex user journeys
- Platform attribution that provides granularity at a campaign and ad-set level
This approach empowers marketing teams to automatically calibrate MMM results with incrementality test results and platform attribution data. It eliminates the need to stitch results together manually and makes it easy to holistically assess impact all in one place.
The MMM component of the Measured platform includes a number of features that help it outperform standalone MMM tools. With Measured, you get:
- Incremental conversions. Model weekly incremental conversions against weekly spend, then incorporate those results into a broader model that includes predicted incrementality from the Measured Cross-Channel Dashboard and observed results from geo-matched market tests.
- Diminishing returns. Explicitly identify diminishing returns within the models.
- Baseline comparisons. Account for the effects of seasonality and pricing with baseline (non-media driven) conversion results for each conversion type and week.
- Advanced features. Capture seasonality, industry category, channel, tactic, conversion type, and brand-specific test results, trained on prior three years of data.
- Enhanced accuracy. Trust standard statistical accuracy metrics such as VIF, R^2, and MSE, as well as gold-standard inputs like geo tests.
- Fully managed, automated API-based data architecture. Measured Marketing Data Warehouse is unique in the industry, allowing brands to onboard within four weeks.
Technology isn’t the only thing that sets the Measured Incrementality Model apart from standalone MMM tools.
We also bring extensive experience to the table, with decades of expertise in both MMM and incrementality (and how they can best work together.) We offer quick onboarding that leverages already onboarded data and test results for training, avoiding duplicate efforts. Perhaps more importantly, MMM is built into the Measured platform at no additional cost. That means you no longer need to pay a fortune in time and money to one-off MMM vendors.
Taken together, our technological edge, expertise and experience, and integrated model deliver real business impact. Measured customers spend less on MMM, get deeper insights with smarter MMM models, and gain a comprehensive, unified view of marketing impact that can’t be achieved with MMM alone.