Introduction
Marketing Mix Modeling or Media Mix Modeling (MMM) has been around for nearly 60 years, but this marketing strategy is more important today than ever before. As a marketer, you want to know how hard your marketing budget is working and how you can get the most bang for your buck for both your online and offline advertising efforts. Marketing Mix Modeling gives you the answers.
In this article, we’ll give you an introduction to MMM to better understand what it is, why it’s important, and how you can implement it to enhance your marketing efforts, optimize your budget allocation, and determine an appropriate investment strategy for the future.
History of MMM
A good way to understand how MMM works is to understand why it was created and how it was first used. MMM became popular in the 1960s-70s when technology and the marketing landscape were much simpler. Kraft was one of the first companies to leverage MMM, but they certainly weren’t the last. When they launched Jell-O, they were able to choose between three or four television networks and magazine ads to promote their product.
Traditional deployment of MMM allowed marketers at Kraft to see how their campaigns were received in different parts of the country, at different times of the year – allowing them to understand which factors affected sales. For example, if Kraft was running TV ad campaigns in 5 different parts of the country and they noticed sales for Jello-O were high across the board during July, they could have used that information to increase TV ad spending in other parts of the country. If only southern markets responded to Jello-O ads during July—when temperatures were already high—they likely would have pursued a more seasonal ad strategy.
Data collected over time (ideally 1-5 years’ worth) and “experimenting” with factors like weather or holidays could help marketers create a strategy for successful future campaigns and sales forecasts.
MMM is about the “big picture” – a top-down statistical method where one or more models are leveraged to extract key information and insights from multiple sources of marketing, economic, weather, and financial data.
Harvard professor Neil Borden formalized the term marketing mix in his 1964 article, “The Concept of the Marketing Mix” and Borden’s inspiration for the term came from his associate, James Culliton, who suitably compared successful marketers to professional chefs. A successful marketer, like a great chef, must be able to work from a tried-and-true recipe, but must also be willing to adapt and experiment based on changing market conditions and customer demands.
Marketing mixes rely on consistently experimenting, analyzing, and gathering historical data to understand the impact of varying marketing elements for predictive forecasting. Marketing success in today’s dynamic world is a moving target, made up of factors that are constantly changing and impacting effectiveness including but not limited to promotions, seasonality, holidays, technology, consumer demand, the list goes on. With MMM, marketers can use these findings to build data-driven marketing strategies, optimize their marketing mix, and forecast future sales.
The only constant in marketing is change.
But why are marketers scrambling to get their hands on MMM? The simple explanation (and most obvious) is that marketers are losing the ability to track conversions via third-party cookies.
Multi-touch attribution has been the go-to for decades (we’ll dive into attribution models in another blog) because marketers LOVE cookies. Cookies have allowed marketers to track website visitors, improve customer experience, and collect data that helps target the right audiences. With iOS 14.5, Apple will require users to provide explicit permission for apps to collect and share data, significantly reducing tracking capabilities. This privacy initiative means only 10% of users will share their unique Identifier for Advertisers (IDFA) with apps, down from 70%.
Google also announced it will no longer use “alternate identifiers to track individuals as they browse across the web, nor will we use them in our products” which spells major changes for MTA marketers.
Marketing Mix Modeling is considered the gold standard technique for measuring marketing effectiveness.
- New policies from tech companies (iOS 14.5)
- Legislation (CCPA, GDPR)
- Signal loss
- Macroeconomic factors
- Cookie depreciation
- Tracking restrictions
Bird’s eye explanation of MMM
MMM is the analysis of aggregate data from a variety of sources – both marketing and non-marketing – over a multi-year historical period, adjusting for external influences like COVID-19, economic data, days of the week, weather, etc. This data is then used to develop a model that correlates specific marketing efforts with specific business outcomes, i.e. sales or conversions. Marketing mix principles are controllable variables and can be adjusted frequently to meet the changing needs of a target market and the dynamics of our unpredictable world.
Rather than relying on granular data other attribution models rely on, MMM provides macro-level insights into campaign performance to answer questions like:
- What happens to our revenue if I increase ad spend on Google Ads by X%?
- How much should I spend on marketing to get the most value for my money?
- What is the ROI for each individual channel?
- What is each channel’s contribution to revenue and conversions?
- How are holidays impacting our marketing channel performance?
MMM has a major advantage over other measurement strategies since it doesn’t rely on PII cookie-level data, making it fully privacy-safe, and it can measure both digital and traditional media as well as online and offline conversion outcomes. MMM is undoubtedly a more analytically sophisticated and powerful tool for marketers who want to create a comprehensive fact base for advanced marketing allocation and activity decisions.
So why haven’t marketers been using MMM this whole time? Well, it boils down to what mix modeling used to be versus what it’s evolved to today.
Traditional vs. Modern MMM
Traditional MMM was historically slow, very expensive, inflexible, not actionable, and showed dated data. Facebook (Meta) estimates manually building a model could take anywhere from 12-22 weeks to complete – yikes. And that doesn’t include maintenance and dev work to automate models on an ongoing basis.
Great news is that MMM is now more modern than ever and becoming more accessible thanks to advancements using artificial intelligence (AI), high-speed computing, and data science to deliver actionable guidance instantly.
Meta/Facebook has been working on Robyn, a semi-automated open-source MMM library to “reduce human bias by means of ridge regression and evolutionary algorithms, enable actionable decision making providing a budget allocator and diminishing returns curves and allow ground-truth calibration to account for causation”. If you’re not a data scientist or engineer with a lot of coding experience though, you’re not going to be able to pop into Robyn, plug in your datasets, and have a lovely model appear. Doesn’t quite work like that.
This is where Forvio comes in.
Forvio’s fully automated integration allows marketers to easily access data across major platforms like Google Ads and Facebook Ads in a preconfigured way that’s powered by Robyn’s statistical and machine learning code.
With shrinking budgets, billions of wasted dollars on allocating budget to the wrong channels, and multi-touch attribution proving to be more futile by the day, the team of data scientists and engineers at Forvio are on a mission to give marketers unbiased, accurate, and accessible data to power decision-making and ensure money is being spent the right way.
It just takes a few clicks when using the Forvio app to start seeing results.
To build a model simply connect to the data source(s), build a dataset, and click run. No need to code in-house or rely on a consulting company to build you a model. Forvio’s advanced app lets marketers easily see what channels are underperforming, where to best allocate marketing spend, and test different if-then scenarios to optimize your ROI.
