Marketing Mix Modeling is a powerful statistical analysis on sales and marketing data to estimate the impact of marketing activities on sales. It is used by companies to measure and predict the impact of their marketing efforts.
MMM is also sometimes explained as Media Mix Modeling.
Let's dive deeper
Marketing mix modeling, or MMM for short, is an incredible tool that can help you identify which elements of your marketing are driving overall performance most.
In the episode of Funnel Tips, Alex breaks down the following:
- What MMM is, exactly
- Some analogies of how the analysis model works,
- Why you should explore MMM
You can watch the video above, or read the extended script below.
A definition of MMM
As Alex points out, MMM is a highly resilient, privacy-friendly, and data-driven statistical analysis that considers how various internal and external factors impact your marketing performance - be it sales or any other KPI.
In a modern multi-channel marketing campaign, you might employ broadcast advertising, Google Ads, paid and organic social media, public relations, outdoor transit advertising (bus stops, billboards, etc.), webinars, co-selling partnerships, promotions, and more.
That’s a lot to keep track of! Plus, the more different channels and complexity that you add to your campaign, the more people you will reach. (Ideally, people from within your target audience.)
However, with more different channels, you will also have a harder timer tracking the contribution of each of these channels. If you aren’t able to measure contribution for your channels, you may not realize that some of your marketing spend is more efficient than others.
An MMM analogy
To better illustrate this dynamic, Alex reaches into his past as a football (or soccer) player in the UK and US. With each team sending 11 players onto the pitch, it would be useful to identify that one player contributed 25 percent of the win from their individual performance. Especially if another player contributed only 4 percent of the win.
It’s incredibly difficult to make this determination just by standing back and watching the game unfold. As Alex rightly points out, the player who scores the winning goal likely didn’t bring the ball up the field all by themselves to score. There were likely a series of passes or a critical defensive challenge. Other players were moving on the pitch, which could have created the perfect space for the scorer to use.
It’s an entire micro-ecosystem that feeds off itself. Here, attribution is very hard to define.
Identify KPI's
However, if the team manager identifies a set of core KPIs that will be tracked for each player — say, number of passes, pass completion rate, assists, interceptions, goals, clearances, etc. This gives us an opportunity to employ an MMM-style approach.
Applying this Marketing Mix Modeling approach to one single game might not give you that much valuable data. In a marketing context, that may be trying to define performance attribution based on a single days-worth of data.
Now, if we view these KPIs across an entire season (or several months to a year in the case of a campaign), we can start to see super valuable attribution insights starting to rise to the surface. Those insights then, can be used to shape the marketing mix for next seasons - or your future campaigns.
Now that we have the basics out of the way, let’s explore the variable you may want to track using an MMM approach with a campaign.
What variables should I analyze for MMM?
The list of variables you can monitor is nearly limitless. However, we can group many of them together in a few categories.
- Calendar-based variables
- Media activities or marketing tactics
- External variables
- Internal variables
First, there are calendar-based variables of the market. Think of seasonal trends and major holidays that have an impact on your consumer’s buying patterns.
Next, we have media activities, or marketing tactics. This category is a bit of a catch all for your advertising. It includes TV, print, outdoor, display, direct, search, social, total media spend, etc. It can also include earned media mentions like those gained from your public relations efforts.
Third, we should consider external effects. This is a sort of “force major” category. It’s all of the factors that are out of your control like macroeconomic conditions, weather, natural disasters, competitor activities and more.
Finally, there are internal changes that arise from alterations in how you do business. This can include a change to your product distribution, changes to the product or service itself, price changes and sales process changes. This category is almost like the old 4 Ps of marketing (product, price, and place), with promotion being covered by our media activities.
By measuring the variables in these categories that are critical to your business, a MMM analysis can begin to identify which variable has the strongest contribution to changes in performance. In this case, performance can mean incremental sales. What sales would you not have had if it wasn't for that TV campaign, or that specific marketing tactic? But it can also mean a KPI such as brand perception, depending on what you want to investigate.
And remember, the more data you feed into the MMM analysis, the more factors you can draw insights from.
Why you should implement MMM
As modern digital marketers, there are an ever-increasing array of different tools and media that we can leverage to get the word out about our product or service. Some of those tools are easily trackable, but some are not. Plus, some tools lose a bit of their attribution tracking capacity as you blend aggregated data with other tools.
This all means that marketers need a way to account for as many below- and above-the-line tactics as possible, while also digging into which is the most valuable for their marketing campaigns.
Now, Alex does point out that the desire to precisely pinpoint exact attribution of each tactic is misguided. For instance, check out his video all about attribution. Especially with 3-party cookies falling out of style, even our digital tactics are becoming harder to track.
A birds eye view for you marketing strategy
With an MMM approach, though, marketing managers and CMO's can gain a wider and more holistic view of their marketing ecosystem. With that larger, more bird’s eye view, they can get a better handle on which “levers” should be pulled at different times.
If we look back to our variable categories, just by monitoring external variables and our media activities, we can gain an understanding of how fluctuations in the consumer price index are influencing the effectiveness of targeted digital spend versus broadcast.
Sounds useful, right? Here are 3 more analyses that can benefit from an MMM approach.
- ROI analysis
- Forecasting
- Pricing
ROI analysis
Which marketing investment is giving us the best return. ROI analysis is probably the most common use for marketing mix modeling. It gives business leaders a quick summary of where our money is being best spent. Then, we can make decisions about how to shift our investment strategies to keep getting the best return.
Forecasting
With forecasting and MMM, it’s less about using a crystal ball to see future sales and revenue, but instead about planning future marketing budgets and overall spend. After all, there are a lot of teams that will want to know what marketing spend will look like in the months or year ahead.
While we won’t get exact future figures from MMM, it can be used to roughly predict what those budgets and spends will be. This is still quite useful, though, since it allows other parts of the business to plan accordingly.
Pricing
If you drop the price of your product or service, will more customers flock to it? Will you steal market share from a competitor? Or, does a price decrease lower the perceived value of said product or service? These are all incredibly difficult and complex questions to ask, requiring equally complex analysis to arrive at a reliable conclusion.
However, with the right data and analysis model, you can begin to make predictions about how the market may react to changes in your pricing structure.
What are some of the disadvantages of Marketing Mix Modeling?
One of the problems with Marketing Mix Models, is that they require a lot of data. That might be a problem if you are working with a small marketing budget, or if your organization has only just began media buying at scale. (Most Marketing Mix Models require at least 2 years of historical data in order to use its forecasting abilities.)
And even if an organzation does invest a lot in marketing, it's often not so easy to collect all the aggregated data needed for an MMM analysis. Data is often collected in different places, like data silos. Only more data mature companies have the right solutions in place, such as a marketing data hub and a data warehouse.
Conclusion
As we've covered in this blog post, MMM can be well worth the investment, since it removes a lot of guesswork from a marketing campaign or yearly plan - and introduces more clarity instead.
Learn more
These might be interesting to you as well: