There are quite a few different methods for understanding our sales - which processes increase sales and which ones fail to help. There is no shortage of data analysis methods to see how our various marketing strategies are faring. Marketing Mix Modeling is one such method.
What is marketing mix modeling?
Marketing Mix Modeling (or MMM) is an analysis method that uses macro-level statistical analysis to ascertain how effective any given marketing campaign is by breaking down and analyzing data.
So, once you have carried out MMM analysis, you should better understand how much advertising channels, campaigns, external factors, and business decisions are driving your conversions. And what can you do with this? Well, if Facebook ads are the biggest contributor to higher sales, then MMM can suggest you spend more of your budget there and less on your webinars, for example.
This technique can help you to figure out which, if any, of your marketing techniques are the drivers behind your sales (including any fluctuations), or whether something outside of your control, such as the economic landscape, purchasing power for your customers, or seasons are major influences.
The pandemic is one such trend that is out of your control (for the most part), but it affects purchasing decisions.
There are a few benefits to using this method:
Some of the benefits of this method include:
Reliable data: this technique applies tried-and-true statistical analysis methods to a huge data set, so you are likely getting a very accurate and reliable picture of what's going on with your sales.
Deep analysis: you don't just get a hunch about what's working. There's no guesswork involved with this method and very little trial and error. You get a breakdown of multiple individual factors contributing to your sales, including non-marketing factors, such as seasons or political changes.
Accurate forecasts: this level of analysis allows you to predict future trends more accurately and prepare accordingly. Budgeting more accurately is a bonus.
Useful for non-digital channels: MMM can help you understand the impact of TV or billboards more than attribution tools which are more designed for digital channels.
Private: MMM doesn't use lots of data about specific individuals. It's more of a holistic approach.
Contribution, not attribution: MMM takes into account that the conversion journey isn't just 1 ad = 1 sale. It understands that all marketing and external factors need to come together to create the perfect storm to lead someone to convert. So instead of looking at each channel individually, it considers the whole marketing and industry landscape for your business.
Credit where credit is due: an in-depth analysis like this can help you to clearly show how much of a contribution your marketing team is making to sales, and how much each campaign or channel is contributing to sales. It might seem during a time of economic downturn that your team's efforts aren't making a big impact, so this analysis can help you to see their true contributions and make adjustments to your sales commission system.
There are a few drawbacks too
Drawbacks of the marketing mix modeling method include:
A little cumbersome: with so many variables to consider, this is not the most fast-paced method for analyzing the factors affecting your sales. It doesn't offer real-time analysis, for example. Depth of information takes precedence over speed. For this reason, it should be used occasionally to dictate upcoming strategy rather than day-to-day optimizations.
Traditional: this method was created with traditional marketing methods in mind, such as TV ads and billboards, which weren't as targeted and specific to different audiences. As such, MMM isn't as useful when it comes to accounting for variations in targeted messaging.
Average: MMM usually looks at averages over time, without accounting for unusual peaks and troughs. This means that some channels can seem like they're doing better or worse, in general, than they are because these peaks and troughs aren't excluded from the data or treated as outliers.
Customer experience: The customer experience, like the calls your inbound services receive, is not considered in the data, so it lacks some qualitative data that could be useful in explaining the reasons behind the performance of different channels.
Not suitable for smaller teams: this method requires a solid understanding of statistical analysis, which might not be realistic for smaller companies.
Steps for running marketing mix modeling
Now that we've explained what MMM is and the pros and cons of this method, let's dive into the steps for how to carry it out.
Gather the data
The first step is to gather your data from your marketing data warehouse! With MMM, we use statistical methods to estimate product demand produced by marketing tactics. Product sales are separated into two types of sales drivers: incremental and base drivers.
Base drivers refer to fundamental factors affecting the company which can't be influenced by short-term marketing interventions. Customer loyalty, for example, is the result of a reputation that has been years in the making. Price and distribution are also base drivers because they significantly affect company outcomes.
Distribution means which items are available, how many items are available, which stores they are located in, how many stores there are, and where those stores are located. Price directly affects who can afford your call management tools, for example, and it also communicates the value of your product to your customer.
Market conditions are also beyond your reach, most of the time. Certain times of the year are reliable for boosting sales, such as the winter holiday period, when kids go back to school, and the start of the summer holidays. We refer to this last one as seasonality.
So basically, anything affecting your sales that isn't the result of direct marketing intervention by your company falls under the base driver category.
So as you may have gathered, incremental drivers are any company outcomes that result from concerted marketing efforts, such as promotions, tv, physical, and YouTube advertising, sales events, and any other kind of promotional marketing activity.
Incremental drivers can be divided into two main camps and a third camp which combines the main two. ATL and BTL stand for Above The Line and Below The Line. ATL marketing consists primarily of brand-building, creating brand awareness, and building your brand's identity and story. This type of marketing is general and not hyper-specific in terms of the target audience. The aim is to create awareness of your brand.
The second type of incremental driver, Below The Line marketing, is the opposite in many ways. It is hyper-specific and targeted to a select audience. The aim is to drive sales rather than build brand awareness. The audience knows the brand at this stage, and it's time to get them to buy. This is more direct marketing to leave a memorable impression on the audience.
The third type of incremental driver is called Through The Line, or TTL marketing, and it involves a combination of ATL and BTL.
So why does all this matter? Well, these are the data channels you are going to analyze as part of your MMM.
Develop a model
For most Marketing Mix Modeling to work, you need at least two years of data to analyze. It can also take several weeks for the data analysis results to be complete, so you must factor that in when conducting your analysis. Because of the fast-paced nature of marketing and business today, many companies will have a continuous MMM on the go pumping out regular data.
One popular model for MMM is called Time Series Analysis. Its basic function is to predict future trends based on past trends. It's often used in various contexts, including highly scientific areas like biology or governments, to predict economic or political patterns. It's a form of forecasting that you could use to predict how the end of third-party cookies might affect you.
This isn't the only model you can use, just a popular one that people in different contexts commonly use.
Analyze the results
Once you have all of the data and your model is up and running, you can start looking at some of the model's outputs. You will primarily be interested in looking at three things:
How efficient each activity is
How effective each activity is
The MROI (Median Return of Investment) of each activity
That is to say, how appropriate each activity is at, e.g., generating sales, how good/fast it is at generating sales, and how much money is being made from that activity compared to how much money is being spent on that activity.
If advertising on social media does not reach many people, it is not very effective. If it is reaching them, but those engagements are not leading to many sales, it's not very efficient. And if they are leading to sales, but the ratio of advertising cost to sales is not optimal, then your MROI won't be great.
The final stage of the process is to look at your results and predict what would happen if you changed the areas you have control over. MMM can allow you to predict how spending more on particular types of advertising will affect sales. It could also help you forecast how a change in your audience's spending power would also affect sales.
This is the part of the process where you can start to see the benefits of MMM.
MMM is a very complex and in-depth modeling system
Many companies choose to forego this method entirely because of the sheer amount of work that goes into it, but it is a viable method for some businesses. Because of the fast-paced and ever-changing landscape of the business world, it can feel futile to put this much effort into predicting which marketing strategies increase sales. Still, it is technique scientists, statisticians, and sociologists have been using for many decades.
Richard Conn - Senior Director, Demand Generation, 8x8
Richard Conn is the Senior Director for Demand Generation at 8x8, a leading communication platform and a workforce optimization software with integrated contact center, voice, video, and chat functionality. Richard is an analytical & results-driven digital marketing leader with a track record of achieving major ROI improvements in fast-paced, competitive B2B environments. Here is his LinkedIn.