Google Analytics 4 is of limited use for attribution modeling. That’s where marketing mix modeling comes in.
Do you see more “unassigned” and “direct” traffic in your Google Analytics account even if you’re careful with your UTM tagging? That’s because data privacy regulations protect the users visiting your website.
As data privacy regulations increase, attribution models may struggle to stay accurate and valuable. Many use Google Analytics 4 for attribution modeling, but it is not GDPR-compliant. Because of this, attribution models like those from Google Analytics will be less effective. They won’t accurately show which channels are working. This is where marketing mix modeling comes in.
This article explores the growing relevance of marketing mix modeling today, how it differs from attribution modeling and how to harness it effectively within a strategic framework.
Comparing attribution modeling and marketing mix modeling
Attribution modeling and marketing mix modeling are two distinct approaches used in marketing analytics to understand the impact of various marketing activities on business outcomes. While both aim to provide insights into the effectiveness of marketing efforts, they differ in their methodologies, scope and application.
Attribution modeling. A set of rules that determines how to assign credit for conversions. These models use digital touchpoints in conversion paths. For example, the last touch model assigns 100% credit to the final touchpoint immediately preceding a conversion. There are also first touch, time decay, linear and data-driven models, to name a few.
Marketing mix modeling. An analysis technique that helps marketers measure the impact of their marketing and advertising campaigns. They can see how different variables contribute to their goals. Goals are often revenue, conversions, form-fills or subscriptions.
Simply put:
- An attribution model tells you what sales or marketing activity gets credit for a user taking a specific action. In marketing, we often use attribution reporting to see which actions caused specific conversions. For example, we can learn if opening an email led to filling out a form.
- A marketing mix model is a large regression model. A regression model is trying to understand the relationship between variables. An example of this could be weather patterns and revenue. You can still try to understand what actions lead to conversions, but a marketing mix model allows you to introduce more data. The analysis can then tell you the relationship between the variables. When the weather is sunny, more people visit your physical store, which leads to increased sales.
Both are valuable in understanding your marketing. You can also execute them using machine learning and coding.
How to approach marketing mix modeling today
To avoid errors, organize your analysis before working with these complex models. For instance, I was recently running through some monthly reports. We run attribution reporting for ourselves and our clients. Because I know our data so well, I flagged what looked to be some inaccuracies.
We spent two hours investigating and found big differences between our website data and Google Analytics. More specifically, the discrepancies were between the data API and the Google Analytics interface.
We know what the metric should be, and yet none of our data sources match. The issue is that we’re restricted to one data set for the attribution model. We can disregard the problematic data if we use a marketing mix model instead because data can be gathered from other systems to tell us what’s working.
Before we can get into the analysis, we have to take inventory. The example shows why you should gather requirements and have good data governance before using a marketing mix model. If you don’t have a good handle on what your data should tell you, a complex analysis won’t help.
To understand what we’re working with, we can use the “5 Ps” framework to determine your:
- Purpose.
- People.
- Process.
- Platform.
- Performance.
Purpose
This is where you’ll state why you want to run a marketing mix model. The best way to organize your thoughts is with a user story.
“As a [persona], I [want to], so [that].”
The user story tells you what the other Ps are.
- [Persona] tells you the people.
- [Want to] tells you the process and platform.
- [That] tells you the performance.
Here’s what mine looks like:
- As a CEO, I want to understand which of my digital marketing efforts are resulting in sales so that I can prioritize budget and resources.
In this statement, I have a lot of information. Let’s keep breaking it down.
People
I stated that I wanted to understand the data, so I’m the first person involved. Knowing that I am not solely responsible for data collection and analysis, I can assume that I’ll need my analyst involved. We will also need our business development resource to bring the sales data.
Process
I stated that my purpose was to understand my digital marketing efforts and sales. In terms of process, this statement tells me that I need to do a couple of things. I need to know how that data is being collected, the frequency and the format. This is where I need to have data governance in place so that the data collection processes aren’t what holds up running a marketing mix model.
Once I identify which systems I need to extract data from (in the next step), I can circle back up to the processes, ensuring that I can export the data needed. If I can’t, I’ll need to develop and work new processes into the overall plan. I will also need to create a process to clean and normalize the data once extracted to analyze data from different sources.
If I were getting the user story from a stakeholder, I would probably push back and ask for a more specific timeframe. This is where you will likely spend most of your time, between process and platform.
You can use a marketing mix model to analyze data from different sources. These sources may not have the same format, so you must create a process to combine them for analysis. The more data you want to use from different platforms, the more processes you’ll need to develop — especially if you want to repeatedly rerun the marketing mix model.
Platform
Using the middle of the statement again, I stated that my purpose was to understand my digital marketing efforts and sales. This tells me which platforms I need to be extracting data from because I want to understand sales data, which will be either my CRM or accounting software.
I also want to understand my digital marketing efforts. This means I need to first know all the digital marketing tactics and then figure out which platforms have data I can extract. LinkedIn, for example, is stingy with data extraction, so that could be a problem if that’s a channel I care about. I could easily end up with data from half a dozen platforms. Whereas with an attribution model, you typically only have data from one or two sources.
If I have a well-thought-out user story, I won’t get overwhelmed trying to collect data from all my systems. My user story states, “digital marketing efforts.” When I have many campaigns and tactics, I can focus on a few channels or a shorter date range to make it easier to handle.
Performance
This is the last piece of the user story. If you aren’t creating a user story with a measurable outcome, try again. In my user story, I stated that I wanted to be able to prioritize resources and budget. Well, that’s not a good outcome. It might be true, but it’s not super measurable. How will I know that I did that, prioritized?
The recommendation would be to go back to the user story and rewrite it to be more precise. A different version could say, “to lower spending on ineffective channels and increase it on successful tactics.”
You don’t have to do People, Process and Platform in any particular order. You might know the platforms which will inform the process and the people. But do not skip these Ps. If you skip gathering requirements and managing data, it can cause costly mistakes and wasted resources.
Looking back at my initial audit, I see that I have a lot of work to do before I can consider running a marketing mix model. Many teams will run a marketing mix model using code and machine learning. Having a plan before you even start with your code will make your execution more efficient. Instead of fixing problems in the data, you can spend your time fine-tuning and creating action plans.
The good news is that I can break it down into smaller, more controllable pieces. I can create repeatable processes to extract data and rerun the marketing mix model. Choosing this route means the upfront development will take longer. However, the process will be much more efficient when I need to re-run the analysis.
Embracing marketing mix modeling for comprehensive insights
A marketing mix model can be a really powerful part of your analysis portfolio. When working on a data project, setting yourself up for success is important. Requirements gathering and governance is the part we all want to speed through, but taking shortcuts here is not worth it. Take time upfront to make a plan; your analysis will be much more valuable and actionable.
The post Preparing for marketing mix modeling: What you need to know appeared first on MarTech.
MarTechAbout the author
Katie Robbert is an authority on compliance, governance, change management, agile methodologies, and dealing with high-stakes, “no mistakes” data. As CEO of Trust Insights, she oversees the growth of the company, manages operations and product commercialization, and sets overall strategy. Her expertise includes strategic planning, marketing operations management, organizational behavior and market research and analysis. Prior to co-founding Trust Insights, she built and grew multi-million dollar lines of business in the marketing technology, pharmaceutical, and healthcare industries. Ms. Robbert led teams of Microsoft Partner Software Engineers to build industry-leading research software to address and mitigate pharmaceutical abuse. Ms. Robbert holds a Master of Science degree in Marketing and Technological Innovation. She is a published researcher in the Pharmacoepidemiology and Drug Safety Journal. She is the Corporate Community Manager and Ambassador for Women In Analytics.
(18)
