In today’s privacy-conscious world, digital marketers are facing a fundamental shift: the traditional methods of tracking user behavior are disappearing.
With cookies crumbling and data privacy laws on the rise, the question remains: how can we still measure the real impact of our marketing efforts?
Marketing mix modeling (MMM) is again recognized as one of the cornerstones of marketing measurement. But choosing the right approach for implementing it isn’t straightforward. Should you invest in MMM software or lean on external consultancy expertise to guide the way?
We’ll unpack the pros and cons of each option and look at a third solution that takes a hybrid approach.
Understanding marketing mix modeling (MMM)
At its core, MMM is a multivariate regression analysis that dates back to the 1960s.
What does this mean exactly? Imagine you're running marketing campaigns across TV, social media and email and you want to figure out exactly which channel contributes the most to your sales.
Multivariate regression takes your data on factors such as ad spend, timing, promotions and pricing and shows how much each factor impacts your results.
The name is originally derived from the 4 Ps of the marketing mix (Product, Price, Place, Promotion). But in the context of online marketing, MMM is mostly concerned with modeling the impact your advertising has.
MMM offers a holistic overview of what is directly impacting your marketing outcomes.
Marketing mix models (MMMs) are like a backstage pass into understanding what really drives sales. They help you make smarter decisions on where to focus your efforts and how you might increase sales by adjusting your marketing (media) mix.
Core concepts of MMM
One of the huge benefits of MMM is that it doesn’t rely on personal data, so it’s privacy-friendly and not affected by GDPR.
Instead, it pulls in non-trackable signals like market trends and aggregated data to give you a better picture of what’s driving results.
This makes it a great option for businesses looking to measure marketing impact without dealing with user-level data.
Putting MMM back in the spotlight
Before new age “measurable” digital marketing, marketers tried to understand the effectiveness of their old-school marketing like TV and print by creating statistical models with advanced analytics.
Then, “measurable” digital marketing came in, and click tracking became possible. Using Google Analytics, marketers were able to see who clicked what and when.
Then came the attribution models, which were introduced to help marketers understand which interactions, such as ad clicks, might have contributed to conversions. However, these models don’t determine with certainty which clicks had the most impact. Instead, they assign weights to interactions — like first-touch, last-touch or multi-touch, offering a structured perspective on how marketing efforts align with conversions.
While this framework has been valuable, it has also been critiqued for its underlying assumption that "if an ad click preceded a conversion, then that ad likely played some role in causing that conversion." Although this relationship may exist in many cases, attribution models typically can’t prove causation. Their insights are best understood as correlational, meaning decisions based solely on attribution data can sometimes be misinformed.
Attribution models have evolved over time. Multi-touch attribution, for instance, attempts to distribute credit (or weight) across several interactions, providing a more comprehensive view but requiring marketers to assign values, a process that can introduce subjectivity.
Data-driven attribution refines this further by using machine learning to dynamically allocate credit. Data-driven models might look like they're generating amazing conversion insights, but they don't reveal any information about how they come up with their calculations. This lack of transparency can make it challenging to fully interpret or trust the results.
Additionally, recent privacy changes and reduced access to user-level data have limited the amount of input data available for attribution models. This has added another layer of complexity to an already imperfect system.
While attribution models are not without their challenges, they provide a useful starting point for evaluating marketing efforts. That said, their limitations underscore the importance of complementary approaches, such as marketing mix modeling (MMM), which can provide a broader perspective and help fill the gaps left by attribution models.
Marketing mix modeling together with incrementality testing are the go-to tools to add to multi-touch attribution (MTA) for a well-rounded measurement arsenal. This combination is known as measurement triangulation.
Measurement triangulation balances the strengths and weaknesses of each method.
Pro tip: We recommend against conducting marketing mix modeling (MMM) in isolation. Instead, it's beneficial to use a balanced measurement approach, such as triangulation, to achieve a more accurate understanding of your overall marketing effectiveness.
Implementing MMM — what are your options?
If you want to implement marketing mix modeling in your business, you have three main options:
- Buy a software as a service (SaaS) tool to manage your MMM as part of a wider measurement triangulation approach.
- Hire an external consulting firm.
- Create a hybrid way forward using both solutions.
So, how do you decide?
Most businesses will find a SaaS tool to be a great fit. It’s user-friendly, gives you valuable insights and saves you from spending a ton of time and resources.
But if your business has more complex needs or lacks the in-house expertise (or resources) to make the most of the software, working with a consultancy firm can be a smart investment.
Let’s take a closer look at your options to see which one might be the best fit for you.
Using MMM software
In short, marketing mix modeling software is capable of ingesting the data required, running the model and providing an output. All of this happens without the user needing a deep understanding of the technical side of the process.
What you will need, however, is a strong domain knowledge so you can bring the nuance of your setup to the tool — things like campaign naming conventions, regional splits or product feeds.
Working with an MMM consultancy
You might already have a sense of what hiring a consultancy entails, but hiring a consultancy for MMM would likely look something like the following:
A consultancy acts as an external data team that is responsible for measuring and reporting your marketing outcomes.
The key difference between a consultancy versus SaaS for MMM is that a consultancy firm becomes accountable for the accuracy of the data. You’ll still need the in-house expertise to draw conclusions from this aggregated data to inform your marketing strategies.
A hybrid approach
If you choose SaaS, you’ll likely need team members or outside experts skilled in data handling to help with the implementation and proper interpretation of your data model. Even with powerful MMM software like Funnel, having the expertise to interpret and make sense of the numbers is essential. That’s why we have a team of data scientists who work closely with our users to validate and identify the best model for their needs.
This hybrid approach combines SaaS with consultancy services. That way, your company benefits from the experience and technical knowledge offered by data scientists and you also benefit from the control that comes from managing some of your marketing measurement in-house with your own software.
A hybrid option can also be a cost-friendly way to get started if you need something more than software alone but can’t fit comprehensive MMM consultancy services into your budget. Additionally, some companies start with a hybrid approach and then move to a more self-service SaaS-focused approach later on.
Considerations when choosing between MMM software and consulting services
Now that we have the basics of each method down, let’s run through the factors that may impact your decision. But first, here’s the framework we’re using to compare an MMM consultancy to MMM software:
Flexibility
The core strength of SaaS is its plug-and-play nature. It's more accessible than the other options but at the cost of flexibility. After all, SaaS at its core is a technical platform serving customers in a way that’s cost-efficient and streamlined for all.
This is usually enough for most businesses that are getting started with MMM and triangulation. However, if you find yourself an outlier in terms of your needs, the SaaS route is likely not the right one for you.
The consultancy business, on the other hand, is predicated on coming up with bespoke solutions. So if you have unique needs, chances are that any good consultancy can sort you out.
Pro tip: Depending on how complicated your use case is, this will likely be reflected on your invoice, so make sure to set expectations upfront.
Skills required
If you find yourself with little to no skills regarding how to implement an MMM solution, interpret its output or use it for decision-making, then the consultancy route is probably right for you.
A good consultancy will not only do all the heavy lifting when it comes to building and tuning the model, they will also hold your hand and help you both ask the right questions and put the answers to good use.
If you have some idea of what you're after, or if you're not afraid of learning by doing, the SaaS route is probably the right one. It gets you up and running the quickest and should come with the lowest price tag.
SaaS vendors invest significantly, often far more than consultancies, into research and development as well as the continuous optimization of their MMM algorithms. By choosing the SaaS route, you benefit from these substantial investments without paying anything more than your subscription fee.
It's also worth noting that most SaaS vendors do provide onboarding services and ongoing help to make sure you're getting value, although it's not comparable to the white-glove approach of your high-end consultancy.
Time involved
The big benefit of SaaS tools is that they are built to be quicker and easier to use. Developers of SaaS tools literally spend their time improving the value a product provides and so this works in your favor for MMM. Whereas typically data ingestion requires a lot of data engineering, SaaS tools aim to reduce the time spent in this area.
The foundation of using a SaaS product as an MMM solution requires that you have all your data in place and ready to use. Funnel’s core product, for example, is our Data Hub, a centralized place for all your marketing data. It will automatically gather new data from your marketing channels and sales systems as soon as it’s available.
Consultants can expedite the process by leveraging their experience and established methodologies. They can often deliver results faster than an in-house team starting from scratch. However, the engagement typically involves a long project setup phase, which can take weeks to months.
The opportunity cost of an MMM project
Investing in an MMM project is like recalibrating your marketing engine. If you’re spending millions annually on advertising but not optimizing effectively, even a small misstep could cost significant revenue, so the recalibration can be well worth it.
However, depending on your total spend and how much of an impact optimizing your marketing measurement is likely to make, you may be better off using those dollars in other ways, such as on more ads.
Think of it this way: reallocating $250,000 to an MMM project only makes sense if it improves budget accuracy enough to deliver measurable gains above $250,000 over your marketing efforts. The key is assessing whether the insights you’ll gain are impactful enough to outweigh the resources you’re committing to such a large undertaking.
It’s a strategic choice: are you fine-tuning for greater efficiency or potentially spending on an upgrade that doesn’t deliver the returns you need?
MMM solutions for a data-informed 2025
For some businesses, working with a consultancy can bring strong MMM knowledge to the table but, they are comparatively expensive. Some consultancies start at around $40,000 just for a one-off analysis and a presentation.
Where SaaS tools are really good at providing easy-to-use, out-of-the-box solutions that don’t require huge time investments to get to value. They leave you in the driver’s seat with a high degree of self-serve.
If you’re looking to take just the first step, you could start by meeting with a leading SaaS provider. With Funnel, you could enjoy a 43% increase in return on advertising spend (ROAS) while saving 125 hours with automated reporting from your Data Hub each week.