Concerned about impending browser policies that could limit cookies? You’re not alone. As businesses evolve to meet consumer demands for more privacy, the fundamentals of media measurement are shifting too. Marketers are still coming to terms with how the potential for cookie loss will affect their businesses in the coming years.

It’s widely anticipated that the need to move away from identity-based analysis will have a lasting impact on how businesses measure media, the data they collect, and the metrics used to understand performance.

In search of solutions, global-reaching brands such as ASUS and Resident have added a time-tested strategy to their measurement toolkits: marketing mix modeling (MMM). Once thought to be too involved, too slow, or matter only for TV advertisers, MMM has proved highly resilient in the modern era due to its reliance on privacy-friendly, aggregate data to measure sales impact and marketing ROI.

MMM can be used to inform cross-channel budgeting decisions, measure holistic impact of media, and support strategic and tactical optimization across platforms — needs that never go out of style.

More versatile than you think

Before we cover implementation, let’s debunk common misconceptions that hold some businesses back.

• Myth #1: MMM is only for brand advertisers that want to measure TV

In the past, MMM was mostly used by businesses that didn’t have granular signal data. Now, it’s gaining traction with direct-response advertisers since it helps bridge gaps in existing measurement systems.

Today’s MMM is more accessible. It’s most often used by clients that have a diverse media mix (with no one media channel over-indexed on spend), and up to a year of performance data to draw from.

• Myth #2: We don’t need MMM since we already use multi-touch or last-click attribution

This may have been true in the past — but it likely won’t be true in the cookie-less future. Most attribution models will be impacted by less access to third-party data. It’s important to use multiple measurement systems so you get more reliable results.

That’s where MMM comes in. It can be used to complement other attribution learnings, bring you closer to a true understanding of performance, support key budgeting and optimization decisions, and bolster scenario planning. And MMM can be continually improved upon too. When calibrated with lift measurement, it can increase your measurement accuracy and prevent overreliance on flawed attribution models.

Walden University, a leader in online education, uses MMM along with multi-touch attribution and other measurement tools. This strategic approach recently increased marketing contribution while increasing brand favorability amongst prospective students by +186%.

• Myth #3: MMM is resource-intensive — and it takes too long

The research and methodology behind MMM have evolved significantly. New innovations have dramatically reduced resourcing and time requirements.

Modeling insights can now be generated monthly, or more often if needed, based on just six-to-twelve months of granular historical data.

Two ways to implement MMM for your business

• 1. Do it yourself

Many businesses start with a DIY approach to MMM, but knowing where to start can be challenging. One option you may want to consider is Robyn, Facebook’s open-source, out-of-the-box MMM code.

Resident, the online home goods business behind brands such as Nectar and DreamCloud, used Robyn to optimize its budget allocation, resulting in a 20% increase in revenue. Faced with challenges brought about by COVID-19, and changes in the ads ecosystem that made it harder to measure effectiveness, Resident needed a way to measure performance across all its marketing channels.

It took the Resident team only five days to implement Robyn, compared with the five projected working months it would have taken to implement its own in-house model.

• 2. Work with a partner

There’s a world of partner options available to meet your business’ modelling needs — from self-serve SaaS platforms, to full-service models that support data collection, interpretation, and recommendations. If you’re interested, the most recent list of Facebook’s verified business partners can be found here.

Enterprise tech company ASUS recently partnered with Analytic Edge, a SaaS solution, to achieve its goals. MMM SaaS models are designed to be more scalable and cost-efficient than traditional MMM, and they rely on less data due to their use of AI. ASUS’ analytics team worked with Facebook and a partner on a pilot test to develop and execute this model across various sales channels. As a result, ASUS shared that they saw an unexpected 40% increase in incremental sales volume.

Look back to move ahead

The term “innovation” doesn’t always describe the newest, shiniest technique. Sometimes, it’s about being resourceful and willing to look at something in a whole new way. While MMM isn’t new, brands are using it to successfully update their measurement and optimization strategies for a cookie-less future.

As the ads ecosystem changes, advertisers have to be willing to change too. MMM is just one piece of the puzzle. Learn more about how measurement can help you optimize your marketing, and read the latest news around how Facebook is enabling privacy-enhancing, personalized ad experiences that help you better measure the value of your ads.

Swetha Subramanian, an advanced analytics and marketing strategy expert, is a Marketing Science Partner with Facebook Canada.