Tech entrepreneur & co-founder at AUDIENCEX with 20 years experience developing integrated digital campaigns for leading brands & agencies.
getty
Here’s a hot take that not everyone will agree with: Multitouch attribution (MTA) is toast. Not today, not tomorrow, but it’s coming. As third-party cookies are phased out, marketers’ ability to track consumer journeys, something that used to have implications on what we promote across different channels and how, is changing dramatically. And in the very near future, attribution models that have been honed to work in today’s browsers simply won’t work.
That’s because attribution models like MTA require cookies to work well. And when cookies no longer track consumers across sites and apps (and they are already on the ropes), only users who are logged into a site (which can be as low as 5% or less of all users) will be reliably trackable.
All is not lost. Instead, we’re seeing the world move toward a revised version of media mix modeling, also called marketing mix modeling, or MMM.
MMM is not new (it’s been around since the 1960s) and works by analyzing sales, marketing and other data that quantifies the benefit of different marketing motions, including channel performance, creative performance and other factors—and how they drive sales over time. Today, many marketers use MMM for prediction and planning/budget allocation.
When done right, it’s a marketer’s best friend for assessing how past performance can help predict the future. For example, MMM can tell you whether you should be spending more on TV or mobile and even on nonaddressable advertising like radio and out-of-home media.
MORE FOR YOU
Deep Insights
MMM used to be done maybe once or twice a year and often requires two years of data to do a time-series analysis. And that’s tricky because you don’t always have two years of data on hand. But MMM today is a different beast than it was 50 years ago. We now have cloud architecture and more computing power, which means you need less data to get a signal, and the intelligence has improved.
While the methodology is the same, the actual tools available to execute those methods are vastly improved, and the ability to do it more frequently and with more automation makes the technology itself more web positive. It also provides marketers with an adequate understanding of where to place their dollars and optimize their media spend at the highest level. And when you do that, MMM offers robust analysis that delivers excellent results.
Back To The Future
Although many companies use both, many forward-looking marketers are starting to use MMM over MTA. At AUDIENCEX, we are helping our clients make the move to MMM so they’re ready for a cookieless future now.
To take advantage of MMM and these new capabilities, you need historical data and a safe place to share it (preferably in the cloud). You also need enough historical data to get a picture of what’s happening, the infrastructure to safely store and share it, and possibly a partner that can be part of your overall media buying workflow.
The best way to get started is to make it a team sport with your ad tech team. Talk to them (and the data engineers) about data availability, access and the current level of detail you have around marketing expenditures and business key performance indicators. You’ll want sales information, order information, subscription data and more. If you want your MMM solution to be automated once it’s in place (and you should), you’ll need to set up data pipelines to catalog historical data for this time-series analysis, setting a foundation to continuously feed the repository with performance and sales data to improve the prediction modeling.
You also need a solid understanding of the historical context of previous marketing campaigns and creative tactics leveraged. MMM offers a high-level, channel view of performance, guiding where strategic adjustments could help your performance goals. But it does not capture all the context or levers available to marketers. What you spend is the first decision toward marketing optimization, but you’ll still need a firm grasp of your tactical mix, your targeting parameters and creative assets in use. MMM shines a light upon where those deep dives ought to occur, and where there’s room for improvement or additional spending.
The Right Partner Matters
Collecting and harnessing data for these types of analysis require very specific technical skills in data engineering and data science, which means you may need to find a partner to help you. The right partner will be able to help you from start to finish. The right partner will also be able to tell you how to get started, the timeline to deployment and the value produced by this type of work. It’s wonderful to find out how each channel performs, but be sure to ask potential partners what actionable guidance they can provide from the MMM results.
Finally, your partner in MMM should be able to tell you where your channel strategy needs adjustments—meaning where to spend more or less—while also helping you determine where to place incremental funds.
If this sounds like a big leap and a time investment, it is. But once you’re there, MMM offers better modeling, better insight and a much softer landing when cookies (and MTA) inevitably disappear.
Forbes Agency Council is an invitation-only community for executives in successful public relations, media strategy, creative and advertising agencies. Do I qualify?
Go to Source
Author: Jason Wulfsohn, Forbes Councils Member
