Keen Brings SaaS to Marketing Mix Modeling
Marketing mix modeling is enjoying a resurgence as signal loss due to privacy changes challenges the popularity of multi-touch attribution. When digitization rendered so much of marketing measurable, the bottom-up approach associated with MTA became de rigueur. CMOs proclaimed (some still proclaim) that every little thing could and should be measured, and mixes could be calibrated on the basis of all that granular data.
With changes like Apple’s App Tracking Transparency and Google’s forthcoming elimination of the third-party cookie, plus new privacy laws such as the California Privacy Rights Act, measuring everything isn’t looking so hot anymore. So, marketers are returning to MMM, a top-down approach that attempts to identify how much various marketing tactics are affecting overall performance. As the name of the practice suggests, marketers can use those insights to calibrate their ongoing mix.
But how exactly is a marketing organization to get started with MMM? Historically, the practice has required working with a consultative partner stacked with data scientists. They dive into reams of data and come back with a lengthy presentation intended to guide marketers based on past marketing performance and, in some cases, market conditions.
This is where Keen comes in. The company believes it can offer a version of MMM updated for the 2020s in the form of a SaaS-based platform (with some degree of managed services). Less expensive than competitive solutions, Keen believes it can also offer more up-to-date, holistic guidance than MMM consultancies. The company aims to get many of its clients to a fully self-service model within 12 months.
“There needed to be less of a focus on strict measurement and more of a focus on complexity and fragmentation of media channels in a way that actually guides decisions and ties marketing decisions to financial outcomes,” said CEO Greg Dolan. The company’s chief executive came of age as a CPG marketer, and some 90% of Keen’s clients are still in that vertical, though it is expanding to others.
As those familiar with the martech managed-service-versus-SaaS continuum know, Keen’s competitors would argue that managed services are necessary not just to deliver properly contextualized, sophisticated MMM-based recommendations but to build trust in those recommendations. It’s one thing for a model to spit out marketing advice. It’s another for a company’s CFO to trust that advice and allocate millions in spend based on it. Data scientists can help make the case.
Keen’s rejoinder is that its model can replicate the accuracy of teams of data scientists “cheaper” and “faster” while being “more predictive,” to quote the three differentiators board member Randall Beard ascribes to the technology.
“What we’ve done is take the brains of all those data scientists and put it into an algorithm and a … model that can leverage information and deliver results that can be trusted,” Dolan said. “We’re building trust over time with the marketer when they see that our model can predict with 97% accuracy, and that’s something our competitors cannot do because if they do have a simulation tool, it’s based on a point in time.”
Other MMM companies that lean more toward the managed service model also contend that they take market conditions into account in addition to past marketing performance when calibrating recommendations.
But what’s clear is that, as MMM surges back in popularity, the space is getting its own SaaS-agency debate — just like ad targeting, campaign creation, and other sectors of the martech universe. And SaaS is cheaper. So, especially with a recession looming, it will be up to incumbents to prove that their teams of data scientists are worth the cost — and up to Keen to prove it can do more for less.