Machine learning is leveling the ad measurement landscape for disruptor brands

For large organizations, marketing mix modeling (MMM)—a statistical technique that quantifies the impact of marketing and nonmarketing activities—has long been a powerful tool for measuring and optimizing ad campaigns across channels. In recent years, as marketers have grappled with declining access to personal data amid a rapidly evolving advertising landscape, MMM has emerged as a vital, privacy-centric alternative to user tracking because it does not not depend on cookies or unique identifiers.

But disruptor brands—startups that are born online and use innovative approaches to challenge established companies—have traditionally shied away from MMM, viewing it as too time- and resource-intensive to build and implement.

Now, a large-scale first-of-its-kind experiment commissioned by Meta and conducted by Accenture shows that MMM, when coupled with machine learning techniques and granular data, can produce budget-friendly and actionable insights for marketers across all sizes and verticals, without requiring the use of personally identifiable information.

The work highlights that technical innovations have democratized marketing mix modeling and significantly reduced the resources required to deploy MMM. Today’s models can deliver highly accurate, cost efficient predictions that can be customized to include variables especially valuable to disruptor marketers—such as purchases, app installs and subscriptions. These insights can then be used to optimize strategies ranging from budget allocation and campaign tactics to creative approaches and messaging.

The experience of Bark showcases how today’s marketing mix modeling can deliver valuable insights for brands of all types. A company best known for its monthly subscription services for pet owners—including BarkBox, Super Chewer, BARK Bright and BARK Eats—Bark also sells its products via direct-to-pet-parent channels, online marketplaces and a network of retailers.