A recent McKinsey article on machine learning quoted Author Ram Charan as stating “any organization that is not a math house now or is unable to become one soon is already a legacy company.”
In that spirit, here are three great examples of companies and industries that are using machine learning for competitive advantage – to predict behavior, drive desired outcomes, a
nd optimize business decisions. In each example, there’s an application to healthcare.
- Amazon - We are all likely familiar with the ‘You might also like…’ or ‘Frequently Bought Together…’ or ‘Customers Who Bought This Item Also Bought…’ sections of Amazon’s site, where the online retailer serves up additional suggested items to its customers. Amazon uses behavioral data (real world purchase data) and machine learning techniques, to continuously learn the best items to serve up to shoppers to increase sales. The recommendations are based not only on the products that are frequently bought together, but also on the shopper’s buying profile compared to buyers with similar interests and purchase histories. Imagine the ability to serve up health plan choices to current and prospective members in the individually insured market based on a person’s demographics, health and wellness profile, and interests.
- Netflix - The on-demand media company uses machine learning to power its recommendation system. Netflix algorithms make recommendations based on the relationship between the characteristics of the content and a person’s watching history (behavior). In healthcare this type of learning could be applied to serving up the right content in healthcare communications to motivate specific health behaviors. Examples include improving medication adherence by getting members to renew their prescriptions; improving HEDIS scores and Star ratings by motivating members to get their preventive care screenings; and reducing hospitalizations during flu season by motivating key demographic groups to get their flu shots.
- Target - The country’s second largest discount retailer made headlines in 2012 for marketing baby products to a high school girl, essentially ‘knowing’ she was pregnant prior to her parents knowing this news. Target uses sophisticated machine learning algorithms to understand the relationship of various product purchases to life events, such as pregnancy, marriage, job changes, college, among others. In this case, Target identifies a few dozen products, that when analyzed collectively, yield a predictive pregnancy score. Even more impressively, the retailer’s machine learning techniques enable it to predict the timing of the birth within a small window, so marketing campaigns can be even more highly targeted. Health plans could use similar machine learning techniques to understand condition-specific behaviors and target interventions accordingly.
There are tremendous learnings that healthcare can leverage from other industries, teaching us how to best harness machine learning to more effectively drive outcomes.
Colin Hill is chairman and CEO of GNS Healthcare. He can be reached at [email protected]