I spent most of my time at SXSW attending data-focused talks and events, which had a clear theme: the predictive power of data. Just a few years ago, analytics within marketing was focused on “bringing data to the masses” via self-serve visualization and dashboard tools. Over the last year we have seen a major shift to the predictive/prescriptive power that large datasets can offer via machine learning algorithms. While we expect to see continued exponential growth in data for the foreseeable future, McKinsey points out that “By 2018, the United States alone could face a shortage of 140,000 to 190,000 people with deep analytical skills as well as 1.5 million managers and analysts with the know-how to use the analysis of big data to make effective decisions”. While the opportunity has grown exponentially, the talent has yet to catch up (STEM PhD’s don’t grow on trees, after all).
Etienne Bernard from Wolfram Research (the brains behind the technical computing powerhouse Mathematica, and one of my favorite tools on the Web, WolframAlpha is working to realize this goal by “injecting machine learning everywhere.” There are countless unsolved real-world business problems (such as customer acquisition/churn, logistical optimization, and recommendation engines) in which the ability to predict the future can help drive millions of dollars in efficiency and untapped revenue growth potential. In many companies, these applications are in their infancy: the true power lies in embedding these algorithms directly into business processes.
What does it mean for your business?
The adoption and growth of open source tools such as R, KNIME, and Scikit-Learn have gained popularity over the last five years in helping marketing teams develop more effective targeting models. We have also seen the rise of advanced machine learning platforms like DataRobot that empower both technical and non-technical users to build highly-predictive models in a fraction of the time of conventional methods.
As user friendly (and highly powerful) tools such as DataRobot continue to gain traction in the marketplace, we predict that the analytical talent gap in McKinsey’s model will shrink over time – for the same reason that data visualization tools such as Tableau have empowered business users to “bring data to the masses” via rapid visualization. This new breed of tools is empowering those same business users to “bring machine learning to the masses” via rapid model training. Over time, companies will become savvier (and more comfortable) in injecting machine-learning-based solutions across the enterprise, and brands will ultimately see the benefit.
In the healthcare space, we have seen this algorithmic approach begin to gain traction. We see direct applications of this approach as low-hanging fruit for non-personal promotion and programmatic media on the HCP and Consumer sides, respectively. SSW Analytics has led the charge with respect to innovation over the last few years in this regard. For example in applying machine-learning based approaches to HCP behavioral targeting models that helps brands target and prioritize HCPs for non-personal promotion based on their likelihood of adoption in the launch phase (and ultimately, their likelihood of loyalty in the mature lifecycle stage). On the patient side, we have developed machine-learning-based models that identify and predict which patient types are least likely to remain adherent to a given therapy that allow us to optimize the effectiveness of patient support programs. We predict that over the new few years, this machine-learning-based approach to predictive behavioral targeting will become the norm rather than the exception.