By: Eric Siegel, Program Chair, Predictive Analytics World for Business
In anticipation of his upcoming conference presentation at Predictive Analytics World for Business Las Vegas, June 16-20, 2019, we asked James Taylor, CEO at Decision Management Solutions, a few questions about incorporating predictive analytics into business. Catch a glimpse of his presentation, Backwards Engineering: Plan Machine Learning Deployment in Reverse, and see what’s in store at the PAW Business conference in Las Vegas.
Q: In your work with predictive analytics, what behavior or outcome do your models predict?
A: The models we build – or that our clients build – vary widely. We always start by defining the business decision our client needs to improve and then identify the analytic models that will require. Examples include model to predict how likely it is that a medical report describes a particular condition, how likely it is that someone has an undisclosed condition, what the most effective product sequence is for each customer segment and the riskiness of an auto loan.
Q: How does predictive analytics deliver value at your organization – what is one specific way in which it actively drives decisions or operations?
A: All the models we work with are targeted to a specific decision. We always begin decisions first – identifying the decision we want to improve and what kind of analytics will improve it. We only build analytics that we know we can use – and that we know how we will employ them once they are deployed.
Q: Can you describe a quantitative result, such as the predictive lift of your model or the ROI of an analytics initiative?
A: We never like to talk about “lift” as an analytic is never a complete solution – you should always think about business value, business ROI, not analytic accuracy. Once client has used a decision model to combine several analytic models and a bunch of business rules to get 14-14% accept rates on cross-sell offers and generate millions of dollars in new business. That was a pretty good ROI…
Q: What surprising discovery or insight have you unearthed in your data?
A: The most surprising discovery we have unearthed is how often the analytics team does not really understand how the business would use the analytic they are working on. Business teams often need much less accurate models than the analytics team assume they do and will often use them differently. Getting the analytics to team to listen to the business, not just the data, is really key.
Q: Sneak preview: Please tell us a take-away that you will provide during your talk at Predictive Analytics World.
A: Building good analytics is not the hard part. Even deploying analytic models is not really hard anymore. EMPLOYING analytics to improve decision-making in the real-world is the hard part. Unless you understand what your business partners will do with your model, and how they plan to do that, before you build the model you are likely to be disappointed in your results. Begin with the end (the decision) in mind.
Don’t miss James’ presentation, Backwards Engineering: Plan Machine Learning Deployment in Reverse, at PAW Business on Tuesday, June 18, 2019 from 10:30 to 11:10 AM. Click here to register for attendance.
By: Eric Siegel, Conference Chair, Predictive Analytics World for Business