By: Eric Siegel, Program Chair, Predictive Analytics World for Business
In anticipation of her upcoming conference presentation at Predictive Analytics World for Business Las Vegas, June 16-20, 2019, we asked Theresa Kushner, Formerly Sr Vice President, Performance Analytics Group at DELL EMC, a few questions about incorporating predictive analytics into business. Catch a glimpse of her presentation, Helping it Stick: How to Overcome Barriers to Predictive Deployments, and see what’s in store at the PAW Business conference in Las Vegas.
In your work with predictive analytics, what behavior or outcome do your models predict?
A: Usually the models that my teams have created are best at predicting a customer’s next move – the product they will purchase, the return they will make, the service they will require, the intention they might have to purchase at all. Of course, there is always the prediction – or forecast — of what the revenue or sales will be for the entire organization.
Q: How does predictive analytics deliver value at your organization – what is one specific way in which it actively drives decisions or operations?
A: One of the best ways to deliver value to an organization is to use predictive analytics to predict which customers are most likely to purchase. This kind of prediction is best managed with inside sales teams who can follow up and record customer responses so that the results of the predictions are quickly and easily measured. Not only does this effect sales results, but it also directs the sales person’s time more effectively by indicating which customers are the best to call on to make quota.
Q: Can you describe a quantitative result, such as the predictive lift of your model or the ROI of an analytics initiative?
A: Actual revenue is the measurement that my teams and I have always tracked. If models don’t create or save $, then what use are they? At Cisco, the predictive analytics models produced $4.2B of actual revenue over a 4-year period. At VMware, the models were expected to generate an uplift of at least 30 to 50%. At Dell, predictive models told us which customers were most likely to return products. The uplift for these models in testing was in the high 70% range, but as often happens, when models are put into real world situations, the uplifts drop. In most cases by half. Getting models used and productive is an art, not a science.
Q: What surprising discovery or insight have you unearthed in your data?
A: One of the most surprising discoveries unearthed in the data used for analytics and prediction was that the 80/20 rule is alive and well – 80% of the revenue is usually generated by only 20% of the customers. In some cases, this was more extreme – 95% of the revenue by 5% of the customers. Although these were large businesses, this revenue consolidation had major effects. For example, with a large networking company during the downturn of 2008, a great percentage of their revenue was coming from the financial industry. Without understanding how much, they ran the risk of misaligning sales and operational resources and not being able to compensate quickly enough for the loss of revenue.
Q: Sneak preview: Please tell us a take-away that you will provide during your talk at Predictive Analytics World.
A:Automation and self-service are the keys to the use of predictive analytics. It must be easy. It must produce results. It must enable the customer experience.
Don’t miss Theresa’s presentation, Helping it Stick: How to Overcome Barriers to Predictive Deployments, at PAW Business on Tuesday, June 18, 2019 from 2:40 to 3:25 PM. Click here to register for attendance.
By: Eric Siegel, Conference Chair, Predictive Analytics World for Business