Predictive Modeling - Chapter 4

Business leaders have been facing the cave wall watching the shadows of an Information Revolution that has shaped our decision-making process in the last couple of decades. We’ve reached the limits of spreadsheets and macros to help with our decision-making. Today, the ability to see the “angles” in a given market requires a higher order of intelligence just to navigate a highly commoditized business environment. Artificial intelligence, in the form of a new Predictive Modeling paradigm, is breaking the chains that bind, and only those leaders who choose to leave the cave and see the new reality that will succeed in the decades to come.

Unfortunately, predicting with certainty what a human being will do at a given time when presented with a given option at a some given price, is impossible. Prediction is all about probability or something close akin (e.g., “likelihood” or “confidence” that some given thing will happen). Machine Learning is about taking Big Data as input and building a Predictive Model that might be used to “score” the likelihood of something happening. 

By construction, the machine is learning to be “creative” in finding new associations and relationships in the data to provide actionable insight you didn’t see, or perhaps you observed the raw data but couldn’t comprehend its deeper meaning.

What are some applications that we can use in selling more effectively in this hypercompetitive market? Well, we know that client retention is more of a predictor of business stability than client acquisition. In other words, keeping clients buying from us is a more certain way to grow the business as compared to attempting to acquire new clients. With that in mind, here are some concrete applications of how Deep Learning can help us retain clients:

Cancellation: What if we could predict ahead of time the likelihood of a client canceling? If we had this list of potential cancellations ahead of time, we could embark on some type of customer reinvigoration campaign. We would call the client and do a free, one-hour consultation to answer any questions they might have, or we would schedule an onsite visit to uncover and address any problems or concerns.

Renewal: If a client is up for renewal, which clients need to be contacted to ensure they follow through? Let’s say we are leasing broadband services to a large B2B company, and we want them to renew instead of opting for a competitor. The machine would alert us about which clients are at risk of not renewing, and the salesperson would then reach out with a renewal discount or other customer appreciation incentive.

Upsell or Cross-Sell: What if the machine could predict which clients are ready to either upgrade their product or service (upsell), or consider buying another product or service (cross-sell) the company offers? The customer wins because they are being offered something they need (but didn’t ask for), or hadn’t considered that another product might help them run their business more effectively. This is a so-called preemptive sales approach that, if done correctly, adds value to the customer experience.

Authors:

Victor Antonio holds a BSEE and MBA degree and is currently an international sales trainer and sales development consultant. Prior to entering into the world of sales, Victor was a telecommunication Systems Engineer designing fiber optic and wireless networks and Product Manager of a software development group. 

James Glenn-Anderson has a PhD in Mathematical Systems Theory and MSEE in Semiconductor Physics. His career includes a twenty year stint in Silicon Valley serving in various systems and hardware architecture roles. He now owns an R&D consultancy specializing in supercomputing, Artificial Intelligence, and advanced image processing.

Read more in Chapter 4 of Sales Ex Machina - How Artificial Intelligence is Changing the World of Selling