Predictive analytics turn uncertainty into usable probability

Following on from yesterday’s post on analytics, let’s talk about predictive analytics. Another phrase I picked up while working at FICO was this one:

Predictive analytics turn uncertainty about the future into usable probability

Again, I don’t know if the phrase originated there or was just in common usage but it always struck me as a particularly powerful phrase. If you think about the future, it is inherently uncertain. You don’t know which customers will fail to pay their bills, will crash their cars or will desert to a rival supplier. If you did know you could do something about it – make them pay in advance rather than offering credit, charge them a higher insurance premium or make a proactive retention offer. And your business would be the better for it. Lacking a crystal ball, however, most companies seem to resign themselves to not being able to do anything about these kinds of things and become entirely reactive – putting customers into collections, raising premiums for those that crash and chasing after customers who leave.

But …

Copyright © 2009 James Taylor. Visit the original article at Predictive analytics turn uncertainty into usable probability.

Syndicated from ebizQ

Following on from yesterday’s post on analytics, let’s talk about predictive analytics. Another phrase I picked up while working at FICO was this one:

Predictive analytics turn uncertainty about the future into usable probability

Again, I don’t know if the phrase originated there or was just in common usage but it always struck me as a particularly powerful phrase. If you think about the future, it is inherently uncertain. You don’t know which customers will fail to pay their bills, will crash their cars or will desert to a rival supplier. If you did know you could do something about it – make them pay in advance rather than offering credit, charge them a higher insurance premium or make a proactive retention offer. And your business would be the better for it. Lacking a crystal ball, however, most companies seem to resign themselves to not being able to do anything about these kinds of things and become entirely reactive – putting customers into collections, raising premiums for those that crash and chasing after customers who leave.

But companies that understand the power of predictive analytics can become proactive. Predictive analytics are not perfect at predicting the future but they allow you to see how likely something is. Take our three examples:

• Historical data about which customers paid on time and on the behavior of customers before they stopped paying on time can be used to predict the likelihood that a particular customer is at risk of not paying on time.
• Analysis of the behavior of cars and drivers involved in accidents can find the characteristics of cars and drivers that increase risk and a model developed to predict the likelihood of a particular car/driver combination having an accident in the next 12 months.
• The changes in behavior of customers who went to a competitor while they were still customers can be analyzed to identify those behaviors that should be considered triggers for proactive retention – the behaviors that make it more likely that this customer is considering leaving

In each case we are moving from simply being uncertain to having some idea of the probability of something being true. This allows us to act, and act proactively. We will likely try and retain some customers who weren’t going to leave or restrict the credit of someone who would have turned out to be a perfectly good risk but overall our business will run better and more profitably. Our decisions will be able to see through the uncertainty inherent in business and will be better for it.

Link to original post