Do the Right Thing...The Use of Predictive Analytics in Business Processes
Many business processes are about doing things the right way. There is not always the same emphasis on doing the right things. Referring to a popular theme in this day and age, there may be a near-optimal process in place to fulfill a mortgage application, but should the mortgage have been approved in the first place? What’s the probability of default, and what is the expected loss to the company? Similarly, all the sales fulfillment processes may be running full throttle, but are the right products being offered? Those that, in the end, maximize the lifetime value for that customer? Should a different product have been proposed, at a different price, or with a different incentive?
This is where predictive analytics comes into play. Businesses have many hidden treasures in their data. The data can be held in operational databases, data warehouses or even census or publicly available data. There is value in the individual data sources, but even more so in the combination. Customer purchase patterns, satisfaction drivers, and future behavior are all “hidden” in this data. The whole purpose of, and motivation for, predictive analytics is to discover these patterns, use them to predict future behavior, and then act on the insight.
Per the mortgage example above, without predictive insight many decisions will be bad decisions. It’s surprising how many bad decisions are made based on hope, gut feel, mere assumptions, or a naïve interpretation of historic trends. Where customers, who are notably fickle, are concerned, it’s almost always impossible to play by ear. Advances in statistical analysis and machine learning have made it possible in many cases to predict customer behavior with a high level of accuracy. Moreover, it is possible to calculate the confidence one can have in those predictions. There’s no more guess work in trying to figure out the right thing to do. And for every possible action, you’ll know the margin of error in advance.
This article will demonstrate the value that predictive analytics can bring. Specifically, it will cover where it can be used, where it should be used, but also where it’s less appropriate. Under what conditions the technology can contribute to the business and when it will be a waste of time. And most importantly, how the discoveries made in the data can be operationalized and automated through BPM suites1 to help those solutions do the right things right.