How to revamp business processes with predictive analytics
Can you build better business processes than what you have now? If you have a data science team at your disposal, chances are you can.
Leaders are always looking for ways to do things better-faster, cheaper, higher quality. Many companies I have worked with have gone so far as building a continuous improvement organization into their structure. These are employees whose entire job is looking for ways to improve the company's business processes. It's a good idea.
However, sometimes you need something more than incremental improvement on subprocesses. Sometimes you need to step all the way back to your highest-level global processes and rethink the best ways to accomplish your mission and strategy.
SEE: Big data policy (Tech Pro Research)
The idea of Business Process Reengineering (BPR) hit the scene in the 1990s with this fundamental philosophy: Start over and challenge everything about your business processes. Like many other management fads, it waned as quickly as it surged, giving way to the next great management idea. Well, it might be time to revisit this idea, because they didn't have data science in the 1990s like we have it now.
How predictive analytics can add value to BPR efforts
Data science and big data analytics are game changers when it comes to BPR. There are several areas where data science can add value to your BPR efforts, though the most compelling is predictive analytics. Predictive analytics paves the way for blowing up existing assumptions about your business processes.
Let me explain: Many times, a business process will wait for information that it doesn't have. Think of a typical order processing workflow. I was talking to a publisher not long ago about improving their order management system. Their process starts when a customer either phones in an order, faxes in an order (can you believe people still use fax machines?), or orders a publication over the web.
This is a reasonable place for the order processing workflow to start, but what would happen if they could predict what customers were going to order based on their past purchasing patterns? They could skip the waiting and just send the right publications to their customers. On the rare occasion when a customer did not want the publication that was sent, the publisher would eat the cost and/or make it easy for the customer to return the publication.
SEE: 60 ways to get the most value from your big data initiatives (free PDF) (TechRepublic)
The first step is taking a critical look at your processes and asking yourself where a reliable prediction engine could alter your entire workflow. It could be at trigger points, like the example above, or it could be in other queues where the process is waiting for information that can be predicted. It may not be obvious on your current process flowcharts, so you will need your business experts to help identify these areas. But before you engage your business experts, think about it from their point of view.
How BPR may affect business experts
A common challenge for all leaders embarking on any BPR effort is engaging the business experts who will eventually be displaced by the new process. More often than not, BPR efforts are fueled by an expectation that some new technology will, at least in part, replace the humans that are currently doing the work; this presents an obvious problem for the humans. So if your plan is to leverage their knowledge, it's best to have a change management plan in place before you raise the prospect of partnering on your new BPR endeavor.
Solving this problem is not as hard as you may initially think, unless your primary objective is to reduce headcount to save costs, in which case I cannot help you. If you would like to revamp your processes and still retain the business experts that house your institutional knowledge, then consider how the predictive analytics will be used. The people who are currently in the prediction business at your company are the best ones to know what to do after the prediction is made.
SEE: Quick glossary: Big data (Tech Pro Research)
In the example above, I would not recommend releasing the customer service representatives because you can predict what and when customers will order-they are the ones who know the customers the best, so letting them go would be a huge mistake. Now that their time processing orders is freed up, they can spend more time servicing customers in other ways and building relationships.
If you are looking to blow up your existing processes and start over with some fresh thinking, you would be remiss if you overlooked your data science department. Data science-specifically predictive analytics-can be a powerful weapon to defeat incumbent assumptions and discover breakthrough efficiencies in the way you accomplish your corporate mission and strategy.
Starting at your highest global process level, systematically go through your workflows asking yourself: "Can I make this better by predicting what's going to happen?" Partner with your business experts to identify the biggest opportunities, but make sure they have a home when everything is said and done. BPR can be exciting, rewarding, and extremely effective when predictive analytics is in the mix.