Predictive analytics has found its bridgehead in Industrial Internet of Things applications. It's asset performance management.
Some think that predictive analytics is somehow superior to traditional business intelligence and will supersede it. I disagree - although there is some overlap, I think they are largely complementary technologies. Increasingly, the Discover style of analytics can work with large data sets to uncover cause-and-effect - but only if the causal relationships are relatively simple. On the other hand, predictive analytics excels at finding complex relationships in large data sets. It's only natural then that predictive analytics will have a big role to play in making sense of data generated by the Industrial Internet of Things (IIoT). There will be no shortage of IIoT data, and inferring meaning from it in a timely way isn't going to be easy because it will come from so many diverse sources.
A number of case studies, partnerships and product announcements attests to the early energy being focused on maintenance applications. For instance:
- We learned from the GE Minds and Machines events last October that insights from a cloud-based, predictive analytics solution had been used to replace a seal on a pump before it failed. The cost saving from avoiding unscheduled downtime were estimated at $7.5m.
- Caterpillar and Uptake just announced a partnership to expand the use of predictive analytics to optimize machine availability - notably for both Cat® and non-Cat branded products.
- Last week, at a Software AG event, I learned about a predictive maintenance application that Software AG has developed with a partner for GE Jenbacher engines.
And on April 15th, I'll be presenting on asset performance management for IBM at a Predictive Maintenance and Quality Innovation Event in Columbus, Ohio. Exciting times - I know Ralph Rio, Paula Hollywood and myself are looking forward to learning and writing more about these solutions as they mature and start to deliver a more enduring ROI.