Having been in manufacturing for 20+ years, I have to admit I’m a little chagrined at the term “big data” – the reason being that manufactures have long been generating and capturing a huge amount of data in the forms of product genealogy, traceability, equipment events, regulatory conformance, etc. But, I get it – data is growing at a far more rapid rate with a set of expectations that the intelligence “locked” into the data should be readily available. More “intelligent” equipment is spewing more data, and there’s greater participation across the supply network in fulfillment processes. The real question is not how “big” your data is (and by the way, I’ve never heard a manufacturer use the term), but how your data enables faster, more intelligent decision support within an ever shrinking decision window.
I heard a statement a couple weeks ago while visiting some pretty sharp folks up in Boston that pointed out to me the incredible value of a small (but important) piece of data when it is available before an event versus 100 times the volume of data made available six months after the fact.
Today, executives and managers want help with agility, responsiveness and risk mitigation. Line of Business managers, especially those new managers that might have been recently promoted, expect and demand data “right now” so as to make smarter business decisions faster, within an ever shrinking decision window. As part of these expectations, combined with an inherent analytical mindset, they expect ease of use and the ability to quickly discover relationships and operational insight from the data themselves. Visibility and speed are the order of the day. These folks aren’t going to wait for corporate IT to give them what they need.
Traditional Business Intelligence Solutions Don’t Work
Intelligence or visibility has the most value when the data is reliable, timely and role-based – and is delivered with the right business context. General business intelligence tools are fine for general business problems. But, there’s a substantial effort to convert these solutions to expose operational intelligence in the right context. As a result, I have found that line of business folks are generally not well served by these applications.
Sure, there are vendors touting Enterprise Manufacturing Intelligence (EMI) solutions, but many users find that layering yet another application on top of a disparate set of operational point solutions just compounds the issue. Further, the applications that EMI tries to cull data from were never designed to support analytics or intelligence, so they often don’t have the required data anyway. What’s needed is something different. More and more, I’m coming to believe that using a separate operational intelligence platform from the operations execution platform just doesn’t work. Forget for a second about the IT implications; the real problem is that if you don’t adjust your manufacturing platform to provide the data that an operational intelligence solution needs, you’ll always come up short.
The Need to Combine Operations Intelligence with Execution
Fact is there’s data needed for analysis that just isn’t needed for execution. Take for example Slow Changing Dimensions, such as when an employee changes departments, or when existing equipment is reconfigured into a new product line. Execution just needs to be updated to the new identities and dependencies. The operational intelligence solution, however, needs to know the relationship between the old information and the new. There are many other examples of data that execution doesn’t care about but intelligence fails without.
If you continue to think of operational intelligence being separate or a “layer” you can add on to a manufacturing execution system, your chances increase that you will end up with “big data” but not the intelligent decision support you need to actually understand and use this valuable resource. Those that take a combined approach will be able to outmaneuver and outperform you, putting you in a disadvantageous position with your competition.