The Chicken or the Egg

Big data skeptics are often subject to haughty derision by those of us “in the know”—the legions who have bought into the vision of a connected world and are seeking to make it real in their enterprises and industries. A recent article on the excellent website Fierce BigData bears the restrained title: “Big data cynics run amok, winners win anyway.” This level of hubris is not uncommon among those of us who have seen the light of the Industrial Internet.

But I run into a ton of skeptics during my travels through the defense world and I completely understand where they’re coming from. A lot of them were among the biggest advocates of leveraging data to increase efficiency and improve organizational performance when they first heard the sales pitch a decade or so ago. They pushed through programs to equip instrument platforms with sensors, collect data in vehicle computers, and transmit that data to massive centralized databases at headquarters. And ten years later, lots of former early adopters feel like they were sold a bill of goods. The data just sits there in expensive data warehouses gathering dust (so to speak).

So here’s the dirty little secret: Big Data is useless.

On its own, that is.

Chicken EggAs a result of investments made over the past decade or so, most military organizations today are awash in data from their platforms. But they’re still waiting for all that insight that was supposed to come with it. Which platform is going to have which issue and when? How should I prioritize the alarms that are popping up on my fleet management software? Does the data show me more than a potential problem? Can it point the way toward a solution? Data isn’t worth much without the ability to make sense of it all, but somewhere during the rush to gather all the data we could, we forgot to find analytics that could actually make our lives easier.

In fact, useless may be putting it mildly. If all you have is data then you’re actually probably worse off than you were before. You’ve added weight to your platforms and steps to your processes while subtracting hundreds of thousands—if not millions—of dollars annually from your budget for database maintenance. Until and unless there are industrial analytics easily available to the military that offer a clear and overwhelming ROI for the warfighter, the voices pushing for more data collection are always going to be at a disadvantage.

And so here’s our chicken-or-the-egg problem: How do you convince the Pentagon to pay (or keep paying) to gather, store and manage their platform data when they haven’t seen the value from analytics? And how do you prove the value of analytics in environments where data isn’t collected or managed well? No analytics that I’m aware of can perform well with bad data—but getting good data requires the right technology and the right processes, and those require investments of money, time and energy. And that requires some level of faith that a significant ROI will result from the investment. And—oh never mind…

The irony of it all is that there are analytical products out there in the market that could make a huge difference for the military by using that data to help reduce maintenance cost, reduce unplanned downtime, and increase readiness and operating efficiency. Products that work out of the box. Today. GE’s Proficy SmartSignal, for example, is deployed on over 30,000 assets globally in environments ranging from mining trucks below the earth’s surface to commercial aircraft flying at 35,000 feet.

But in order to realize the value from even mature, proven products like these, they need good data as a feedstock. Some military platforms, like many Navy ships, are well instrumented and supported by a data architecture and organizational processes that facilitate robust analytics. But Abrams tanks, by contrast, can’t even store sensor data onboard. Some ground vehicles can store data but it is only offloaded once every two weeks or so, and most folks I’ve talked to have concerns about the reliability of even the data that is uploaded.

In cases like the latter ones, a major change in technology and organization would be required for those asset users to benefit from the data revolution. How do you convince those users to make the tough changes without a concrete business case based on actual, not speculative, results from analytics?

In this regard, we’re all lucky that there are folks at OSD and within the service commands who still believe passionately in the transformational power of data analytics and are willing to invest to prove it out. What we need is to arm those true believers with the results of concrete field trials that show obvious value from analytics like SmartSignal, on platforms that are already well-instrumented. We’re having a bunch of conversations with the military today about how to do just that and we hope the results of those trials will help resolve the chicken-or-the-egg dilemma.

Only when we have the right data, the right processes and the right analytics, all at the same time and in the same place, will the vision of the Industrial Internet become real for military users on the ground. And only at that point will the skeptics melt away.

Todd Stiefler

Todd joined GE from the world of Washington politics, and in no time at all has moved on to his second assignment, which sees him managing business development for the services GE is increasingly looking to offer to customers, including the Proficy SmartSignal predictive analytics software.

More Posts