Why IoT-enabled predictive maintenance hasn’t taken off

“Two years ago, predictive maintenance was forecast to be one of the most promising uses of the industrial Internet of Things (IoT).”

That’s the lead of report based on a recent Bain & Company survey of more than 600 high-tech executives (Beyond Proofs of Concept: Scaling the Industrial IoT, by Bain partners Michael Schallehn, Christopher Schorling, Peter Bowen and Oliver Straehle). The report goes on to note that identifying precisely when equipment might fail “seemed like a no-brainer.” And yet, the report concludes, “predictive maintenance has failed to take off as broadly as expected.” In fact, industrial leaders were not as excited about predictive maintenance as they were back in a 2016 survey.

Predictive maintenance hard to implement, hard to derive value

According to Bain, there have been problems on the both sides of the ball: First, implementing predictive maintenance has been harder than expected, and second, deriving valuable insights from the data gathered has also turned out to be unexpectedly challenging.

But wait, it gets worse.

It seems that “predictive maintenance is just one of many IoT use cases that customers have had difficulty integrating into their existing operational technology and IT systems.” While investment in proof-of-concept projects continues, the Bain report said, actually turning that into successful mainstream implementations hasn’t been able to keep up. Long-term enthusiasm for the technology remains strong, the survey showed, but many industrial organizations now foresee implementation taking longer than initially predicted. (Bain still predicts the industrial IoT market to double by 2021, topping $200 billion.)

An evolution in industrial IoT concerns

Those delays come along with a shift in the concerns of industrial IoT users. In 2016, top concerns centered on security, return on investment, and integrating IoT solutions with existing IT systems. After a couple years of PoC projects, security remains a key concern in 2018, and worries around gathering sufficient technical expertise, dealing with multiple data formats, and transition risks have become even more intense. ROI concerns, meanwhile, have faded a bit. (Could it be that industrial IoT projects are actually paying off?)

None of this should come as a huge surprise. The transition from promising vision to practical success often faces unforeseen roadblocks. But while these impediments can loom large during the early adoption phase, they tend to be quickly forgotten once the technology achieves widespread success, Indeed, the Bain report notes that “given the progress in sensor technology, 5G connectivity, edge computing and edge analytics, and an estimated 20 billion devices connected by 2020, there’s little doubt of the potential for technology to vastly improve efficiency and no doubt that the IoT will have to manage it.”

Suggestions for IoT vendors

To overcome these issues, Bain says, analytics firms, industrial technology makers and cloud service providers must help their customers gain deeper experience in industry-specific applications and offer more complete, end-to-end IoT solutions.

In the short run, though, that may not be so easy, for a variety of reasons.

First, industrial IoT vendors and users don’t always agree on what’s important and what’s ready for prime time, the Bain survey revealed. For example, while quality control and remote monitoring and tracking of equipment ranked high among both groups, customers wanted augmented reality/virtual reality (AR/VR) and energy management solutions, which vendors were less ready to deliver. On the other hand, vendors said they were ready to deploy predictive maintenance solutions, but it’s not clear customers are waiting in line to buy them. (Notably, smart-city use cases, which have received a lot of press attention, were not top of the list for either vendors or customers.)

Software deficiencies also remain a problem. According to Bain, “Device makers and other vendors of industrial and operational technology need to dramatically improve their software capabilities—not a historical strength for most of them.” The report notes that many vendors are spending freely on acquisitions to accumulate the required capabilities.

Finally, Bain suggests that industrial IoT vendors focus on enabling key use cases and overcoming critical barriers, plus leverage partnerships to fill the inevitable capability gaps.