The Slow Pace of IIoT Adoption
The slow pace of Industrial Internet of Things (IIoT) adoption offers a lesson on technology getting ahead of its customers. Unlike pure software plays, IIoT requires a sophisticated network of sensors, along with meaningful data collection and algorithms that turn insights into actions. According to a Cisco study “60 percent of IoT initiatives stall at the Proof of Concept (PoC) stage and only 26 percent of companies have had an IoT initiative that they considered a complete success. Even worse: a third of all completed projects were not considered a success.”
Serious Insights has identified eleven factors that contribute to the slow pace of IIoT adoption.
1. IIoT security. Given the number of security breaches across the Internet, industrial firms, facilities managers, and individual technicians often doubt the ability of vendors to secure IIoT data and connections. While access to vibration data about a piece of equipment isn’t the goal of hackers, breaching operational networks and incapacitating equipment may be. IIoT security is of particular concern with cloud-based approaches and multi-plant implementations. Even the perception that the IIoT network brings new vulnerabilities to operations creates doubt and delays adoption.
2. Complexity. There is nothing simple about the IIoT. It involves installing sensors or replacing equipment or machines with newer versions that already include sensors (though practice may prove a need for additional complementary sensors to provide a more comprehensive operational picture), along with various layers of software that includes data collection and sensor management. Outputs range from basic reporting and analytics dashboards at the entry-level, to machine learning models with predictive analytics at the high-end. The skills in many organizations don’t align with those required to implement or to leverage IIoT. Complexity results in a combination of lengthy implementation schedules and a slow road to results that many firms decide to park or drive off of before they reach the destination.
3. Run to failure and preventative maintenance perceived as good enough. In many cases, the unknown cost of transformation and the lack of solid data about benefits and returns keeps firms doing what they have always done, running equipment to failure and replacing it, or running standard preventative maintenance programs. While moving to condition-based maintenance or predictive maintenance may still represent the ideal, organizations remain skeptical that the investments required to reach those aspirational levels of automation and performance monitoring will yield positive returns.
The idea of “good enough” also reflects back on processes and practices that might require updates, retirement or replacement. In some cases, “good enough” also translates into “required” when, for instance, paper logs remain as part of a regulatory framework regardless of their cost or of automation in place that makes them irrelevant. Change is hard, and it’s even harder when it includes multiple parts of an integrated system changing simultaneously.
4. Slow displacement of existing equipment. The industrial sector faces a number of uncertainties, from Brexit to the potential for trade wars. While anticipatory choices under uncertainty can lead to competitive advantage, in most cases, it leads to inaction. Maintaining older equipment longer may argue for investment that reduces the cost of that maintenance, but it may equally argue for staying with what’s known and keeping costs consistent by not introducing change. Organizations that lean toward the former are good candidates for IIoT proof of concept. Those seeking to maintain current practice will likely result in a high cost of sales to IIoT vendors, and if they do strong-arm them into a POC, it will probably be a half-hearted one.
5. Expense, logistics and reliability of retrofitting existing equipment. Placing inexpensive new thermal or vibration sensors on existing equipment may yield new insights that help prevent failure earlier than periodic maintenance or monitoring. How to scale those placements, however, and how to keep the emergent infrastructure up and running is a task most organizations aren’t ready for. Given the dearth of data on value and returns, many opt for pilots if they invest at all.
6. Not all industries are equally ready. Industries exist at different levels of maturity relative to each other. GE’s recent decision to focus on oil & gas, aviation and energy suggests that the strategic Venn diagram, at least for GE, includes the ability for their products to offer quality solutions in the industries they serve, and a readiness of those industries to adopt such solutions. Taking a scattershot approach across industries may result in more potential leads and test cases, but it will also result in more failures as vendors discover that they often lack the industry-specific knowledge required to deliver a viable solution. Failure may be a learning opportunity, but it must be balanced with maintaining positive brand equity.
7. Displacement of workers. While technology companies want to dispel the “myth” that technology will displace workers, the truth is, technology will displace workers. The jobs promised by vendors and the industry won’t likely be given to people in existing positions. The typical factory maintenance worker is not the customer of IIoT solutions. New skills are required to manage data collection, data security, machine-learning, predictive analytics and the ability to see the factory as an integrated whole rather than a collection of individual pieces of equipment. To some technicians, putting a sensor on a machine in order to eliminate the need for a maintenance route feels like putting a nail in one’s own coffin.
Ultimately, the technology sector promises to deliver systems that reduce the need for data interpretation to issuing an automated CMMS work order to a maintenance worker based on equipment condition, along with explicit instructions on how to remedy the symptom or solve a root cause. In this scenario, even though the maintenance technician may still exist, it becomes a job with increasingly fewer upwardly mobile opportunities. This suggests that maintenance technician outsourcing and gig economy models may grow. Technicians may either align with an outsourcing organization that focuses on the value of their core competencies, or technicians may opt for self-management and control of their own schedule and livelihood.
The uncertainty over the pace of change becomes an IIoT adoption factor, with those leaning to slow adoption seeing that as a way of maintaining the status quo, including keeping their jobs.
8. Emerging managed models. Many equipment manufacturers are now moving toward a model of total performance, meaning that they offer guarantees on equipment uptime. This eliminates the need for organizations to integrate the data for some pieces of equipment. These managed models often include lease options that reduce initial expenditures, provide a clear cash flow forecast and help ensure that as technology changes the organization will be able to upgrade to newer equipment. The PMMI Research 2017 Trends in Food Processing Operations reported that leased equipment grew at over twice the rate of food equipment purchased directly for manufacturing use.
The potential for manufacturer managed equipment in some sectors, along with a more rapid turnover of equipment can stall investments in retrofitting current equipment. In the long-term, however, these models may fuel IIoT as they facilitate the replacement of older equipment with more connected versions.
9. Integration with operations. The uncertainty in manufacturing in general may well lead to an extended span of control between operations and maintenance, with operations taking on factory or facility reliability as a key performance indicator. This simplifies the organizational structure and forces an integrated view of output and the means of production. That these two organizations often exist separately today, reinforces the stalemate on general investment in maintenance as a source of value and returns, which includes reluctance to move toward IIoT.
10. Consumer perceptions. Consumers with devices that don’t stay connected, who need to deal with incompatible protocols that require multiple hubs or the adherence to a single competing “standard”, and dealing with the remediation of states for lights and appliances following a power outage, along with hacks of consumer IoT, increase skepticism about the maturity of IIoT. Translating the consumer experience to the work setting, it would not be unfair to say: If IoT vendors can’t get my house to work consistently, then how well will similar technology work in complicated industrial settings that will rely on IoT for reliability and safety?
11. The software isn’t good enough yet. From the connectivity of devices to the integration with operations to the lack of good data driving machine learning algorithms, the entire IIoT market remains in start-up mode, delivering subpar software. The quality programs that pervade modern manufacturing need to be adopted by IIoT vendors as a key principle. Until software can reliably and consistently gather data, provide meaningful insights and transform those insights into meaningful actions that reduce overall costs by increasing reliability, IIoT will remain as a vision stuck in prototypes and proofs of concept.
The eleven factors also suggest an underlying issue of trust. Organizations must trust their equipment. Any change that undermines that trust creates risk. They must trust their vendors. Any prototypes or proofs-of-concept, or worse, shipping product, that fails to meet expectations, erodes trust in the vendor. Any vision that conflicts with actual results, such as the displacement of workers against a vision of higher-value employment, eats away at trust. When massive amounts of sensor data get pulled from storage, analyzed by a machine learning algorithm and a work order gets assigned to an engineer who will shut down a production line for a shorter period than a catastrophic failure, the only relevant factor between acting and not acting is trust.
Recommendations to IIoT Vendors
Don’t overpromise. Do over deliver. If you can’t deliver on the vision, separate the vision from the offer and make sure that the offer delivers on its own promise in a reliable way that meets or ideally exceeds customer expectations. Many IIoT offers appear as prototypical as the implementation situations. Vendors still working on a solution should be honest that the POC is really a prototype in order to set customer expectations about the reliability of the software components.
Provide value quickly. In an environment where technology has failed to prove its value, vendors need to concentrate on clear, repeatable offers that provide quick wins for the adopting organization. Vendors should not lead with large, multi-layered, multi-year projects, but with small, localized and effective offers that bring value to all internal stakeholders. The vision may still include a larger solution, but if the initial phases fail to produce results, projects may get canceled and depending on the deal, the vendor may be held liable for some of the costs associated with the failure.
Focus on layer mastery. As GE examined the issues around its Predix products, it concluded that it bit off more than one company could chew. IIoT vendors need to decide what core strategic value they offer the market and deliver that with high quality. They need to openly partner with others who can complement their core value.
Design for an ecosystem. The IIoT industry needs to get together and figure out how to create protocols and interfaces that allow multiple vendors to coexist and integrate. Vendors need to align their offers to match their capabilities and strategic intent. No vendor should feel the need to create a strategic obligation to build out an entire solution simply because a solution doesn’t yet exist that serves its core market.
The end game for IIoT is likely a systems integration play with multiple technology vendors being brought together by a third party to offer an industry-specific solution that helps a company overcome its unique operational and reliability challenges.
The sooner vendors adopt an ecosystem solution rather than see the future as a number of competing full stack solutions, the faster they will be able to apply their investments to develop complementary niches where they can offer superior value. Vendors should also able to provide value independently of the ecosystem within their niche, eliminating the need for the ecosystem to develop before finding a path to profitability.
Focus on industries. Most equipment and software remains industry specific. IIoT, which offers some horizontal opportunities for sensors and data collection, will be an easier sell with industry-specific solutions and partnership.
Help customers with their vision. Vendors need to ensure that the customer’s vision of the outcome matches what they can deliver. They also need to be courageous enough to challenge the customer if the customer’s vision doesn’t fit their actual needs. This means successful IIoT vendors will require some level of professional services in order to support implementations.
Partner with a systems integrator. As noted above, a systems integration ecosystem will likely evolve, absolving hardware and software vendors from customer vision concerns as they concentrate on delivering their core value to a project. Systems integrators, however, need to make sure they manage the customer’s business objectives, internal alignment issues, cost expectations and implementation pace to ensure successful projects.
Own security. It doesn’t matter if a security breach comes from a flaw in the IIoT software or something that happened on the customer’s infrastructure. The failure will reflect back on the IIoT implementation. Vendors need to own security to the point that they are willing to halt implementation if they discover an issue that will put their customer at risk.