One of the biggest fears about Internet of Things (IoT) initiatives is that they’ll fail. After all, it’s not hard to find examples of over-hyped projects or underwhelming results. Many IoT technologies are experimental, new, or simply untested.

So before investing in IoT solutions, it’s worth asking some hard questions. How often do IoT projects fail? How often do they succeed? What does success or failure look like, in either case? And how can we make sure future projects deliver more of what they intend?

The answers in this post draw on 35+ research studies, interviews, best practice guides, and similar resources to see what’s happening. Note that many of these focus on the industrial IoT (IIoT), while some look at more general IoT projects. Given the overlap reported across the board, this post will often lump them together.

Here are the main topics this post addresses:
1. The adoption of IIoT
2. IIoT failures
3. Warnings about failure and best practices for success
4. Figures on success

1. The adoption of IIoT
More and more companies are testing the IoT waters for themselves.
Companies like TagoIO for instance, have undertaken projects ranging from maintenance improvement to agricultural condition monitoring to fleet management. New IoT organizations seem to appear every month—and the ecosystem will probably need these new companies to keep up with the explosive market potential.

In 2019, Sierra Wireless forecast that as many as 50% of factory assets would be “connected” by 2020. Another figure says we can expect to see as much as a 40% annual growth rate in IIoT by 2024.

Many estimates like these are based primarily on industry buzz, of which there’s no shortage. For instance, one survey of manufacturing executives reported that 92% saw smart manufacturing as “the way forward” for their industry. This doesn’t, however, illustrate how many of them were actually ready to invest in these solutions.

A study by PWC found that not only did 93% of manufacturers believe that the potential for IoT success exceeded the risks, but also that 71% of industrial manufacturers were “already building or testing IoT-related solutions in both active and in-development projects.” Furthermore, 68% of PWC’s surveyed manufacturers reported plans to increase their IoT spending soon.

Other studies align remarkably close to these numbers. For example, 94% of participants in a large Microsoft study expressed the intention of using IoT solutions before the close of 2021. Approximately 85% were already pursuing IoT projects, and 88% called IoT “critical” for their business’s plans and interests, “with a majority of those who’ve adopted IoT predicting they will see a 30% return on investment two years from now (inclusive of cost savings and efficiencies).”

Whether for the promise of cost savings or worries over old processes failing, “the vast majority (84%) of project-oriented organizations have already adopted formal IoT project implementation practices” or plan to by 2024, according to Gartner.

These figures are even more promising when you consider that IIoT operates in an industry that’s particularly scared of new risks, where equipment lasts for decades and purchasing decisions seem to take just as long. Perhaps IIoT solutions are finally maturing.

2. IIoT / IoT failures
The big questions, though, are not about how often people adopt IIoT technologies, but how often they succeed or fail—and why. A few studies focused on IIoT industries get cited especially often here:
● 74% of IoT projects fail, Cisco reported in 2017
● 30% of projects fail in the proof-of-concept stage, according to a 2019 report by Microsoft
● Only about 17% of companies see big payoff from IoT projects, according to McKinsey
● 58% of surveyed IoT adopters found their projects to be unsuccessful (Beecham Research)

One of the challenges in interpreting these figures is that there’s little consensus on what ‘failure’ or ‘success’ means. For instance, Cisco’s survey reported that 35% of IT executives said their projects were totally successful. On the same projects, however, only 15% of business executives reported success. It’s also usually unclear if there’s a difference between “total success” or large payoff, as McKinsey addresses, versus more modest returns. IoT leader PTC indicates that 89% of IoT-adopters “expected to move from pilot to full production within a year of purchase.” Whether pilot purgatory means the project failed or not, however, is impossible to say from outside the project.

Although it still doesn’t define either failure or success completely, Beecham’s report gathers a particularly solid list of reasons behind IoT failures. Interestingly, nearly every failure they list centers on planning and project management. Only a few are technical in nature.

For instance, out of 18 different reasons for project failure, Beecham’s only three citing technical issues are: 1) IoT projects using the wrong technologies in the first place; 2) lack of involvement, support, or management from IT; and 3) lack of technical understanding. No one in this survey reported cybersecurity, outright technology failure, or lack of interoperability, all which are common concerns in the IoT conversation at large.

The top reasons given are poor management. This includes unclear or unstable project requirements, lack of buy-in from executives, lack of careful planning or project management, not doing enough, or, the single most common point of failure, “failure to change the operational model.” Some of Beecham’s more qualitative inquiry into IoT failures brings up the same issues, citing unrealistic expectations, unclear or unfeasible timelines, and a lack of scalability planning as top reasons a project didn’t succeed.

One other matter to consider: The recency of the data. Cisco’s report in 2017 is one of the internet’s most-cited studies on IoT failure and success, but this study was largely assessing technologies from 2016, 2015, or even earlier. This doesn’t mean it’s not valuable today. It just serves as a reminder that newer technologies need newer research.

In any case, it’s clear that the IoT graveyard isn’t empty. It’s also clear that project success is difficult to define, let alone to achieve.

3. Warnings on failure and best practices for success

There’s at least one silver lining in so many IoT failures: People are learning from them. In fact, Cisco’s 2017 report says that more than 60% of survey participants sped up their IoT investments “thanks to learning from stalled and failed IoT initiatives.”

Most studies of failed projects give some advice on how to avoid the same mistakes. Beecham, for instance, identified several specific challenges for future IoT initiatives:
● Budgeting
● Setting project parameters and success metrics
● Planning for scalability
● Testing technologies prior to purchasing
● Lack of expertise in solution architecture and design

A survey by Cap Gemini, referenced by Beecham, outlines three additional concerns: Limited abilities to analyze or make sense of any data produced, cyber security, and lack of clear business value for IoT solutions. Microsoft similarly warned that advancing from a proof-of-concept or pilot often failed without a clear understanding of the project’s monetary value.

All these warnings about failure—based on so many actual examples—have led to a huge number of best-practices, with valuable resources from Tech Republic, Industry Week, Computer Weekly, Digitally, the Association of Equipment Manufacturers, Manufacturing Business Technology, and Very, as well as Beecham’s report and other resources mentioned above. Across best-practices resources, the most common themes are:
● Plan your collaboration: This could be both internal and external, involving IT, end-users, buy-in from executives, and an ecosystem of partners like TagoIO or WISER Systems. One especially good idea here is to involve those whose jobs may be affected by new technologies. In the end, this is about connecting people—not just things.
● Focus on the actual business value and solving real business problems: Perhaps because of the huge interest in IoT, many companies seem to feel the pressure just to implement new technologies. It’s little wonder that they sometimes find these projects unsuccessful, not having real metrics of success in the first place.
● Actually use the technology to influence and change operations: Beecham reported this as being an issue nearly twice as often as any other single thing. This is likely a challenge since most organizations govern their day-to-day operations across many different leaders and organizational branches, whereas IoT projects often get assigned to just a few individuals.
● Don’t neglect security: It’s worth noting that few companies actually reported having security issues—they simply worried and warned about them. This is likely an especially common thread since the risks of failure are particularly high where security is concerned.

TagoIO and Wiser System integration for visualization and analytics in realtime

Other concerns mention the lack of professionals with a background in IoT and the need to address data infrastructure. As with the case of resources on IoT / IIoT failure, these best practice guides focus primarily on organization and implementation and not on the state of IoT technologies themselves.

4. Figures on Success
Clearly, plenty of projects have not delivered what people expected, whether because of tech failure, overhyped solutions, or poor organization. But there’s also a growing corpus of successful projects.

Earlier in 2020, Smart Industry shared the prediction that the IoT would leave “the Trough of Despair” this year. IoT Analyst Michael Kanellos says, “IoT projects are often portrayed as the quickest way to get yourself fired. Have you heard the ‘fact’ that three fourths of them fail? In reality, success is the norm.” Kanellos cites 451 Research’s Ian Hughes with the figure that only about 7% have actually delivered a negative ROI. He also references Gartner’s Eric Goodness saying that 97% of projects do indeed meet their outcomes, with 57% “achieving outcomes better than anticipated.” In both cases, the counts of project failure number in the “single digits,” Kanellos concludes.

Last year PWC indicated that 81% of industrial manufacturers were already seeing efficiency improvements thanks to IoT. Breaking this figure down more specifically, 41% had already validated supply chain gains thanks to IoT, 39% had seen security improvements, and 43% saw improvements in asset management. In every case, earlier-stage companies expected to see similar gains within 2 years, with 68% of manufacturers planning to increase their IoT spending in the near future.

A 2018 report by McKinsey indicated that 58% of respondents saw a 5% revenue increase—or more. Also, 46% reported a 5% or larger reduction in operating costs. While this doesn’t show whether the companies in question defined these as successful projects or not, most companies will jump at the prospect of 5% revenue gains or cost reductions.

Even as early as 2017, a study by BDO indicated that 72% of surveyed manufacturers “increased their productivity” and “69% of manufacturers increased their profitability” within one year, thanks to IIoT applications. That same year, the Bsquare annual IIoT maturity survey reported that 84% of manufacturers found IIoT “extremely effective.”

Whatever their success criteria are, these figures make it clear that IoT projects are succeeding regularly.

Here are a few specific examples of successful projects:
A Deloitte IIoT project added 10% throughput to a manufacturing line, saving approximately $26M; a global energy company reduced their unit production costs by 33%, totaling over $9B in savings across five years.
● Siemens shortened a routine machinery programming process from two hours to just one minute (Internet of Business).
● A plastic manufacturer boosted yield rates from just 36% to 99%—nearly a 200% increase—by implementing a real-time monitoring solution (Forbes).
● One manufacturer saved $3M per year on each production line by implementing an asset location and tracking system (Wired).
● An aerospace facility saved dozens of technician hours every month by tracking the real-time locations of their tools (WISER).
● Other asset management and tracking projects saved an average of $220K per projects—and generated nearly $9.5 in new revenues for each project (IoT Agenda).

In sum: What is success?
The examples above include plenty of cases where ‘success’ is a moving target. If you’re working on an IoT project, you’ll need to determine what success actually looks like in practice. It could be streamlining an existing process. It could be enabling a new one. It could be coaxing a few more years of life from old machinery via predictive maintenance.

The key is that you should decide what constitutes success based on your own use case, industry, and business needs. If you make this determination before your project is underway, you’ve got a much better platform to achieve that success in the first place.

By: Stephen Taylor, Director of Communications at WISER Systems, Inc., a leading ultra-wideband (UWB) provider of precise localization. When he's not at work, he likes to jam on his violin, write creatively, or wander through the forest. Someday he'll try all three things at once.