Internet of Things

Where, really, is the ROI in IoT?

Many industry prognosticators (and not a few vendors) are pushing Internet of Things (IoT) technology but are somewhat vague as to how financial results for actual businesses will be materially improved. Instead, they tend to focus on concepts like “digital transformation,” which sound promising but is difficult to quantify. This places many businesses in a quandary—they are aware of the fact that IoT holds great promise but they can’t really move forward without being able to identify a tangible ROI. And make no mistake, unlike consumer IoT, and even government-sponsored IoT initiatives (think smart cities), without tangible and quantifiable improvements in financial outcomes (i.e. an ROI), businesses will be hard pressed to move forward with the vigor that industry prognosticators think they should.

Unfortunately, this lack of focus on tangible business uses cases has the potential to stall growth in IoT and deprive businesses of many of the benefits they expect to realize from it (and growth in M2M should not be interpreted as growth in IoT; M2M is simply connectivity while IoT typically involves multifaceted systems incorporating machine learning, data analytics, complex rule generation, and automated orchestration of actions).
Fortunately, there is an approach to IoT deployments that can, depending on the nature of the business, demonstrate fast and meaningful payback. This approach focuses almost exclusively on actual business-oriented IoT use cases but it differs from the “platform-first” strategies being pushed by many vendors and analysts.

The platform-first approach

Proponents of the platform-first school argue that organizations of all types should first make a corporate-wide decision on an IoT platform. Although the definition of IoT platform is somewhat broad, in general it is the system upon which applications can be built to take advantage of data being collected from myriad dissimilar devices. After standardizing on an IoT platform, organizations then need to purchase or develop applications, potentially integrate other enterprise systems and ingest device data. Only after they’ve done all of that can they begin to focus on IoT use cases that actually benefit the business and presumably, generate an ROI.

The challenges associated with this approach should be obvious. In most cases businesses are being asked to make significant financial investments with no clear view as to what, if any, payback will result. They are also assuming substantially more risk than would otherwise be required since they are being forced to make enterprise-wide decisions on what is still, in many cases, evolving technology. Finally, platforms are by definition incomplete systems; they need applications to be developed on them before they can be deployed in production. This aspect carries the downside that time-to-benefit is lengthened.

The use-case-first approach

The reality is that, as has been the case with most major technologies adopted by businesses over the last quarter century, initial production deployments of IoT technology will almost always be related to individual business initiatives. These initiatives are typically not viewed, nor should they be viewed, as IoT initiatives. Instead, they are focused on driving specific business outcomes—for example, improving asset uptime, reducing service and warranty costs, improving food safety, complying with government regulations, adding new revenue generating services, etc. It is only as organizations evaluate technologies that can help them meet these business objectives that they frequently discover data generated by various distributed devices can be harnessed, analyzed, and used to automatically drive business processes. In other words, these ROI-producing business initiatives begin to take on the aspects of IoT initiatives.

The important point here is that successful—and successful means that they (a) work, and (b) provide a financial return—IoT initiatives are those for which the main goal is a quantifiable business outcome and in which IoT only plays a supporting role, albeit a critical one.

How did we get here?

It is possible that the chief reason we see a bifurcation in approaches to industrial IoT is that different businesses may choose different departments to spearhead these efforts. For organizations that look to IT to lead the charge, the platform-first approach will be more attractive as it potentially represents a uniform architecture that can be deployed across the enterprise. However, it also represents an “IoT for IoT’s sake” approach that may fail to deliver a payback and leave the business somewhat frustrated with the results.

Use-case-first initiatives, on the other hand will almost always be sponsored by OT (operations technology; really any revenue generating line of business). As such, there is by definition an ROI-producing business objective driving things and IoT, while critical, is incidental to the initiative.

Is there a downside?

Some will argue that the use-case-first approach could result in dissimilar systems being deployed in different parts of the organization, long the bane of IT’s existence. While this could be the case, the potential downside is mitigated by the fact that—unlike the early days of computing (Mac versus Windows) and local area networking (Ethernet versus Token Ring)—IoT systems use well-established protocols and standards. This allows dissimilar systems not only to coexist but even to exchange information and leverage multiple data sources.

While this battle is likely to continue for some time it is also likely that OT will prevail, at least in the near term. The reason is simply that ROI drives everything for businesses (again, consumers and governments tend not to care whether their investments are well thought out) and IoT platforms, by themselves, are hard-pressed to generate meaningful financial returns.

Like this article? Subscribe to our weekly newsletter to never miss out!

Previous post

"Fashion needs ‘datafication’ that clearly benefits fashion." - Interview with Data Scientist Jessica Graves

Next post

Top Virtual Reality Blog for VR Lovers and Developers