When you’re ready to make the most of your data through an IoT Platform, you have options on how to get started. This article will help you determine if you need a Proof of Concept (PoC), a pilot or a full production instance of the IoT Platform.

Since forming the Watson IoT business unit in 2015, IBM has worked with hundreds of customers on their IoT projects. Through those engagements we’ve learned a lot about how a typical Internet of Things (IoT) solution is conceived, built and deployed to generate new business value.

The 2018 Gartner CIO Survey reveals only 12 percent of enterprises have invested in IoT, and 24 percent are in short-term planning. (See  “The 2018 CIO Agenda: Mastering the New Job of the CIO”.) Thus, few enterprises have large-scale IoT initiatives. Business alignment remains the primary challenge — many enterprises are clarifying how to incorporate IoT into their digital transformation.[1]

We’ve seen many organizations struggle with the same challenges. In turn, we have developed capabilities and methodologies to help our clients overcome those hurdles and achieve success. This article aims to explore those common challenges, and share what we’ve learned.

Introducing old themes

Let’s start with the “bigger picture” for the Internet of Things. Since I started my career in IT over 25 years ago, one of the greatest opportunities for increasing business efficiency and productivity has been through integration.

Business used to run on memos typed and sent around between individuals as pieces of paper. Nowadays we have email and a message can reach someone on the other side of the world in the blink of an eye.

In the same way, information flow between IT systems used to be a challenge. Networks connect the physical hardware where these systems run. But there was always a challenge to make the systems understand each other’s data formats and cope with different processing speeds.

We developed software solutions that solved those problems (message queuing and brokering). These then became standard practice across the industry. With IoT, the integration problem is still there. But now it’s integration between tiny computing devices, embedded into sensors and devices, and IT systems that need to access and understand the data.

Integrating other things

Not surprisingly, the same technologies that solved the problems for application-to-application integration have been put to use in the IoT space. What’s moved on since they were first developed are the standards in the industry – now much more mature and widely accepted – these open up a world of compatible devices, gateways and IoT platforms. All of which can be put to use more rapidly and with less risk.

What technology has not solved is the business side of the solution. Investing in technology can improve productivity, efficiency, customer experience, and more. But does the return justify the investment? With IoT, we see a lot of clients struggling to evaluate this equation.

While the costs of connected sensors and IoT-enabled devices has tumbled, and the choice has proliferated in recent years, there are other unknowns that make it hard to assess:

• The ongoing costs of running the solution, storing IoT data, maintaining devices
• The true impact of greater visibility derived from IoT data

Test the waters with a Proof of Concept

To address the uncertainties on both sides of the equation, we see many of our clients who choose to start with a small-scale project. Often, this is referred to as a Proof of Concept or PoC. These projects provide a low-cost way to “fail fast” and learn what works, what does not, and where the benefits truly lie.

The idea of “fail fast” is inherent in iterative strategies where learning from mistakes and missteps is more valuable than reaching a predestined end point. When you’re not sure where the end point should be, this is the only smart approach.

That’s not to say that you shouldn’t have a goal in mind. A precursor step to starting a project is helping customers think through what outcomes they would value. This is based on where they see pain points and opportunities in their current operation. Then, we think creatively about how they could address them. We use a powerful, flexible methodology called “Enterprise Design Thinking”[2]to facilitate these workshops. It brings together stakeholders from across the client’s business to draw from a wide and diverse set of views and experiences.

Through this type of workshop, we aim to identify one or two focus areas. We can then start to work on how to move forward and implement a technology solution. What’s important, even for these early stages, is to have some clear “challenges” that we can “overcome” and the “outcomes” that will be demonstrated – with a “wow factor” – for an important stakeholder.

The criteria for a Proof of Concept

In terms of the technology foundation for PoC projects, there are some important criteria:

• Quick to deploy
• Flexible and able to handle multiple use cases and scenarios
• Extensible and scalable (thinking ahead to the steps following the PoC)
• Secure (never something to overlook)
• Low-cost entry point
• Predictable future costs (thinking ahead to expansion following the PoC)

Taking the plunge: navigate new territory with a pilot program

Let’s assume the conceptual benefits of IoT are better understood, having iterated through several versions of a solution. Now we are in a position to understand how the technology can be applied, what the costs are and how the benefits will impact the business. At this point the business and technology teams inside the client are usually collaborating on a business case to take forward the next step in the development of the IoT solution.

Most clients we have worked with will be looking at how they deploy a well-understood – both in terms of technology and business – solution to a more “real world” environment. Whereas a PoC could be undertaken in a lab or testbed environment, sooner or later, the solution will have to meet the more demanding conditions of typical users, real assets, unreliable infrastructure (power, networks, etc) and paying customers. This is where a pilot phase begins.

With a carefully selected scope that can be managed and measured to limit risk and continue the development of a verifiable business case, the pilot builds on the PoC and implements the solution on a larger scale. We’ve often seen clients choose a single building or factory or cell within a factory, or perhaps a cohort of selected customers or one class of asset.

Key questions for a pilot program

The questions that the pilot project aims to answer are typically:

• Do the business impact expectations identified in the PoC play out in the real world?
• Is the solution stable over an extended period of time?
• What are the “non-functional” issues that need to be taken into account – for example:
  – How do devices behave in “real world” conditions?
  – How will devices be installed and maintained?
• What is the impact on real users of the solution, whether these be employees, customers, partners? Note: this is a useful point to revisit the outcomes identified in the concept workshop
• Are there unforeseen costs in managing the solution?
• Are there challenging implications of IoT – for example, privacy issues?

The pilot phase typically lasts several months, so there’s time to gather results and assess progress. There can obviously be unforeseen aspects that come to light during the pilot. And this is a period during which the solution continues to evolve and be refined.

When you’re ready for results: move to a full production instance

With the business case clear and the technological implications well understood, it’s time to look at moving the solution into the “production” phase, where the benefits can truly be garnered.

Now we’re talking about rolling out the solution more widely, perhaps in a phased approach. But with the ultimate goal of deploying it as extensively as possible to maximize the benefits. The business case is well understood and often we’ve seen projects pay for themselves through the savings that can be achieved by squeezing greater efficiency out of existing assets – like optimizing maintenance schedules so money is not wasted on unnecessary engineers’ time and spare parts or eliminating manual procedures and data collection.

The ground work that was done in the PoC and pilot phases are invaluable when you move to full production. You will better understand how things will change as the solution is scaled with vastly greater numbers of devices are connected, particularly in the pilot phase. You will also want to:

• Monitor and ensure that the costs of the solution are controlled
• Maintain the highest level of security standards
• Structure the IoT solution so that it is managed as a part of the extended IT estate
• Ensure that service levels are delivered with support from the client’s own IT staff plus the solution vendor’s helpdesk staff

What’s next? Go back to the beginning

Lastly, this can also be the phase in which you return to the conceptual phase and use the insights and expertise that have been acquired on the journey to investigate the next area where IoT can be applied.

Where are you on your IoT journey?

If these phases resonate with your experiences and aspirations for IoT then we’d like to work with you. The IBM IoT Platform provides a cloud-based solution. It incorporates the capabilities that the majority of IoT projects require, based on our experience with many customer engagements. Our Platform is delivered as-a-service, so it’s quick to get started, and hassle free to manage and scale.

What’s more, we’ve thought through how to address security and scalability, so you don’t have to. And because the IoT Platform is only the starting point, we make it easy to extend into analytics and advanced machine-learning solutions, integrate with existing applications or build your own.

The Watson IoT Platform starts from as little as $0.50 per device per month. To learn how to start your own project, visit our site and schedule a consultation.

Resources:

1. https://www.gartner.com/document/3883066?toggle=1
2. https://www.ibm.com/design/thinking/

About the author:Mark Swinson is the Worldwide Go-To-Market Leader for the Watson IoT Platform. He joined IBM in 1995 and has worked many different areas of IBM’s software portfolio. The locus of his career has been in integration – starting with application-to-application integration and evolving through high-performance, low-latency messaging into the realm of IoT. He has been a part of the Watson IoT Business Unit since 2016. There, he works on helping our clients realize value from IoT through cloud-based platforms and industry-aligned solutions.

Begin your IoT engagement