A 3-step process to overcome the IoT solution log jam
I was caught by a recent statistic from 451 Research that almost half of Internet of Things projects fail to deliver a return on investment. As a person who spends every day immersed in shaping IoT solutions for IBM clients, I stopped: how can that be?
And… no wonder there’s so much chatter out there right now about whether or not to move pilots to production. (The same report noted only 54 percent of IoT solutions move from proof of concept to production.)
I think a big part of the problem is the fragmentation we see in the IoT solution space. Let’s face it: with more than 900 vendors in the IoT ecosystem, how can one reasonably keep up?
My opinion: we’ll see more return on investment from IoT solutions when teams become more disciplined in their approach to building those solutions.
That discipline is a product of three fundamental steps:
- Continuous management
First: Why fragmentation?
Before I jump into each of the steps, let’s consider the natural consequence of walking onto a carpet of more than 900 solution providers….
It’s overwhelming! And it leads to point solutions.
Consider, for example, the case of a perishable goods tracking supply chain. Companies in this business strive to get the products to customers (1) in the shortest possible timeframes, while (2) keeping the food at a safe temperature.
IoT devices at work in this scenario may be temperature sensors, asset tracking (such as: RFID, GPS, and beacons) to make them implementation more successful.
Today, teams typically buy solutions based on these point requirements (e.g., temperature, location tracking, etc.). Doing so leads to creation of a siloed device architecture—and to that middling ROI stat I mentioned earlier.
A better way
With growing adoption of IoT and the vast number of data points that IoT can provide, I believe companies need to abandon the point solution mentality and adopt a services framework to rapidly deploy multiple IoT solutions.
So, what do the three steps of a services framework entail:
Step 1: strategize
Start with the big picture. Every IoT solution should be built from the ground up, based on the desired business outcomes and including edge devices. Start with the outcome you want before jumping into the specific devices and communication technology that will best achieve that outcome.
Step 2: design
Design a solution that can be implemented to business scale. And scale means not just rolling out to production, but ability to reuse the same solution across different use cases.
Then conduct proofs of concept and pilots. Use this time to define clear production rollout specifications. Clearly lay out an implementation plan based on the desired ROI. Encourage the adoption on emerging standards like LWM2M or OneM2M so that you have a device OEM agnostic service framework.
Step 3: deploy and manage
IoT edge devices need to managed like corporate-liable devices under a common services framework. These devices need continuous monitoring and management to ensure uptime and availability of the service, and again, ROI.
What you get
So, back to the case of a perishable goods tracking supply chain, here’s how the services framework would play out.
The company would begin by defining a value statement, such as: Reduce the amount of food that goes bad (also known as inventory loss).
Connected to that would be a measurable business outcome—let’s say, reduce inventory loss by 20 percent.
- What’s the current loss? How much should the company target to save through this implementation? There’s such a wide range of costs for IoT devices. Having a clear understanding of budget for the solution will guide allowable expenses for devices and technology.
- How will solution be used? Decide what dashboards or other visualizations the end users (such as retail store managers and supply chain managers) may want, need or otherwise benefit from to help manage operations.
Next up: make architecture decisions and create a solution.
- Make the solution more cloud centric versus edge or hybrid?
- Should the sensors and other IoT devices talk directly to the cloud (a point solution) or do we implement a gateway for enhanced compute capability at the edge?
A gateway may make the solution more scalable, so it’s a centralized hub capable of taking on more devices (such as pressure sensors, cameras or motion sensors) for future implementations.
- Infuse AI and analytics to give actionable data points (such as status of shipments, health of inventory packages) to employees in the field?
- Create a new dashboard to visualize data from the sensors or render the data to an existing application?
Since the IoT market is so fragmented and there could be issues with interoperability, the design phase should include testing selected devices in a lab, as a proof of concept. Then piloting the solution at selected sites. And before a full rollout, creating clear production specifications.
Deploy and manage phases
Finally, systematically deploy the devices, based on production specifications and an agreed upon project plan. Then employ a service management methodology similar to those for other critical IT assets—perform over-the-air updates, monitor device health, etc.
Anticipated spikes in the next 5 years in the number of installed IoT endpoints will no doubt strain IT infrastructure. Businesses will need to adapt, and they’ll need a systematic approach to their rollout and management.
The services framework I have outlined is a step towards that.
Learn about IBM IoT device management services.