A lot has been written about why IoT is going to be absolutely essential to your company. If you are convinced as to the why but wonder about the how, pay attention. Based on all the lessons learned from my engagements around the globe, I am going to tell you how. It basically comes down to three basic phases: preparation, the “real stuff,” and embedding it in the organization.
Phase 1: Preparation
First of all, I presume the reason you are reading this is that you are responsible for IoT in your company. If so, start off by tightly defining the business need. As you embark on IoT, you need to ensure that your IoT-based solution fits into your overall business strategy. In order to do that, you need to define the immediate objectives of the solution. If possible, try to understand the future business requirements as this might help to define the additional capabilities required going forward—paving the way for broad-scale adoption.
Once you understand the solution to kick off with, create the team to make it happen. Teamwork is needed to succeed in implementation. Very few solution providers have an end-to-end capability to implement IoT solutions. (Don’t confuse IoT Solutions with IoT platforms, an IoT platform can cover part of a solution but never 100%.) To mitigate the implementation risk, you should involve multiple providers and encourage best-in-class in solution components provided by various expert players. Think about (cloud) IT infrastructure, data management, analytics engines, and various other technologies.
To continue with IT infrastructure, you might not have the right processes and data needed to support IoT solutions. Therefore, it is imperative to improve business processes and upgrade IT infrastructure to support any IoT-based solution before it is deployed. You can choose to follow a step-by-step approach where you first implement your IoT capabilities, addressing existing process and infrastructure needs, and then gradually roll out advanced functionalities.
Finally, embrace a data-driven culture. Presently, most companies follow a knowledge worker-centric approach. They completely rely on the experience of their domain experts. Given the fact that, in most industries, the workforce is aging, the time has come to adopt a data-driven approach, as many of your experts will retire or leave the workforce within the next decade.
Phase 2: The ‘Real Stuff’
So this is all well and good as a preparation plan, but what about the real stuff? You know—when the going gets tough, the tough get going. Let’s go through the steps with an example.
Imagine that you have a manufacturing company, and let’s presume that you see a business need to implement a predictive maintenance case (not an uncommon use case).
You put a team together, and the seasoned engineer in your team tells you that putting a hand on the machine to feel the vibration generally provides a good indication of the machine health. After all, the most frequent faults come from faulty bearings, imbalance, misalignment, and loose parts, which all result in vibration disturbances.
With your data-driven mind, you quickly come to the realization that having the ability to measure vibration and then analyze the vibration to make informed decisions seems to be a sensible predictive maintenance case.
You ask the engineers to find a vendor that can deliver a vibration sensor that can be suitably mounted on an appropriate point on your machinery. This will likely open up a whole range of discussions with the team:
- What type of vibration should you measure? There are many types of vibrations, for example, loose foundation bolts, imbalance, misalignment, fan blade faults, gear mesh faults, etc.
- Which machinery to measure? If possible, go for the critical ones that have historically the highest failure rate. Those will bring the biggest benefit to your organization and allow you to make your project visible in the company.
Don’t forget that the vibration measurement device you choose must have the capability to transfer the data generated to your IoT platform for further analysis. While the engineers figure out the best way to place the device on the machine, the data scientist, together with the engineers, has to discuss what the measurements should look like and what data elements and units are included. A good practice is to ask the suppliers of the vibration device for a recommendation—you might be surprised how much they know.