The Seven Habits of Highly Successful IoT Projects


In the past few years, IoT has been greatly overhyped; hence, many IoT vendors had to convince the market of their pedigrees by touting use cases and customer success stories. In reality, IoT was only getting traction in very few markets—manufacturing being one—with the rest trailing behind.

Now, however, there are a few indications that the IoT market is maturing. First, we see the number of platforms consolidating. Second, we see more respected analyst firms coming to terms with what IoT platforms are supposed to be and consequently ranking the platforms. Third, customer use cases are becoming available in abundance.

So, it is time to move from the convincing stage to the educating stage, sharing experiences of what works and what doesn’t—and where to focus to start the IoT or IIoT journey.

Based on my personal experience in the last years with numerous IoT opportunities and projects across the globe, I came to the realization that implementing an IoT project is hard—very hard. IoT projects are complex, with many stakeholders involved, and although IoT has a lot of technology inside, if you cannot capture the business value with it, the project is doomed to fail.

I came to the conclusion that, although the reasons that projects fail are many, there are some traits that successful projects have in common. Taking the lead from businessman and author Stephen Covey, I like to call them “habits.” And, as seven is the magic number, I shall limit them to that extent for the purpose of this article. Taken one-by-one they are not revolutionary—together they are powerful.

Here are the seven habits:

  1. Lead with use cases.
  2. Work in multi-disciplinary teams.
  3. Put the platform at the heart.
  4. Work backward.
  5. Be obsessed with data.
  6. Embrace AI and machine learning paradigms.
  7. Plan your business continuity.

Although some of these might sound a bit cryptic, never fear; in my upcoming articles, I will dive deeper into each habit and provide further insight on why some companies continue to deliver consistent success in the domain of IoT.

Let’s start with habit one: Lead with use cases. Use cases have proven to be a very effective way of modeling software systems. Use cases allow organizations to build a mutual vision of the problem at hand by bridging the gap between the subject matter experts who understand the problem and the IT teams who understand how to build a solution.

They provide a powerful tool to quickly inform an organization as to what management wants to achieve. Still, a use case can go astray if management doesn’t take the time to understand the implications. Consider the predictive maintenance use case. Many companies I have worked with consider this to be the most important outcome of an IoT project. For example, at the end of 2016, I met with a customer who said, “We want to do predictive maintenance on our refinery installations.”

I asked if the company had historic data on their machines available. I asked if it had a data scientist that had already created some models that could be used to create a basic predictive maintenance project. All answers were “No” or “Not yet.” Still, they insisted on having a predictive maintenance program in place within 3 months. Needless to say, there is still nothing in place.

Why? Two reasons. One, it is important to understand that use cases don’t live in isolation. Use cases need to be communicated to and adapted by the organization. One of the easiest starting points is the business model canvas approach. This simple approach defines 10 aspects of the use case which you need to address to come to a complete picture of what the company wants to achieve with the use case. Relevant aspects include the customer value, channel, stakeholders, and revenue models, to name a few. They are often mapped in a visual representation, making sure that the whole model can be viewed at a single glance.

The second reason is that use cases have to adhere to the maturity curve; the use case complexity has to be aligned with the maturity of the organization. If any use case is too advanced for the organization, the learning curve becomes too steep and adoption thus too expensive.

An easy way to think about this is to view your organization’s competency on a scale of three stages.

  1. The data-centric stage: This stage means that your organization can handle data volumes generated by devices at a large scale, along with connectivity and device management issues. It can store the data generated by those devices continuously in a reliable way. On top of that, parts of your organization understand how to get access to this data and how to visualize it in one form or another.
  2. The process-centric stage: Once the data stage is passed, you might want to alter existing processes to take advantage of the insights generated by the data, or even create new processes. A relevant example might be how an automatic replenishment process can take advantage of real-time inventory data feeds, ensuring that customers are never out of stock.
  3. The analytics-centric stage. In this phase, data scientists are analyzing the data and creating new advanced analytic models such as machine learning and AI models to create predictions and actions that automatically optimize decisions and processes. Here the biggest gains can be made in terms of customer experience but also in optimization of the workforce—and in reducing spare parts inventory levels.

Bottom line: If you are thinking of going after a use case that requires advanced analytics, then consider the fact that Rome wasn’t built in one day, and neither will your IoT infrastructure. Take the time for your organization to adopt the stages and start reaping benefits from day one.

In my next several articles I will go into more detail about habits two through seven. See you next time!



Newsletters

Subscribe to Big Data Quarterly E-Edition