No guessing games: three steps to take for effective IoT implementation

IoT implementations often do not get past the pilot stage. One reason is that organizations fail to take action to realize their return on investment (ROI). Think about how much data you receive from your phone, your smartwatch or your video doorbell, and how much of it you ignore. The same goes for an operator in a warehouse or on a factory floor.

Three steps to effective IoT implementation

To make IoT work, there are three steps to follow: digitalize (collect the data), analyze (turn the data into insights and recommendations) and act (turn the information into action).

Each step must be guided by a business outcome. So, what are you trying to achieve?

People love technology, and IoT is no exception – but just because you can connect something, doesn't mean you should. (How useful is a connected toaster, really?) Start with the business outcome in mind. This could be, for example:

• Augmenting workers’ skills through automation while improving their safety and security
• Cutting operating costs and downtime
• Increasing yield or productivity
• Tracking inventory
• Measuring carbon for sustainability

Then, move on to the three steps.

1. Digitalize the physical environment: collect the data

What are you digitalizing? Is it a factory, a warehouse or a hospital being turned into a smart building that uses a network of sensors to know whether a door is open or closed, what the temperature is in a freezer, or if a pallet is sitting unattended in a loading dock?

We turn an environment into data and then we transmit that data through the right kind of network. Do we take a highway, a side road or a bike path to get the data to where someone can make sense of it? For a farmer in a cornfield, a public 3G or 4G network might work better than a private network. A factory or a hospital with an overwhelmed Wi-Fi network may require stronger security and performance to accommodate use cases such as high-definition video feeds, autonomous systems and robots, or augmented and virtual reality, and these needs would be better met through a 5G network.

2. Analyze: turn the data into information

What does a temperature reading of 40 degrees Fahrenheit mean to you? If it’s for a refrigerator, it means the refrigerator is broken. So, we also need context to determine whether the data means something. A key issue is that there are many siloed point solutions and data sites when data must rather be aggregated. Take the temperature example: what is the weather like? What is the location? Is a door open and the product being exposed?

Then, do we store the data or process it at the edge? Edge processing increases response time and efficiency when there’s a high volume of data and a need for low latency. An analogy is your video doorbell. You can scroll through all its cloud recordings to see if anyone visited. But it’s more efficient to process the data at the edge, so you are alerted to a visit right away. The average factory generates petabytes of data regularly. Does it all need to go to the cloud?

3. Act: turn the information into action

So, IoT is about getting the raw data, analyzing and contextualizing it to determine an action to be taken, then sending the right amount of information to the right person at the right time and in the right format (SMS, mobile app and so on).

IoT is ineffective unless the right person actually takes action. To achieve true operational excellence, you have to follow the right processes all the way to the end.

For example, you’d ensure a proper workflow through a tool such as ServiceNow to get a certified plumber to fix the leak immediately – and, oh, it’s indoors and it's humid, so place a dehumidifier there for two days to prevent the growth of mold, then do a quality check and close the ticket.

Because IoT is about data, we also have to consider issues of data privacy, security and governance. Is a camera picking up people's personal habits at work? Is an audio sensor that monitors the sounds of a machine also recording conversations between workers? These issues will vary by region but must be taken into account.

NTT DATA has all the IoT building blocks

IoT implementations cannot take a one-size-fits-all approach. They have to be fit for purpose and rely on the right components for the environment (a hospital, an oil platform, a large farm or a factory) and the specific use cases.

Our approach to IoT is to deliver the full stack, from strategy to execution. Others may focus on just one layer, such as the devices, the connectivity, the analytics or the application, but then the client is left with a point solution and an unclear path to ROI.

By providing the full stack, we can also implement security by design across all the layers. This is important, as any new technology introduces risk.

We are therefore perfectly positioned to deploy and manage IoT networks because of our strong network and security capabilities.

At NTT DATA, we have a full set of infrastructure building blocks. These include Edge as a Service and Network as a Service. That’s everything from our Transatel connectivity options (Cat M, 2G, 3G, 4G, 5G) for things out in the field to private 5G for secure and reliable high-performance connectivity. We can bring in private networks such as LoRaWAN, Wi-Fi and others to connect mission-critical sensors.

We also provide edge computing and help our clients determine what needs to be processed at the edge instead of the cloud.

We also offer Multicloud as a Service. As a leader in data centers, we can define the exact mix of private and public cloud you need.

We have the operational building blocks to make the data actionable. These include our Smart Platform that integrates previously siloed data to provide deep insights and recommendations. We are using this for everything from the City of Las Vegas to the NTT DATA INDYCAR SERIES. We focus on getting the right amount of information to the right person at the right time to make that data actionable.

And, finally, we have our systems integrations with partners such as SAP, Microsoft and ServiceNow to facilitate enterprise resource planning and automatically execute workflows in case management systems.

Adapt to thrive

Organizations should not deploy technology before understanding its impact on their operating model. Processes need to change, incentives to take action must be put in place, and roles and responsibilities have to be redefined. For example, who will maintain the sensors and change their batteries, and what happens when that person is unable to do it? Organizations must be able to adapt to their new solutions.

In conclusion, to make IoT work for your organization, ensure that your data is secure, trusted, relevant and actionable, and that your infrastructure is fit for purpose and scalable.