Successful IoT adoption requires iteration, learning

Digital transformation is about leveraging technology to close the loop so that visibility to the products and devices and the processes are continuously expanded and internal mind set. Machine learning and data analytics help.

By Andy Wang August 5, 2021
Courtesy: Prescient Devices

 

Learning Objectives

  • Integrating Internet of Things (IoT) components into products and processes will increase operational efficiency and create competitive advantages.
  • The three IoT adoption phases are visibility, discovery and transformation, and all three feed into one another.
  • Companies adopting IoT programs should learn from their mistakes during their small projects and embrace digital transformation.

The Internet of Things (IoT) industry is undergoing a strategic shift.

For years, the promise of IoT and the next stage of digital transformation to bring greater visibility and automation out to every product, device and process at the edge has been thwarted by current piecemeal approaches. While digital acceleration went into overdrive as the COVID-19 pandemic pushed us faster, and further, into the data age, advancements scaling IoT has remained slow.

IoT as part of products, processes

For digital transformation to occur, IoT solution components will become an integral part of the company’s products and processes. This is similar to integrating embedded software components into products and services.

Integrating IoT components into products and processes will increase operational efficiency, change customer experience, and create competitive advantages.

Successful IoT adoption requires iteration. It’s essential to realize IoT is not one solution. Instead, IoT is an entire process to collect data, analyze data and key extract insights and then manage and update systems and devices out at the edge based on the ongoing analysis. As soon as insights are generated, it drives changes in the process again, forming an ongoing feedback loop to refine, learn and optimize. It is this platform learning loop that drives continuous improvement and much greater adoption rates.

Successful IoT adoption requires machine learning (ML). An organization cannot invest in an IoT product and expect to achieve digital transformation all upon its own. The organization needs to understand the meaning of the data it collects, optimize the data analysis process, and ultimately derive the insights needed to apply lessons learned back into the system itself.

Figure 1: The three IoT adoption phases of visibility, discovery and transformation are not independent. Courtesy: Prescient Devices

Figure 1: The three IoT adoption phases of visibility, discovery and transformation are not independent. Courtesy: Prescient Devices

A study by BCG and MIT Sloan showed the overall success rate of organizations adopting artificial intelligence (AI), with IoT use cases, is only 11%. It’s amazing that a noted trillion-dollar industry has been experiencing only 11% successful adoption.

However, with organizational learning enabled, that likelihood increases to 73%. This is a drastic difference in adoption success rate and it is fair to say iteration and learning are essential for IoT to drive significant financial returns. However, it’s the approach as much as the IoT solutions platform being leveraged that can make all the difference.

Three IoT adoption phases

As companies start their digital transformational journey using IoT, they typically undergo three key phases

IoT phase 1: Visibility

Companies collect data from sensors and equipment so they can perform simple monitoring and detection. The amount of data collected has skyrocketed over the last few years and the challenge now has become managing and extracting value out of this data.

IoT phase 2: Discovery

After leveraging initial results, companies go deeper and generate new realizations. They see how they could change their products or processes to achieve greater outcomes, and this drives them to add more sensors, use more advanced data analytics, and also start to modify other parts of the business that were beyond the initial scope of the IoT project.

Figure 2: Because IoT adoption requires iteration and learning, the initial solution is typically not optimal and might be abandoned before real problems can be resolved. Courtesy: Prescient Devices

Figure 2: Because IoT adoption requires iteration and learning, the initial solution is typically not optimal and might be abandoned before real problems can be resolved. Courtesy: Prescient Devices

IoT phase 3: Transformation

When analyzing results from the discovery phase, companies may find one or a few results that are transformational – perhaps to the products, business models, product mix, support or sales. Companies typically do not realize IoT was transformational when commencing their project. However, once implemented, they may realize they have valuable information and use the data to make strategic and transformational changes. This often involves major changes in products, processes, operational models or business models.

The three IoT adoption phases are not independent. They form multiple feedback loops within and essential to the IoT process itself. Closing these loops are the key to achieving transformation that actually moves a business and its key initiatives forward.

Getting to the transformational process is also not within the immediate initial moment of analytics, feedback and discovery. It’s a commitment to create the next essential step forward. This could take key months to years to achieve. The iterative process of ever-discovery and refinement will continue. In all cases, while technology evolves and pushes us ever forward, internal teams, collaborating, are the ones to drive the ultimate discovery and transformation that changes the world around us and the business landscape.

Shifting toward an agile IoT adoption framework

In a traditional IoT adoption framework, a company is required to come up with a detailed adoption plan and allocate budget and resources at the onset of the prototype and adoption process.
This is a major challenge because it is rare the initial dedicated team and company understands the expectations and implications of IoT adoption at the very start of the process.

The company is then challenged with finding a system integration team to do the work according to the plan. Because IoT adoption requires iteration and learning, the initial solution is often not optimal and closer interaction with the system integration team over a long period of time is required. This process is inefficient and frequently unfortunately results in project abandonment before the IoT program even begins to provide feedback.

Figure 3: The company identifies one initial key problem the IoT can solve as a proof of concept for large-scale projects down the road. Courtesy: Prescient Devices

Figure 3: The company identifies one initial key problem the IoT can solve as a proof of concept for large-scale projects down the road. Courtesy: Prescient Devices

Today’s IoT successful adoption is shifting to an agile framework.

Instead of coming up with a complete plan at the onset, the company identifies one initial key problem IoT can solve. The company then solves this problem at minimum budget and resources, with greater agility and at maximum speed. Based on the learning through the process, the company defines the next action item, which either augments or expands the existing solution to solve a different problem.

The agile framework does not require the company to define the complete adoption plan and allocate a large budget and significant resources right away. The agile process also improves the quality of each project as the company iterates through the loop. With this framework and coupled with today’s low-code IoT technologies, the company can often execute this agile framework with its own internal engineering team. The team with the real-world experience of their company and product lines and that knows the unique particulars can make a project more impactful.

Enabling iteration and learning for IoT projects

The complex technology has presented challenges to companies working with IoT. This also has prevented them from performing iteration and learning. However, this is changing. The advent of low-code tools is reducing technology complexity. Low-code refers to software tools that enable engineers and integrators to work with data without advanced programming or data science skills. Low-code tools often present themselves in forms of tables, workflows, or functional-block diagrams that are simple and intuitive to work with.

Low-code IoT development tools enable engineers to work on data at the edge and in the cloud. They simplify deployment, management, scaling and secure communication, which used to make IoT development inaccessible to engineers. Low-code AI platforms also make training and using ML models easier than ever before. Engineers can provide a set of training data, and these platforms will generate an optimized model automatically. Low-code software can even be used to build web applications that interface with databases or enterprise resource planning (ERP) systems.

Finally, when adopting IoT, companies should take the following strategic approach:

Action matters: Take small, but quick steps

The long-term outcome of IoT and digital transformation is usually not well understood at the onset, but companies should take small measurable steps that keep them moving forward one step at a time. Commit to the process, with the goal of insights learned along the way as you scale and expand into greater full-scale deployments.

IoT trial and discovery with full production in mind

IoT is an evolutionary process and some experiments will fail. The key is not to avoid failures, but fail quickly, iterate and learn from past mistakes. The real breakthrough winners will emerge through continuous experimentation and commitment to iteration and further discovery.

Embracing digital transformation

Digital transformation is not only about technology; it is about the internal mindset. It is about leveraging technology to close the loop so visibility to the products and devices and the processes are continuously expanded. With the right data analytics, visibility leads to discovery which leads to actions to improve performance or efficiency. Further coupled with the right strategies, transformational impacts can be achieved across business and operational models.

Andy Wang, founder and CEO of Prescient Devices. Edited by Chris Vavra, web content manager, Control Engineering, CFE Media and Technology, cvavra@cfemedia.com.

MORE ANSWERS

Keywords: Internet of Things, digital transformation

CONSIDER THIS

How are past IoT mistakes help with future projects?


Andy Wang
Author Bio: Andy Wang, founder and CEO of Prescient Devices.