The Industrial Internet of Things (IIoT) incorporates advanced technologies to harness sensor data, machine-to-machine (M2M) communication in various industries.
In the digital world, data can be stored, operations can run non-stop 24 hours a day, seven days a week because everything is connected, and tasks are logged and tracked properly. IIoT allows such tasks to be completed automatically on an industrial level, rather than manually, to save cost and decrease error.
Yet recently research from Cisco found that only 26 percent of IIoT projects are successful. How can this be?
This article, IIoT in Manufacturing—Technology Evolved, explores how businesses can determine whether IIoT is a good investment and what best practices they should follow for implementing IIoT projects.
Content Summary
The Real World
The Stamping Press
Root Cause Analysis
Industrial IoT Disruption Is Not Limited To Technology
Recommended Best Practices For Starting IIoT Projects
Start From The Root Problem Of The Business
Involve Key Stakeholders
Start Simple, But Plan For Growth
Adapt Using Iterations
The Internet of Things (IoT) connects advanced software with sensors and other digital end-devices on a communications network to help transform products and services to work better, digitally or otherwise. Gartner estimates the number of connected devices will grow by 50 percent over the next two years.
Perhaps what’s lesser known than IoT is IIoT. The Industrial Internet of Things (IIoT), sometimes known as the industrial internet, or Industry 4.0, incorporates advanced machine learning (ML) and big data technologies to harness sensor data, machine-to-machine (M2M) communication, and automation technologies in various industries. In the digital world, events are never forgotten because data can be stored, operations can operate for 24 hours, seven days a week because everything is connected, and tasks aren’t overlooked because they are logged and tracked properly. IIoT allows such tasks to be completed automatically on an industrial level, rather than manually, to save cost and decrease error.
Yet recent research from Cisco found that only 26 percent of IIoT projects are successful. How can this be?
The Real World
Traditionally, Industrial IoT is an impressive technology revolution, but it requires more than just technology. There are stakeholders from all levels of an organization (users), business leaders (business) who need to ensure a positive return on investment for the projects, and potential experts in new and complex technologies (known as technologists). Users, business, and technology are the three pillars of data driven systems used in IIoT.
Because IoT system architectures are relatively new, technologists need to prototype solutions to ensure an architectural mistake doesn’t ruin the outcomes of the project. For example, prototyping predictive maintenance algorithms is quite common. However, technology itself doesn’t need a proof-of-concept (PoC). In our experience the stakeholder journey (user experience) and change management process (business process changes) is much riskier than the technology implementation. Much of the technology in the IIoT revolution is built on top of mature technologies (networks, databases, and more) but put together in new and novel ways to create new potential when combined with new techniques in data science.
Focusing solely on a technology POC ignores two other key aspects to IIoT projects—the business and the users. Let’s take a look at a real-life example.
The Stamping Press
A manufacturer of stamped steel parts uses large stamping presses to turn rolled sheet metal into useful widgets. The company was experiencing capacity issues because of unexpected failures with presses causing too much downtime.
The head of the factory led the conversation with agreement from the company’s head of maintenance, which proved to be a good sign of having buy-in from many levels within the business. Not having executive buy-in (business consensus) is another common reason IIoT projects get stuck in PoC purgatory.
The company was on the market to buy a sensor- based predictive maintenance solution to track and improve up-time of their presses. Part of the discovery process required analysis of the most expensive failures, including, when an eight-inch steel drive shaft sheared clean through. The press was out of service for a month for repairs after that failure.
Root Cause Analysis
Excess contamination in the press’ oil caused bearing failure. The failed bearings caused unnecessary load on the shaft which eventually caused it to fail. The company’s maintenance process required periodic analysis of the oil in each press. The lab reports showed an unacceptable amount of rust over a month leading up to the failure.
Did the failure happen because the oil analysis document wasn’t read by anyone? If so, IIoT dashboards and alerting may have prevented this failure to happen. Unfortunately, that was not the case. The staff had read the report and decided they didn’t want to lose the production time to change the oil and repair the bearings.
Industrial IoT Disruption Is Not Limited To Technology
Data has a high value attached to it, but it must be used carefully through effective communication and action. IIoT technologies are great at finding patterns, detecting anomalies, and providing insights. However, businesses are still required to create functioning organizational structures and accountability policies.
Dashboards are a common solution for IIoT systems. The can provide a quick snapshot of an entire operation at a glance. Even without machine learning or advanced data science, simple dashboards and basic rules can provide critical visibility into the health of a system and generate notifications to react to issues in a timely manner. However, just providing a dashboard (technology) doesn’t mean that the proper people (users) will be looking at it when needed. Also, different users will need different information and at different times.
IIoT solutions provide unprecedented visibility into operations but cannot establish new operations procedures on their own. Executives and budget owners will have to accept new cost and training cycles.
Companies that are not willing to change their business processes are unlikely to be successful in their IIoT projects.
Many companies overvalue their current production processes rather than embrace more streamlined and agile ways of working. In fact, most companies won’t find an IIoT solution that doesn’t require a culture change. This cultural philosophy won’t change overnight.
Recommended Best Practices For Starting IIoT Projects
How can any business determine whether Industrial IoT is a good investment? Let’s take a look at some of the best practices.
Start From The Root Problem Of The Business
Many Smart Manufacturing projects are doomed to fail because they start out with the wrong goals. “Let’s collect some data and then figure out what to do with it,” is unlikely to be a winning strategy. One strategy we have adopted is to start with a design thinking workshop. During this workshop we dive deep into the business case to uncover the root cause, not just the symptoms that the business is experiencing. Then we ideate to come up with cross-disciplinary solutions to solve that problem.
Involve Key Stakeholders
Without executive buy-in, teams will not get the support to implement a disruptive technology such as Industrial IoT. Also, identify key stakeholders. It’s not always obvious who the stakeholders are at the beginning of the discovery. In the case above, the company was focused on a maintenance solution, but the root cause was production managers overruling maintenance. The production team was the key stakeholder and they weren’t originally involved in the search for IIoT providers.
IIoT solutions should be seen as another useful tool within the business—not just the IT or maintenance team. The maximum value of the solution can only be achieved when the whole business adapts to the possibilities enabled by IIoT. This will require change management throughout all levels of the operation so it is important that clear expectations are set for all stakeholders.
Start Simple, But Plan For Growth
Industrial IoT solutions have a different lifecycle than traditional enterprise solutions. Identify where errors in the business’s process are likely to occur and focus there first. Identify where business processes can be changed—and where they can’t—and focus on small changes first. Then the technology will be an easier problem because the technologists aren’t fighting process resistance. Then prototype at the lowest fidelity possible.
Industrial IoT software must be easily upgradable to handle security vulnerabilities. Instead of trying to solve all problems up front, focus on the highest valued problem first. Then leverage that software update process to roll out new features as business value is proven.
Adapt Using Iterations
Approach your IIoT projects with clear milestones and short iterations in an agile way. Within each iteration, attempt to prove some business value and ensure all stakeholders are aware of the results for each iteration.
Identify where process is or isn’t changing— one problem can have multiple solutions and the first attempt isn’t guaranteed to be the right one. Focus on adoption and measure it. Address lack of adoption as early as possible— the solution can only solve the problems if all key stakeholders are using it.
A good IIoT solution should easily scale to add more sensors or data points in the future. Sensors and installation are one of the largest costs for a proof of concept project. There is bound to be uncertainty about which sensors are necessary. Having additional sensors to mitigate that uncertainty increases the risk to the project cost and complexity without a clear path to success.
Source: SoftServe
