iiot data silos

October 8, 2019

By Ben Dolmar, Director of Software Development, Nerdery

Industrial Internet of Things (IIoT) adoption continues to accelerate. Eighty-six percent of companies had IIoT solutions in place in 2018, many of them for the first time. Roughly half had been using the solutions for less than 12 months. Companies breaking new ground in IIoT have a unique opportunity to prevent data siloing from the start to avoid backtracking to patch past bad behavior.

How new data silos can happen when adding IIoT solutions

IIoT allows manufacturers to better track data originating from machines making it more valuable for all internal stakeholders. However, value only exists if data is translated into complete, accurate and actionable information.

When piloting IIoT, manufacturers frequently start planning around a single piece of machinery. The temptation at that point is to ignore the processes and other machines that also will be impacted. That’s where new data silos form.

Common technical pitfalls that result in data isolation include:

  • Producing data only from new machines
  • Leaving data in competing inconsistent formats from multiple vendors
  • Not translating raw data into common higher-order constructs

When data is left in silos rather than an integrated system, businesses miss out on improvements in the opportunity to drive efficiency across production lines.

How to avoid creating new IoT-related data silos

Starting with a small test across a non-essential process is a good idea; however, in order to ensure that the smaller case lays the groundwork for future expansion, it is critical that the planning and design look past the initial build. To do so, ensure that you take the following steps:

  • Determine your KPIs and goals early. Create early goals around testing a hypothesis instead of bringing a product to market. More than 90 percent of IIoT users cited machine health-related goals like device management and optimization as the impetus for their investment. Meanwhile, just 18 percent of users were looking to increase production.
  • Use data intelligently. While it’s easy to get excited or overwhelmed by all the new data at your disposal, manufacturers must zero in on the precise pieces that actually help optimize a process.
  • Look for, learn from and correct existing data silos. Ignoring this important step inadvertently contributes to new silos, but correcting existing silos will help optimize the rest of your manufacturing process.
  • Retrofit older machines to provide a holistic view. Most industrial equipment has a 30- to 40-year lifespan. As deployments continue, manufacturers need to retrofit legacy machines to work alongside new IIoT-enabled machines.
  • Consider long-term usage. Standardize the formatting, processing and storage of your new IIoT data from the very beginning to sidestep siloing and ensure long-term value. Remember that your current workforce won’t be the only one interacting with the data, and that future employees will depend on its accuracy in making future business decisions.

While all these technical steps are important to ensure data silos are avoided when implementing IIoT solutions, they’re only half the equation.

The workforce side: Integrating solutions and eliminating silos

Frequently, data silos develop when companies deploy solutions that take a narrow, technical and task-oriented view of the problem. For example, a problem statement like, “I want to get data off this machine into the cloud,” is a limiting view to the problem. It leads to a solution that puts the data into a database in the cloud but does not result in operational changes.

If instead the problem is framed in terms of business impact and human decision-making, companies can look for ways to pull together all of the relevant data. For example, a problem statement like, “I need to understand which of my machines have capacity so that I can appropriately limit my work-in-progress,” encourages pulling in metrics from multiple machines and systems. It might lead to analysis of transit time and operator efficiency.

Recently, Nerdery was working with a client on a system for integrating IoT data into a factory management solution. Prioritizing users and their needs drove the shape of our solution and its technical implementation. When we approached the problem we started by:

  • Auditing and verifying end-user needs. In researching their IIoT solution, Nerdery and our client toured manufacturing facilities to observe real-world workflows. We met with workers on the factory floor to understand the information needed to optimize IIoT for the necessary tasks.
  • Including domain experts and key stakeholders. While external research was important, the client also included its own employees’ opinions and needs throughout the process — a crucial step to ensuring compliance. An intuitive solution, similar to the tablets or smartphones an operator might use at home, led to the best feedback during development.
  • Connect end-user needs and business needs. Don’t add IIoT technology for the sake of bringing in something new. IIoT should serve a real, demonstrable purpose. In the factory management example, the solution was deployed to metalworking machinery capable of producing large amounts of complicated information. However, in meeting with machine operators and managers, Nerdery found that the most valuable data was a much simpler subset that provided an insightful look at each step of the manufacturing process.

Using those tools we were able to design a solution that focused on the data that helped people make better decisions both at the line- and operational-level.

The biggest opportunities for IIoT lie in marrying a broad spectrum of data types. Avoiding data silos from the onset is the crucial difference between companies who find success in IIoT and those who flounder.

Ben Dolmar is the Director of Software Development at Nerdery, a Minneapolis-based digital consultancy of strategists, designers and engineers. In his 11 years with Nerdery, Ben has contributed to more than 500 projects and worked on mobile and IoT apps for entrepreneurs and Fortune 500 companies. A graduate from the University of Wisconsin–Madison with a dual major in journalism and political science, Ben joined Nerdery in 2007 as a Senior Developer and served as Principal Software Engineer before becoming Director of Software Development. 

Contact Info: info@nerdery.com