77729047 | Wavebreakmedia | Dreamstime
Dreamstime M 77729047

Implementation: The most overlooked part of digital transformation for midsize manufacturers

Oct. 16, 2023
The merits and drawbacks of an "all-at-once" approach vs. "small steps, one plant at time." Plus: How "sprint implementation" can succeed.

It’s no secret that the manufacturing industry abounds with Software-as-a-Service (Saas) platforms. And if we’re being honest, many of them can be hard to tell apart on paper. After all, most offer some combination of the following: connecting siloed data, reducing the amount of time spent jumping between platforms, and processing data for relevant insights.

I can think of about a dozen platforms that fit this description. Nevertheless, most people spend a great deal of time pouring over the technical differences between these platforms without thinking about implementation. I think that’s the wrong approach.

See also: Sharing is caring: Webinar covers how pooled data amps up performance

In my two decades in this industry, it’s not the companies with the flashiest marketing materials or the impressive-sounding features that win in the long term. It’s the companies that consistently implement their solutions quickly and with minimal disruption to operations that come out on top because, at the end of the day, no amount of “Industry 4.0” innovation is worth it when it stalls production or if adoption can’t be achieved on the factory floor.

Two approaches to implementation

There are two main ways of adopting digital transformation in manufacturing. Both have SaaS platforms at the heart, but the implementation plans differ widely. On the one hand, some seek to change everything in one action. On the other hand, some take a small-steps approach, breaking implementation into bite-sized pieces.

All-at-once approaches

Recognizing that change can be difficult, some manufacturers try to rip the Band-Aid off, make several drastic changes to production at once, and roll out a new SaaS platform at all—or as close to all—of their locations simultaneously.

Of course, this has an obvious appeal. On the surface, it seems like the fastest route from point A to point B. Unfortunately, it’s often more complicated than it sounds. First, given the complexity of manufacturing, it’s rare that significant customizations to a platform aren’t needed—not just for an enterprise but also for its individual plants.

See also: AI coupled with 5G improves industrial network operations

The process of trying to fix a major software system that also has associated business practices all at once is known as “rip and replace.” This might work when rebuilding a commercial building that has been gutted by fire, but it rarely works well with an enterprise-level software system.

For example, one of our customers came to us after they found out the hard way after attempting to move off an Enterprise Resource Planning system for more than four years. Their time was filled with product and planning meetings, and they still do not have an approved template for the new ERP system to which they are moving. Despite this disruptive effort, the users have not seen any benefits yet from the new system.

Secondly, sudden and drastic changes rarely go over well on the factory floor. In my experience, the more you try to change, the more pushback you get from operators. No matter how much you believe in a change, if your people on the factory floor won’t adopt it, it will be dead on arrival.

Small-steps approaches

The small-steps approach, or as I prefer to call it, the “sprint” approach, focuses on addressing one pain point at a time rather than trying to solve multiple problems simultaneously. You would want to make steady progress by fixing one user issue at a time, rather than replacing the entire underlying system. For example, rather than trying to address digital forms, predictive maintenance, and quality control improvements simultaneously, you would focus on solving one issue first and then move on to another.

Carefully replacing pieces of a system allows for gradual change and minimizes disruptions and system downtime. The only way to move this quickly is to connect to a SaaS model that has the latest software and AI engine updated throughout the system.

See also: Cybersecurity: ‘Largest obstacle to adoption of smart manufacturing technologies’

The key to a small-steps approach is planning implementation so that each step builds on the previous one and can use the ROI from the last step to help fund the next. The latter point is vital for midsize manufacturers who can’t afford large upfront investments and must maintain healthy cash flow throughout a project.

While this approach often looks slow from the outside, much like the tortoise and the hare, it can prove faster in the long run. There are several reasons why the sprint approach works best. First, incremental approaches are more likely to gain employee buy-in, preventing projects from stalling on the factory floor. Second, the sprint approach tightens feedback loops by breaking implementation into short periods, usually less than 90 days, allowing for faster customizations. Finally, by improving cash flow, this approach is more sustainable, allowing manufacturers to see their digital transformation initiatives through to the end rather than running out of money—or executive-backing—halfway through.

Sprint implementation in action

Recently, I worked with an industrial manufacturer to bring much-needed updates to their operations. About 25 years ago, they built an Excel-based application that combined manually entered QC data with sensor data from their data historian. Back then, this was cutting edge. A quarter-century later, it was time it went the way of the dinosaurs.

They initially took an all-at-once approach. But the complexity of the transformation meant that 18 months later, they were still at square one. Then, they tried our 90-day sprint approach.

The goal was to replace their outdated reporting with FactoryEye’s dashboard, which gave them real-time reporting across their assembly lines at one plant. They kept initial costs low by focusing on one pain point at one plant. Once they gained proof of concept (POC), they rolled out solutions to eight more pain points, still at just one plant. This process was completed within the first 90-day sprint. During the sprint, the team was careful to keep data flowing to both the old system and the new system until they could verify the reports were accurate. At the end of the sprint, they were able to turn off the old system, proving that the sprint goal of replacing the reporting system had been achieved.

After the first sprint, the team regrouped and evaluated the solution and spent another 90-day sprint refining based on feedback at the first plant. This set the stage for the next set of sprints, where they rolled out all nine solutions at 18 plants in just 18 weeks. Overall, the key to success was that the manufacturer set clear goals for each sprint. At the end of each sprint, they ensured the goals had been met, and that the solution was actively in use at the relevant plants.

The other benefit was that the plant floor managers saw how quickly FactoryEye was able to provide a dashboard that gave them actionable data, and just as importantly bring it to a fully functional state where it can be used in the plant. This worked wonders for employee buy-in to the new system. The plant managers can now view real-time data from plants from throughout the USA, giving them a way to manage the variables on the assembly line better than they had ever expected.

Questions to ask before implementing manufacturing SaaS

While everyone asks about things like integrations and pricing when looking into a new platform, don’t forget to ask about implementation as well. And don’t settle for vague timelines. Ask about how a provider has helped manufacturers gain employee buy-in during previous projects, the ROI timeline for comparable projects, and how they’ve shaped their implementation strategy. Chances are the answers to these questions will tell you far more than the features page on the website.

About the Author

Jonathan Kaplan | COO, Americas, FactoryEye by Magic Software

Jonathan Kaplan is chief operating officer for the Americas of FactoryEye by Magic Software, which enables data-driven smart manufacturing.