Blog Series Part 10/10
We've been working on quite a few industrial digitalization projects lately, and there's this pattern we keep running into. Companies invest in sensors, set up data collection, build dashboards, and then nothing really changes.
The data is there, being collected, sometimes even visualized quite nicely. But when we ask what they're doing with it, the answer is often just "monitoring" or "compliance checking."
That's expensive noise.
Recently we were working with two different companies at almost the same time. Both had decent automation. Both were collecting production data. But the outcomes couldn't have been more different.
The first company has a really advanced setup with tight quality control, full automation, and data flowing everywhere. But when we dug into what they actually do with all that information, it turned out they're basically just using it to verify they're following their predefined process correctly. No one's interested in learning anything new from it or curious about what the data might reveal.
The second company is almost the opposite. Their automation isn't as sophisticated, but they're obsessed with understanding what's actually happening. Every piece of data becomes a question: why did that happen? What does this pattern mean? Could we do this better?
The difference comes down to culture.
In our experience, there are a few common reasons why data never makes it from collection to decision:
No ownership. Someone needs to be responsible for turning insights into action, not as extra work on top of their real job, but as their core responsibility. We've seen too many cases where data just sits there because nobody's job is to do anything with it.
Missing context. Data without operational context is just numbers. You need to understand why a temperature spike occurred, how it connects to product quality, what it means for the maintenance schedule. We've seen cases where operators each adjust their processes differently, all producing acceptable results. Until you capture that context, you're just recording values without understanding what they mean.
Cultural resistance. This is the big one. The Finnish business mindset especially struggles with investing in knowledge that doesn't have an immediate euro value attached. "Show me the ROI by next quarter or forget it." But understanding your process better has value even before you know exactly how to monetize it.
Here's a case that really stuck with us: a food processing company spent years building energy consumption dashboards with beautiful visualizations showing which equipment used the most power and when usage peaked.
Workers barely looked at them, and behavior didn't change.
Eventually, they just stopped asking humans to act on the data. They built controls that automatically optimized equipment usage based on the patterns they'd identified. Problem solved: significant energy savings without requiring anyone to change their habits.
It worked, but it also meant they never built the organizational capability to actually learn from their data.
We're not saying automation is wrong. Sometimes it's the pragmatic answer, especially in environments with high turnover or other constraints. But there's a difference between automating away a problem and building a learning organization.
We've noticed that data utilization tends to happen in stages, and most companies get stuck somewhere between stage two and three:
Stage 1: Closed-loop automation. The system reads sensors and adjusts automatically. Like maintaining the optimal ratio of virgin to recycled material in real-time extrusion. No human intervention needed. This works, though the optimization stays focused on one specific thing.
Stage 2: Intelligent alerts. The system spots anomalies and tells someone about them, maybe even recommends an action. Predictive maintenance lives here. Still requires humans to decide what to do.
Stage 3: Strategic learning. This is where it gets interesting, and where most companies struggle. Using historical data to optimize processes, identify patterns, make better business decisions. This requires analytical capability AND organizational willingness to change based on what you discover.
Stage 4: Cultural integration. The rare companies that get here have teams at every level naturally incorporating data insights into decisions. It's just how they work.
The companies that reach stages 3 and 4 aren't just optimizing existing processes. They're finding opportunities they didn't even know existed.
Data analytics and AI amplify human expertise rather than replace it.
We've heard from manufacturers where every operator adjusts the process slightly differently, all producing good results. That operator who's been on the line for 30 years can hear when something's about to fail and knows intuitively when to adjust parameters.
That knowledge is invaluable and should be guiding your data strategy, not being replaced by it.
The best implementations we've seen combine that deep process expertise with data infrastructure. The veteran operator's knowledge informs what to measure and what actually matters. Analytics reveal the patterns behind their intuition. New operators learn from validated best practices instead of starting from scratch.
But this only works if you involve those experts from day one. You can't just install sensors and expect to reverse-engineer decades of experience.
So how do you move from data collection to actual decision-making?
Start with use cases that matter to the people doing the work. Don't chase theoretical maximum ROI. Find problems that make workers' lives easier or processes more reliable. Success builds momentum.
Get the right ownership in place. Someone needs to own the insights and the actions that follow, with the time and authority to actually do something.
Involve your process experts. The people who know the work best should help define what's worth measuring and what good looks like.
Be willing to invest in understanding before optimizing. Finnish business culture wants immediate returns, but sometimes you need to learn what's happening before you know how to improve it.
Build feedback loops. When decisions are made based on data, track what happened. Feed that learning back into the system.
After ten posts in this series, one thing has become clear: the value of IIoT comes from connecting people, processes, and data in ways that drive real outcomes. The technology enables that, but the transformation happens when organizations build the culture and capability to actually learn from what their data is telling them.
We're still figuring out the best ways to make this happen. Every company is different, every industry has its own constraints. But the pattern is consistent: the companies that treat data as a learning opportunity rather than a compliance checkbox are the ones that improve.
In the end, the decisions we make with data matter more than the data itself.
