The real blocker is the gap
Digital talk is easy. Getting value on the floor is hard. Many dairies still feel a gap between central IT and the people who run lines, tanks and utilities. Different goals. Different vocabularies. Different definitions of “done.”
Erik Vedfald from Arla captures it bluntly: “We had a shop window full of products, but we forgot to invite the customers in.” Technology by itself does not move yield, energy or OEE. But collaboration does.
Shift the starting point: business case first
Arla’s lesson is clear. Pick use cases that pay. Then fit the tech. For Vedfald, the right place to start is with a business case, not a tool.
“We will never be able to keep the speed that the world is evolving around into our production environment. What is important is that we find the right cases… and actually lift technology as part of the business cases that are being deployed into the production.”
This means looking for use cases that clearly connect to production performance; power optimization, raw-material tracking or tighter process windows. When you prove value in one area, you build the confidence and data foundation to expand.
Build zones of trustworthy data
Every dairy collects data. The real question is whether that data is good enough to make decisions on. Many sites still chase volume instead of quality, ending up with massive data lakes that look impressive but deliver little value.
What matters is building smaller, trustworthy zones where data quality is verified and relevant to the business case at hand. The data that feeds analytics should come from the processes that truly influence yield, energy or product quality. Not from every available sensor.
Erik Vedfald puts it simply: it’s time to move from collecting everything to collecting what counts. The goal is to understand which data is needed to support a specific use case, and then ensure that this data is consistent, valid and complete. That often means combining production data with laboratory results to create a fuller picture of process performance.
These smaller, high-quality data zones are where real progress starts. Once the data is aligned and reliable, insights become trustworthy, and automation can scale.
Intake and pasteurization are strong starting points. These are areas where good instrumentation already exists and data quality can be improved without major system overhauls. From there, the zones can expand naturally, one process at a time, forming a connected network of reliable data across the plant.
Building trust where it matters
Technology only delivers value when people believe in it. The fastest way to kill a new tool is to let it fail the first few times it’s used. Vedfald explains: “If that tool gives the first three times the operator tries it a wrong answer, then that tool is going to go into the desk, and it will require ten times as much effort to take it up again.”
Trust isn’t built in a workshop or a slide deck, but earned on the production floor. Operators and engineers need to be part of the process long before launch day. They should see how the tool supports their work, not replaces it.
When a system proves reliable in real operating conditions, adoption follows naturally. But when tools confuse or overcomplicate daily tasks, they’re abandoned just as fast. For medium-sized dairies, that makes early user involvement as critical as any sensor or software license.
Turning local ideas into scalable value
Some of the best innovations in dairy production don’t come from headquarters. They actually come from the floor. According to Vedfald, local sites are increasingly driving digital progress:
“There’s a big flip on bottom-up that’s coming in. We have so many amazing people at the production sites… they come up with amazing ideas in sustainability, in optimizing energy.” He continues: “Our job is to capture those and make them scalable.”
For production managers, this mindset changes everything. Instead of waiting for the next “corporate project,” sites can test and prove their own ideas, and let IT handle the scaling.
The key is structure. Governance and IT should act as enablers, not gatekeepers. They make sure that local improvements can be rolled out safely, with data integrity, security and performance tracking in place.
Vedfald reminds us all; IT is a support function. The goal isn’t digitalization for its own sake, but rather using technology to make production better, cleaner and more reliable.
From talk to traction
If there’s one thread running through it all, it’s focus. Start with a business case that matters. Secure the data that proves it. Involve the people who make the product, not just the ones who make the slides. That’s how digital turns into better yield, steadier quality, and lower energy per ton. Big visions are fine, but real progress happens tank by tank and line by line.
For a medium-sized dairy, the next move is close to the floor. Pick one area where you can win quickly; intake, pasteurization or another process with good instrumentation. Map the signals. Pair them with lab results. Build a small decision aid the team actually wants to use.
And keep talking. Communication matters as much as technology. When people understand why a new tool is coming and what it helps them achieve, they make it work. Prove the gain, document it, share it, and move to the next line. That’s how digital becomes part of daily production.
