How better data collection drives smarter manufacturing decisions

Modern manufacturers face enormous pressure to stay competitive. Accurate and efficient data is the key that can unlock better automation, quality control, safety, and decision-making to stay competitive in the increasingly digitized future.

In this episode of Manufacturing Matters, I spoke with Finn Hunneche, co-founder of Emendo Consulting and Factbird, to discuss how manufacturing can learn from human intelligence.

We talked about how factories need to be more like people and excel at sensing and taking actions. Finn likens this to “developing an immune system for a company.” In effect, fighting off problems when they come or at least raising the temperature so that you are aware of it. Enjoy the podcast!

Key Takeaways:

• High-quality data is crucial for making informed decisions in manufacturing.
• Human intelligence serves as a model for manufacturing intelligence: sensing and remembering, deciding, and acting.
• Experienced workers can help set control charts and data and learning can help retain this knowledge.
• Start digitizing holistic parameters in businesses in order to automate the normal, not normal analysis.

Introduction: How data powers smart manufacturing

In episode 13 of the Manufacturing Matters Podcast, we explored why improving data collection is critical for manufacturers and how we can make machines do better.

As Finn emphasized, manufacturers should use data to sense what is going on in operations at all times, much like humans rely on their senses to navigate the world. Inaccurate or incomplete data can cause significant disruptions, from poor quality control to machine failures. Finn explains that manufacturers must focus on collecting better, more reliable data to thrive in the competitive future of manufacturing.

The connection between human and manufacturing intelligence

A key theme of the episode is how manufacturing intelligence can mimic human intelligence.

Finn explained that human intelligence involves four main components:

1. Sensing and remembering
2. Deciding
3. Acting

These same principles apply to manufacturing systems. Just as humans need accurate sensory information to make informed decisions, machines rely on data inputs to guide their actions.

Finn explained that humans use approximately 70% of their brain capacity for sensing whether things are normal or not. This comes from our evolution and the primal need to quickly understand whether we are in danger, like a tiger, or not.

Then, based on the sensing and our analysis, we decide on what action to take:

1. No action – We don’t do anything.
2. Corrective action – We decide to change something.
3. Sustained action – We make the corrective action habitual.
4. Explorative actions – We investigate further. We are programmed to explore as humans, and this seems to have helped us to evolve.

So, breaking it down, human intelligence is sensing and acting. And smart people are really good at this.

In the same way, manufacturers need to ensure they prioritize collecting high-quality data that reflects the true state of their operations so that they can make better decisions.

Without reliable data, the entire system becomes prone to errors and inefficiencies.

But, as Finn points out, even if you have all the data in the world, it’s not worth anything if you don’t make good decisions with it.

The role of data in digitization

So you can see the point that Finn is driving at: at the heart of smarter manufacturing is data collection. Whether it's automating a process or improving quality control, having accurate data is essential.

Data is at the core of digitization. A common theme for German Chancellor Angela Merkel was that you need to digitize or become history. She was mostly talking about the staid German bureaucracy and their love of paper documentation (believe me, I lived through it), but she also discussed the need for this to happen in business as well. In her 2018 Davos speech, she stated, “For the next four years, our job will therefore be to bring digital technology to our education system and bureaucracy, to give the public the option of communicating with their state via digital technology in the digital age, and to create a better ecosystem for start-ups so that we remain a good place for innovation." Interestingly, in the same speech, she said, “Data is the raw material of the 21st century.”

She was not wrong. The innovations in automation and machine learning have pressed forward at breakneck pace. If you are not on the digitization journey now, then you will find it difficult to keep up.

On the whole, the more digitized your factory is, the more things you will be able to sense. It’s not only about manufacturing data, it’s data about span of control, quality data, and material data. According to Finn, companies that can do this will perform better, and “instead of having a review every three months, they will have a review every day. And they will outcompete the other companies.”

Keep in mind that the challenge is not just collecting data, but also ensuring its integrity and reliability.

Things like establishing data processing agreements, automation, sticking to SOPs, and continual learning within organizations are helpful for improving data quality. You want to ensure that the data collected is consistent and reliable, which is crucial for avoiding missteps in decision-making. Without such a focus on data quality, manufacturers risk making decisions based on faulty or incomplete information, leading to poor outcomes.

Experience and learning from data

Much like human intelligence benefits from experience, manufacturing systems must leverage experience and historical data to improve decision-making.

Finn pointed out that capturing organizational knowledge is essential, especially as experienced workers retire or leave. These experienced colleagues know whether or not things are normal from their history and senses. In order to become more robust as organizations, “we need to digitize; we need to build systems where we’ve built sensing into the computer systems.”

To help get us there, we can use control charts, for example, with upper and lower limits of where things should be. The experienced colleagues can help say where things should be, capturing their knowledge, and then we develop and improve from there.

Using the analogy of how humans work, he points out that we don’t typically think about our feet. We put our shoes on, and our feet do their thing throughout the day without us worrying about them. But as soon as something is wrong, like a stone in the shoe, we become very aware of it.

Finn wraps up his point by saying, “We need to start digitizing the holistic parameters in the businesses in order to automate the normal, not normal (analysis), and capture the experience of what is normal inside systems rather than relying on people to stay in their jobs for a long time.”

How to implement smarter data collection in manufacturing

So how can manufacturers begin implementing better data collection systems? According to Finn, the first step is to start by asking yourself, “Was it a good day yesterday?” And in order to do that, you have to know what a good day is.

Then you can go from asking questions about machines to asking the same questions about the people working in the factory. You have a quality process, planning process, and maintenance process that you will also want to account for.

His other piece of advice is to just start. It’s not just about performance management and driving things forward. It starts with just detecting whether or not things are normal, so just start gettig the numbers. As he said in an earlier podcast, "It’s impossible to survive in the manufacturing environment unless you have numbers."

From here, a future step is to look at how to integrate artificial intelligence (AI) and machine learning (ML) systems into your operations. These advanced technologies can help automate decision-making by allowing machines to process data in real-time and respond to changes without human intervention.

Better sensing for smarter manufacturing

Finn’s insights make it clear that smarter manufacturing begins with better data collection.

If factories were more like cars, so self-enclosed systems sensing thousands of parameters and automatically adjusting and alerting when things are not normal, then meeting goals for output, efficiency, safety, and quality would become a lot easier.

By improving the collection, quality, and reliability of data, manufacturers can make more informed decisions and stay competitive in an increasingly digitized world. Manufacturing systems must get better at sensing, understanding, and acting, much like humans do. By building a strong foundation of smarter data collection, your business can unlock new opportunities for growth and improvement in the years ahead.

For more insights and practical tips on continuous improvement, lean manufacturing, and other industry topics, stay tuned to Manufacturing Matters. If you have any questions or topics you’d like us to cover, feel free to reach out to us via LinkedIn or email.

And remember to subscribe to Manufacturing Matters to get notified when the next episode is available on Spotify, Apple Podcasts, YouTube, Amazon Music, and other popular podcast directories.