Several years ago, I was working in a large dairy manufacturer’s control room, helping implement our product. We were observing the manufacturing process running, when they had an unexpected stoppage on a production critical asset. While several employees rushed to address the issue on the shop floor, I noticed a shift supervisor copying down some notes in a small notebook. I asked her what she was writing down. She explained that she tried to keep track of a specific in-line quality reading at the point in time when this particular asset failed. She was performing a manual and basic form of data analytics and continuous improvement.
Looking back on this simple interaction, I realized this moment told a compelling story. The shift supervisor was improving and transforming her job skills, creating a mini-database and mental models for analyzing correlations between quality readings and asset failures. This capability is not in the job description of shift supervisors. But the best ones naturally gravitate towards solving problems. This example of workforce transformation and self-taught skill generation allowed her to take on new responsibilities because she wanted to improve her plant’s performance. Giving her better data tools and making data accessible would leverage her existing curiosity and make it a superpower, I realized at that moment.
Manufacturing Workforce Transformation with Data Democratization
With all digital transformation journeys in manufacturing, there are expected and unexpected benefits. Most of the time, benefits are projected in terms of KPIs. Those should lead to better productivity, availability, quality, and improved EBITDA.
Having spent the last three years implementing Sight Machine at the factory level, I’ve come to see the journey through a different lens: how ways of working and problem-solving are transformed when we democratize manufacturing data. Doing this changes the game and makes it easy for shift supervisors like the one I met to solve problems, test hypotheses, and drive incremental continuous improvement. In turn, this can add up to real outcomes and benefits in productivity, availability, quality, and sustainability that are unplanned but incredibly valuable.
Having watched numerous plant teams adopt data-first and data-driven approaches to optimizing production processes, I’ve noticed emerging patterns in how their approach to work changed as they became more digital and democratized manufacturing data. Specifically, teams that achieve financial benefit and measurable outcomes from Manufacturing Workforce Transformation frequently exhibit two significant changes in their approach to work:
- As data becomes more accessible, more team members begin making data-driven decisions in real-time.
- When the bulk of data engineering is automated, useful data analysis happens both more often and closer to real-time.
More team members begin making data-driven decisions, in real-time.
To expand on the story about the shift supervisor at the dairy plant, I wanted to know why she took those notes and how Sight Machine could help with manufacturing workforce transformation. The shift supervisor showed me a running tally she was keeping of viscosity measurements at the exact moment this machine failed. She had a hypothesis that stoppages on this particular machine was correlated with the product’s viscosity. The shift supervisor’s reason for wanting the data made sense. What struck me as odd was the need to note the information manually. Why would she do this when she is literally surrounded by screens displaying data from a set of machines measured with robust, accurate sensors? The information she wanted was available digitally. So why write it down?
Once I asked that question, the answer was obvious. Getting access to production data in an understandable format required a lot of effort. At her plant, historical viscosity data was captured in a separate table from the downtime data. To correlate the viscosity against downtime events, she’d have to manually align the downtime records with the viscosity time-series data in a spreadsheet. Furthermore, she did not know SQL, so she would need to ask another employee to pull the data for her.
From that perspective, it does seem easier to write down the value every time the machine stops. The problem is, she was left essentially eyeballing a table of data from only shifts that she worked, when she was there to see the machine go down. What she really wanted was data comparing viscosity and downtime in an easy to navigate web interface. Easy access to this data would let her quickly ask and answer the question; Is there a trend in my viscosity data ahead of downtime events on this machine? Better still, she could run this correlation on years-worth of data from all shifts to understand whether her observations are outliers or represent a true pattern. She could filter the correlation inputs by a particular product type or material supplier. She could do all of this without writing a single line of code or performing any manual calculations.
As this shift supervisor became a power user of the Sight Machine platform, I watched as her way of working changed. When she came up with a hypothesis based on her observations, she tested it right there in the control room. No asking for help pulling reports. No tedious spreadsheet manipulation to align data sets. No SQL queries. The only limitation on what she could test was what was data was tracked and collected.
This is a trend of subtle but powerful workforce transformation I’ve seen at all facilities undergoing successful digital transformation: More roles within a factory organization move to using data-driven hypothesis testing. This shift makes the entire organization smarter, more capable and more resilient. It also transforms the jobs of ambitious shift supervisors and other curious first-line workers, making their roles more interesting and fulfilling.
High-value analysis happens more often and closer to real-time.
As data becomes not only more accessible, but also contextualized in real-time, employees conducting analysis spend more of their time doing value-added work to improve operations. Compared to our shift supervisor, process engineers tend to have more skills in pulling and blending data. Still, the challenge of procuring a data set built for analysis is time-consuming and inefficient. Data democratization and tools like Sight Machine make it much easier to look for relationships amongst data from different sources and reduces the need for time-consuming data engineering (AKA wrangling). For example, good data democratization platforms can automatically map a set of streaming viscosity data with specific temperature readings taken at discrete intervals so that they are easy to compare and correlate. This is true not only for dairy and food manufacturing but also for chemical manufacturing and pharmaceutical manufacturing, as well.
Thus, when production data is streamed in real-time and automatically engineered and contextualized, the work of blending multiple sources together is also handled automatically. This allows process engineers to spend less time wrangling data and more time analyzing it. This, in turn, drives new behaviors in several ways. The volume of questions a process engineer can ask of their data increases. Process engineers can spend much more of their time looking for answers rather than wrangling data. Their job becomes more detective work and less grunt work – and more creative and analytical. Perhaps more importantly, any analysis they build can be easily updated and presented on real-time data, without additional work from the engineer. That analysis can be shared and improved on by shift supervisors, process engineers, or even data scientists.
Moving From Three-Ring Binders to Constantly Updated Data Analysis
In many plants we visit, we see the results of analysis contained in three-ring binders in the control room, circulated in staff meetings, or shared via email. “Best practices” are arrived at via analysis of production data that may be 30 or even 90 days old. That analysis is then passed to operations teams as an execution playbook.
Because those best practices are derived from data that becomes increasingly stale, the most groundbreaking finding could change in subtle or dramatic ways as conditions in the plant change. To update these best practices, engineers have to maintain and revamp their analysis with new data. Without an easy way to acquire, aggregate, engineer, and normalize data, this step requires additional effort and distracts from further advancing the frontier of continuous improvement.
We have seen when engineers can conduct their analysis in a digital platform like Sight Machine, they quickly adapt to saving their process analysis and periodically checking the automatic updates as new data becomes available. In this way, changes to what an analysis might recommend for a setpoint on a machine or a “golden run” for a production line become immediately apparent. Ripple effects of changes upstream and downstream – both in the plant and in the supply chain – can be automatically taken into account. This constant updating creates best practices and manufacturing recipes that live on in perpetuity, updating recommendations based on the latest data.
Conclusion: Manufacturing Workforce Transformation Makes Better, Happier Employees
In my experience, these types advances are precursors to significant improvements in an organization’s bottom line. While that is ultimately the most important benefit of digital transformation, it is not the only one. Self-directed workforce transformation for manufacturers may be just as important.
As more roles have access to data and spend more time using data to drive improvement, their positions create more value for the employer and for the individual. I can say without hesitation that this is my favorite part of bringing about digital transformation – watching individuals seize the opportunity to expand their usual scope of work and impact on their companies. Similarly, as analysis becomes more efficient and more time is spent actually tackling hard problems (not merely preparing the data to do so), I’ve thoroughly enjoyed seeing individuals’ job satisfaction rise.
So next time you’re thinking about the benefits of a digital transformation project, please consider the added benefits beyond financial ROI. The operator who can now conduct correlation analysis. The process engineer who can focus on problem-solving rather than data wrangling. The curious supervisor who has a hunch and wants to see if the data agrees with their idea. Successful Manufacturing Workforce Transformation will yield financial improvements and improve the productivity of manufacturing companies. These transformations will also yield valuable improvements in the way people work and how much they enjoy their job.