Manufacturing Data Is Worthless Without This…
Sort of a salacious title, right? It’s true. Manufacturing data is worthless without one simple thing.
Every year I watch manufacturers ignore important issues and spend hundreds of thousands of dollars trying to solve manufacturing challenges using data, but still, completely miss this important component.
What’s more, this data is responsible for influencing downtime, throughput, scrap rates, scheduling, and more. If manufacturers miss the mark on this, I’m afraid that all of these things are affected as well.
What am I even talking about?
Manufacturing data in context is everything
I know that might not seem like a sexy answer at first.
It might also seem like some kind of self-serving sales proposition meant to highlight SensrTrx. But, it’s not.
Context is fundamentally the most important thing for manufacturers to understand about their machine data and manufacturing processes. It will ultimately determine whether or not something on the factory floor is fixed, broken, or actively costing money as I type this.
Any software, solutions, and services that help manufacturers do this aside… Below, I will show you exactly what kind of context is important. It doesn’t matter how you are able to achieve this as long as you do it.
First, a classic sports analogy
Before your eyes glaze over, let me show you why context matters… a lot.
Quarterback #1 in rookie year
- First pick in the draft
- Completion percentage = 63%
- Interception percentage = 3.3%
- Avg. yards per completed pass = 8.6
- Avg. yards per game = 233
Quarterback #2 in rookie year
- First pick in the draft
- Completion percentage = 56%
- Interception percentage = 4.9%
- Avg. yards per completed pass = 6.6
- Avg. yards per game = 164
Both quarterbacks won roughly the same number of games this year as well.
With this data, it’s safe to assume that Quarterback #1 is outperforming Quarterback #2.
We all can probably guess where is headed.
If I asked you to choose which you’d pick based on this data, you’d know I was setting you up for a gotcha moment.
You’d be right.
However, If I told you that prior to selecting which Quarterback you wanted I would provide any additional information you wanted, you would probably ask me to supply a lot more data.
Getting more data about those quarterback’s first seasons probably isn’t going to help you understand which quarterback is the best choice; as we know Quarterback #1 looks to have had all-around better stats in their rookie year.
You would probably want to know a little historical background, but you’d really want to know what happened next. How did their careers pan out? What were stats like in years 2,3, and 4? What about Pro Bowl selections?
You get it. I could go on forever.
You need contextual information.
It is super easy to be fooled when all you are looking at is data on its own.
You need context otherwise people and machines can trick you.
Quarterback #1 was Tim Couch and Quarterback #2 was Peyton Manning.
(wouldn’t it have been frustrating if I would have left that out).
Why manufacturers have data out of context
A lot of manufacturers are missing this context (more than you might think).
This is partly due to collection methods and partly due to how the data is stored and used on traditional manufacturing shop floors.
Often data is manually collected or collected from individual machines and stored in multiple systems. This usually means that machine and performance data are looked at independent of overall operations.
It can be really hard to marry all of these snapshots of data together into something actionable or meaningful.
What’s more, even if this data is organized together somehow, most manufacturers only look at it retroactively — in meetings or reviews. This means all the opportunities that may have existed to fix something, or prevent a costly issue, have already passed.
The machine downtime example
Imagine you have a particular machine cell with higher than average downtime.
Here’s the data you would want to be able to properly solve this problem:
- How does it compare to other cells?
- When is it occurring?
- When did this start?
- What is the cost?
- Who might know what is happening?
- Have we seen this issue before?
- Other information specific to your manufacturing business processes…
This is all about looking at this problem in context. Understanding history, evaluating trends, and tapping into all the sources of knowledge of an issue is critical to diagnosing it.
However, diagnosis is just one step.
Imagine you discover it is the 2nd shift causing all the problems. Perhaps it is understaffed. That cell may be having longer changeovers during the 2nd shift due to staffing issues.
You could easily do the math to determine the cost of solving this problem. You could also highlight the money saved from fixing it long term.
This allows you to solve the problem and justify the costs of solving it.
But more importantly, you can now monitor it into the future understanding if the problem is actually fixed.
Monitor, Diagnose, Solve, Monitor
These 4 steps are how you need to be thinking about manufacturing data contextualization.
Ask yourself these questions about your manufacturing business.
- Can we monitor our machines and processes in real-time?
- Can we understand problems in the context of history and trends?
- Is someone able to actually diagnose and understand these issues as they happen?
- Can we calculate what these things are costing us?
- Can we use this data to solve the problems we identify before they cost us money?
- Once solved, can we monitor this moving forward to ensure it is solved?
- Can we measure the ROI of solving this problem?
If the answer is “No” to any of these questions, you can work backward from there to understand what gaps exist in your ability to understand manufacturing data in context.
If you want to truly affect downtime, performance, throughput, and other manufacturing KPIs, this is the information you need. You need to be able to answer all of these questions.
Need some help? SensrTrx does is it all easily, affordable, and automatically.