The manufacturing world is drowning in data. A recent study published by The Economist1 evaluates the current state of manufacturing data. It finds that manufacturers have significantly ramped-up shop floor data collection in recent years. The effort has paid off with significant reductions in the cost of quality, but at a price.
Here are some of the study’s key findings:
- The vast majority (86%) reported having problems with managing the data they were generating.
- The good news is that the data that is being used is generating very positive results. Two-thirds reported savings of at least 10% on their total cost of quality.
- Only 42% of manufacturers reported having what they consider to be a well-defined data-management strategy.
- 35% reported difficulty integrating data from multiple sources and formats.
- Only 40% said they used data analytics to find solutions to production problems.
- 62% said they are not sure they can keep up with the large volume of data they can collect.
This study is consistent with anecdotal estimates from the field that 70% of the data that is collected in manufacturing companies is never used.
This is going to get a lot worse in the near future with the explosion of data collection on the Internet of Things (IoT, or IoE – the Internet of Everything). One study by IC Insights2 forecasts a 21% Compound Annual Growth Rate in investments in the IoT in the five-year period beginning in 2013.
What are we going to do with all that data? How do we keep from becoming completely inundated with irrelevant and unusable data?
These six questions will help you connect data to your business purpose and provide you with actionable intelligence.
1. What is my aim?
At the risk of oversimplifying, manufacturers seem to divide themselves into two groups. One group collects data as a defense shield. The other transcends that function and drives continuous improvement. While the first is content with a protectionist stance, the second group is never content. They are always looking for ways to make it better.
If you’re only concerned about protection, then collecting and having data is enough.
If you’re concerned about continuous improvement, then the idea that 70% of the data you collect is unusable should grate like fingernails on a chalkboard. Making data visible and actionable is absolutely key to your success.
2. How does my aim support our business strategy?
Take a look at your organization’s strategic initiatives. There are probably a few things that your senior leaders and board believe must be accomplished in the next three to five years. Ask yourself, how does my aim support that strategy?
If you can’t draw a direct connection between your aim and at least one of those strategic initiatives, you need to rethink your aim.
If it does connect, can you explain the connection in 15 words or less in a way that anyone—your mother, your uncle, your teenage nephew, your boss, your boss’s boss—can understand? The goal is to reach the simplicity that lies beyond complexity.
3. Does my organization have a culture that is ready to make good use of data?
My guess is that if you’re reading this article, the chances are good that you have some work to do on this front. Becoming a data-driven company doesn’t just happen, and when it does, it doesn’t happen overnight.
This is like any other change initiative: Start very, very small. Don’t make a big splash or program out of it. Put at least one believer and one naysayer on your team, and accept as gift what anyone brings to the table.
Look for early wins. When they happen, celebrate and share them with anyone who will listen. Then slowly build on them.
What may be unique about a change initiative that involves data is that technology can definitely make a difference in your organization’s readiness to take advantage of the data. In one recent study, 67% of data scientists say they don’t have time to do meaningful analysis.3
Their biggest challenge?
They’re spending too much time cleaning data. There are tools out there that ensure you have a ready supply of clean actionable data, and I don’t mean limiting yourself to Excel.
Technology can’t change culture, but when the culture is ripe it can help.
4. What should I collect?
All too often we look at what we can collect instead of what we should collect. Don’t get caught in that trap.
Take the time to identify (through rigorous statistical analysis if possible) the key metrics that matter. Stop collecting the other stuff just because you can. Certainly one of the key reasons 70% of the data that is collected is never used is because it is useless. Stop.
Having said that, I know there are times when you need to log data for bona fide legal reasons or to have available “just in case.” I suppose that is unavoidable. My recommendation is to weigh this one carefully and not throw something into a database just because it is easy to do.
And as we come to the fifth question, remember this famous quote from Peter Drucker: “Nothing is more wasteful than making more efficient what should not be done at all.”
5. How can I optimize the cost of this info?
One of the reasons we’re seeing such an explosion of data is because there are some great technologies that are readily available at a low cost. Manufacturing—especially the quality function in manufacturing—has been leading this curve for years.
Quality people have been connecting digital gages, weigh scales, and other devices to computers for years. Those connections have reduced or eliminated the recurring costs of data collection.
Take advantage of new technologies such as MtConnect and XML that further the capacity to automate data collection and remove the costly human element from the equation.
6. Can I contextualize this data?
Data in your manufacturing company operates at different levels. At the highest level, we have information about how the business is performing. At a medium level, we have information about orders and jobs. At a low level, we have information about specific products. At the lowest level we’re looking at equipment performance.
For the most part, data at the lower levels doesn’t know much about the anything at the higher levels. The opposite is also true: data in the upper levels doesn’t know much about data in the lower levels.
If we can build bridges between these levels—especially by providing context for data at the lower levels—you can make the data at the lower levels much more valuable. An example of this would be to link material lot and supplier data to machine performance so that you make more informed decisions about suppliers based on the performance of their materials in your equipment.
The data wave is building. Chances are you’re already swamped with data. By paying attention to these six questions you’ll be better able to maximize the business benefit you get from that data.
You’re part of a supply chain. Use the data to better understand the links across the supply network.
1 “Manufacturing and the data conundrum - Too much? Too little? Or just right?” http://www.economistinsights.com/sites/default/files/Manufacturing_Data_Conundrum_Jul14.pdf (Accessed on 24 March 2015.)
2 http://www.icinsights.com/news/bulletins/Internet-Of-Things-Boosts-Embedded-Systems-Growth/ (Accessed on 24 March 2015.)