Would you share your data with another manufacturer?
Data sharing in manufacturing is, frankly, a radical proposition. And yet it’s one that has the potential to transform how we create value.
In a recent white paper, the World Economic Forum, with the Boston Consulting Group, outlined what the future of manufacturing could look like with data sharing.
Here are the 10 things you need to know.
10 Things you Need to Know About Data Sharing
1.) What is Data Sharing?
Data sharing is simply when manufacturers (selectively) make their data available to other companies and organizations.
Data sharing can take place between manufacturers in,
- The same industry (e.g. pharmaceutical manufacturers sharing best practices)
- Across processes (e.g. machinists sharing failure data for a particular asset)
- Amongst all of the companies participating in a common value chain.
2.) There are more use-cases than you might suspect
Some use-cases for data sharing come quickly to mind.
For example, pooling visual quality data to create a more robust training set for computer vision algorithms. Others are perhaps less obvious.
The WEF report outlined 5 core areas for data sharing.
- Enhance asset optimization – Better data to help machine learning and AI improve uptime, efficiency, and quality
- Track products along the value chain – a more robust understanding of where, when, and why
- Track conditions along a value chain – for better visibility and quality, as well as simplified compliance and reporting
- Exchange product characteristics – give life to digital twins
- Verify provenance – Collaborate for full supply chain visibility.
3.) Applications are the foundation for data sharing
There are many challenges to data sharing. Among them, interoperability and data collection rank high.
Therefore, the report argues that applications are the foundation for data sharing. Applications collect good data for a circumscribed process. This level of detail and scope is essential when it comes to harmonizing across companies.
4.) The biggest 4IR Advances Require Data Sharing
At this point, you’ve probably heard the hype about predictive maintenance and autonomous control.
The fact is, none of the most exciting advances of Industry 4.0 are possible without truly colossal data sets–data sets that no manufacturer can amass alone.
In sharing data, manufacturers can create a more robust record of events and processes. With this extra data, there’s more grist for advanced algorithms to work with.
5.) Sharing doesn’t mean peer-to-peer
There was one point the report was quick to clarify: sharing data does not mean exchanging peer-to-peer. The security risks are too high, competitive advantage too precious, and trust too low.
Rather, the analysts outline a model in which 3rd parties act as organizers and intermediaries, ensuring a safe and confidential exchange. These intermediaries can be large IT firm, OEMs, or forward-thinking data startups.
No matter the broker, having a third party to facilitate is crucial.
6.) Trust and technical capacity are the biggest barriers
Could you have guessed this? While the potential of data sharing is transformative, trust and expertise remain barriers.
For one, manufacturers are reticent that existing services are safe enough to ensure security and compliance. And many organizations aren’t staffed with the data scientists and engineers necessary to participate in a data sharing program.
The report authors, however, expect this to change as large manufacturers realize value from data sharing.
7.) Paper processes don’t cut it anymore
This might be obvious. But sharing a warehouse of paper reports won’t do anyone any good.
The report suggests that an enduring hindrance to data sharing is the industry’s slow embrace of full digitalization of paper-based processes.
8.) Neither do information silos
Manufacturing, more than many industries, suffers from data silos. Implementing a true data sharing program requires horizontal, as well as vertical, integration of data streams.
9.) Data sharing is a business problem, too
Of the manufacturers surveyed for the report, many stated that determining business value was a limiting factor.
Given the resources involved, and difficulty quantifying data-sharing improvements prior to the program, many organizations struggled to 1.) make the case that the the program would be worth it, 2.) accurately scope the resources required to execute, 3.) anticipate what the value of the program would be.
All of these are fundamentally business matters, and yet they’re crucial to designing a data sharing program that creates value.
10.) The expected value is real
The authors of the report estimate a lower bound for data sharing value to be $100 billion. Roughly 80% of this value is expected to come from improved asset performance and supply chain visibility.
Given that these are areas that all manufacturers could stand to improve, I want to end with a question:
How could data sharing impact your operations?