Johannesburg, 17 Sep 2019
Manufacturers are aware of the benefits their operations can gain from data. Few other sectors can boast the same planning and precision required to maintain manufacturing schedules. There is an appetite for concepts such as Industry 4.0, where connectivity, automation and robotics aim to vastly improve manufacturing performance.
It’s a journey of many steps, starting with software improvements such as Advanced Planning and Scheduling (APS) solutions to make planning faster and more flexible. But manufacturers don’t always see the value they expect from their data and consequently blame the technologies they invested in.
“The data story creates a lot of expectations,” says Lance Zikalala, MD of nCoded Solutions. “This creates the impression that once you have the right systems in place, you should start getting the results you expect. When the output doesn’t match those expectations, they think the technology has failed them.”
But if stakeholders only look at the output, they are missing a very important component of the data-manufacturing value chain.
Admin in the middle
The expectation is to capture data accurately, then deliver information through good analysis. Both are important and have encouraged investments from barcode scanners to visualisation dashboards. APS and ERP systems help connect many of these dots.
But what manufacturers often overlook is the importance of vetting and cleaning data. This is the domain of data administrators. Their jobs can seem very traditional, fussing over the condition of data as it enters storage then cleaned up before being used for analysis. It’s a role that doesn’t get much notice, but the boom of data is changing that.
“Capturing data is inherently messy,” Zikalala explains. “You can improve how data is captured, but there are more data points today to collect. This can cause more mistakes in the data, which ironically makes it harder to manage larger amounts of data better, such as with machine learning. If you can’t spot mistakes in your data, you can’t take advantage of it to modernise your operations.”
Data administrators help keep data clean. They interact with various systems to ensure data used for analysis is clean. Administrators work with benchmarks to define and identify abnormal behaviour. They strike a careful balance between automation and the current systems, further improving the prospects of modernisation gains.
Benchmarks can be used in different ways, such as using programmable logic boards at various points to capture and monitor data behaviour. But it requires an administrator in the mix to orchestrate the vetting strategy, also called the ‘staging’ area of the data value chain. If this staging area works effectively, it can identify problems such as bad capturing practices.
Know your business
What are the functions of data administrators? As mentioned above, they oversee the flow of data inside an organisation. This not only means maintaining databases but also developing the policies that control data. Data administrators don’t - or shouldn’t - operate in silos. Their work supports all other data activities. Then why are data administrators overlooked?
Zikalala offers two reasons: “When companies bring in a new technology that is supposed to do great things with their data, they think the technology will take care of it all. But these systems can’t interpret bad data as intuitively as people can. They are also too focused on the output and forget the fundamental rule: rubbish in, rubbish out. They don’t realise the place to clean out that rubbish is in the middle where the administrator would be.”
These are robbing manufacturers from valuable strategic insight. It is assumed that all the business insights sit with analysts who create information dashboards and reports. Yet those vetting the data have a much closer view of the action. There may be more noise upstream in the data pipeline, but the view is also bigger. Data administrators can be very valuable sources of insight on a manufacturer’s capacity and other factors.
Data administrators shouldn’t operate in silos but rather work with the rest of the business to develop data-related processes and workflows. This puts them close to many important areas inside the company, says Zikalala.
“The process of checking data and developing data rules requires interactions that give administrators better business knowledge. It shows them where the problems could be because that’s how the baselines are determined and tested. Vetting data is about understanding how the data behaves. That reflects how the business behaves.”
Manufacturers can benefit greatly from this. Already familiar with data’s benefits - from asset tracking to maintaining flexible schedules - they are eager to see returns from modern data-driven technologies. But technology is just a lever. It takes the focus and discipline of a data administrator to bring order to data across the manufacturing environment. That data is very telling about the business and, as a result, the data administrator becomes a valuable strategic resource.
If your data is not working for you, maybe you should start by asking: who is working with your data?
Share