Earlier this month, Drishti teamed with A.T. Kearney to announce joint research on the gap in human analytics in the factory. We uncovered interesting statistics that show just how critical humans still are in manufacturing: 72 percent of tasks in the factory are still carried out by humans, and they create 71 percent of the value on the floor.

Despite the importance of humans, respondents lamented their inability to accurately measure the tasks they perform – leaving a person-shaped blind spot that manufacturers have no choice but to work around.

A few weeks ago, I traveled to Berlin for A.T. Kearney’s Digital Operations Executive Summit, where I spoke about these report findings alongside Michael Hu, an A.T. Kearney partner focused on digital transformation in the supply chain. After our presentation, a number of manufacturing leaders talked to me about their real-life frustrations with the lack of human factory analytics.

From my conversations, it’s clear that a lack of data is massive problem for manufacturing. And it’s equally clear that manufacturing leaders know exactly how they would apply greater data on tasks performed by humans. The conversations I had settled around the following themes:

  • Data can help improve productivity: For 110 years, time and motion studies have been the primary method of deriving data on human tasks, but they’re inexact and biased. The data these studies yield provides some level of insight into factory activities during a few discrete points in time; with that data, the industrial engineer performs the art of generalizing that data across weeks or months. He or she then uses that generalized data to try – emphasis on try – to balance lines, identify star performers or reduce overtime shifts. Manufacturing executives know the data has weaknesses, but they’ve had no clear alternative. That said, they also know exactly what they’d do if they had a better data source: They’d use those larger, more exact datasets to easily identify anomalies and specifically target productivity issues that otherwise remain undiagnosed.
  • Data can reduce defect rates: Humans create a lot of value on the factory floor, but they also introduce variability. Our survey found that 68 percent of defects were caused by humans, not machines. But humans are also still the best source of traceability on most lines: Operators perform scans during the cycle, which are good for tracing the bill of materials, but do little to trace back the bill of process. With granular historical data on what happening with every action at every station, manufacturers could quickly identify deviations from the standard sequence, speed root cause analysis and target pinpoint training customized to each operator’s needs to help them avoid quality issues that lead to costly rework.
  • Data can make your operations more transparent: Manufacturers know that recalls are expensive, time-consuming and potentially dangerous (to customers and the brand reputation). But no factory has the ability to travel back in time. Which means that when a recall situation arises, the manufacturer has little choice but to be extremely conservative. Manufacturers in this situation know they may be recalling 50,000 units even though only 50 are defective; the problem is that they lack the certainty to identify the 50 specific units that matter. This, again, shows that manufacturers know exactly what they’d do with process traceability: they’d quickly mitigate the risk of recall.

With any new technology that’s solving a problem never before addressed, questions about ROI can be difficult to quantify, but as my conversations in Berlin showed me, fairly easy to answer.

There’s a lot more data in our full report on the State of Human Factory Analytics.