Have big box stores learned lessons that can be applied by electronics manufacturers?

One of the big takeaways from the Future Compute conference on the campus of MIT in May was a definitive “yes!”

There, we heard about how some of the large retail chains like Target use software, hardware and data in all kinds of customer experiences.

Almost every employee has handheld devices tracking the billions of sensors and cameras in use across some 1,900 stores and 50 regional distribution centers. At each store, it runs about 100 different software applications. They look at traffic trends: When is the peak? When is the lag? And how can they be modulated?

Now consider an electronics manufacturing operation. There could be hundreds of operators, thousands of PCBs, millions of components, billions of solder joints, each one needing traceability.

When making its purchases, Target sources from end-products all the way back to raw materials. It sources some of the cotton used in Target brand clothing directly from farms. In this era of instant outrage, every public relations debacle is just a leak and a tweet away. Say a farm used a verboten pesticide. Target might also need to know what was close to that source: nearby waterways, for example.

Likewise, PCB fabricators and assemblers buy everything from standalone process equipment to raw tin and copper. In our industry, we are looking into hard-to-reach places like the Democratic Republic of Congo, for instance, to trace the origin of certain metals and minerals.

And like Target, we use all sorts of software and systems to do so. A distributed ledger like blockchain could be a perfect tool for managing these transactions. It allows companies to track parts quickly and monitor traceability at a granular level.

The trick is to turn all that data into something useful, of course, and the cost simply to track and house all that data can overwhelm many companies. It goes a long way toward understanding why most supply chain decisions are reactive.

Where Target stumbles is how to manage its multiple cloud environments with local-level computing applications. The cloud is seen as too slow for responding to situations on the ground. Meanwhile, local networks have low latency, but lack the capacity to perform the immense analytical computations needed to adjust to changes in real time.

Edge computing, which loosely describes everything in-between, helps the processing of data gathered at local nodes. But, for now, it is seen as expensive, especially relative to the local sites, requiring a GPU at every store. Even a $100 billion entity like Target says it can’t afford it.

That’s where industrial manufacturing has an advantage. While we process millions or billions of pieces of data per day, we do so across much fewer production sites. Target has 2,000 sites to gear up, but even the largest of the Tier 1 EMS companies generally have fewer than 100 factories, and the NTI-100 PCB fabricators even less. They are in a better position to leverage edge solutions and – potentially – evolve into predictive supply chain management.

Another eye-opener at Tech Compute was the notion put forth that, for the greatest efficiency and implementation, artificial intelligence must be commoditized.

To do so, top-level companies need to coordinate a solution for others to follow. We see this to a degree in our industry, where Cadence, Zuken and others have been leading the effort to develop a comprehensive approach to capturing design intent and sending that downstream to fabrication, assembly and test, an effort known as IPC-2581. At PCB West (pcbwest.com) this year, they will join companies such as Koh Young, which will talk about leveraging the IPC-CFX standard (for Connected Factory Exchange) that establishes requirements for sharing information among the manufacturing processes and software systems used in PCB assembly.

As the amount of information that can be captured and shared grows, it begs for a holistic approach. What we see, however, is companies focused on the data often lose sight of the strategy. They are defining the solution – and customer – before they’ve defined the strategy.

This is where Target shines. The firm has spent two years of the pandemic under consumer siege. Today it’s a run on hand sanitizer. Tomorrow it’s toilet paper. It has to treat almost every day like “Black Friday.” Its use of blockchain to trade beef with supermarkets, for example, could be adopted by electronics manufacturers for component exchanges.

From traceability to data processing solutions, best practices might just come from outside our manufacturing environment. Don’t discount them. It may be the low-cost retail chain that has the answer to industrial manufacturing’s problems. 

• Prev
• Next