Experts say that to build a demand-driven supply chain, manufacturers need more timely and accurate clues into customer demand, plus better ways to spot patterns in the data -- all in the service of more accurate sales forecasts. By propagating these “demand signals” back through the supply chain to the shop floor and beyond to suppliers, companies can make more of what customers really want -- and do it more efficiently and profitably.
That’s the theory, anyway. And while most of the heavy lifting is still focused on re-engineering business processes, deploying supply chain management software and fostering collaboration, a handful of emerging technologies are making it easier to capture and analyze demand signals and manage inventory.
Analysts point most often to recent advances in automatic identification and data capture (AIDC), especially item-level radio frequency identification (RFID) and smartphone-based social media and scanning applications for consumers, as well as new applications of supply chain analytics.
Item-level RFID focus moves upstream to warehouses
Cost and system-design issues have slowed RFID’s use on individual products. RFID tags are still used primarily on cartons, pallets, and high-value capital equipment, where their cost is easier to justify. But the past two years have seen major deployments of item-level RFID, and both advocates and analysts say it could be primed for a breakthrough.
“This is moving into large-scale adoption,” said Justin Patton, managing director at the Fayetteville-based RFID Research Center at the University of Arkansas. “In 2012, it’s going to be a lot more prevalent.”
Last year, Wal-Mart Stores Inc. required all manufacturers of jeans and men’s basics -- socks and underwear -- to apply item-level RFID tags, and it is expanding the mandate to new categories, Patton said. Other retail chains including Dillard’s, JC Penney, Macy’s and Target have followed suit with similar mandates for the same categories, as well as shoes and women’s intimates.
“Those are the categories where they have the most difficulty with replenishment and in-stock issues,” he said.
Item-level RFID makes sense for those lines because they can’t easily be reshelved, yet it’s important to keep a broad selection of sizes in stock. “It’s very difficult to tell from a cursory glance what’s out of stock because everything looks the same,” Patton said, and it’s not easy for retailers to substitute similar items, as they do in other categories.
Forced by retailers to add item-level RFID, manufacturers are looking for other ways to derive value from their investments, according to Patton. Now that the retailers are convinced of the technology’s value, the center is focused on investigating upstream supply chain applications.
“We think there are 40 to 60 different points in their supply chains where they can capture the data they’re getting from the RFID tags for their own benefit,” Patton said.
Preliminary results indicate that the most popular new application is in using item-level RFID to achieve 100% accuracy on audits. Audits are currently done at distribution centers through manual “pick-pack” operations that require opening a sampling of boxes to confirm their contents. Audits can help avoid expensive -- and sometimes questionable -- claims that retailers submit when they find shipments that don’t match their orders, but the current process is prone to human error and theft. RFID-based auditing could also improve the accuracy of electronic proof of delivery systems that shippers use, Patton said.
Yet barcodes are still the cheapest way to manage unopened boxes and will continue to be the preferred tagging mechanism at distribution centers, said Bert Moore, director of technical communications and media relations at AIM Inc., an association of RFID, barcode and mobile computing vendors that is based in Cranberry Township, Pa.
“When it goes on the sales floor, that’s the first great opportunity to read [RFID] tags,” Moore said. Inventory accuracy will improve, and theft can be minimized, he said, but data on actual sales will still come from barcodes scanned at the point-of-sale (POS) system.
Meanwhile, the Voluntary Interindustry Commerce Solutions Association (VICS) says it will release an item-level RFID roadmap and supply chain best practices that are being developed by a committee of manufacturers, RFID vendors, academics and major retailers, including early adopters like Wal-Mart.
Analytics for S&OP and social media
Manufacturers continue to look for new ways to apply business intelligence and analytics to their supply chain data. One area of focus is the sales and operations planning (S&OP) strategy and software that some companies use to synchronize forecasting between their marketing and production sides.
“You can’t do anything to become demand-driven unless you have the best possible forecast,” said Jeremy Friedler, senior consultant for strategy and operations at Maxiom Consulting Group, based in Waltham, Mass.
The pharmaceutical industry, for example, is interested in using supply chain analytics and analytics-enhanced S&OP to become more cost-effective as patents on branded drugs expire, and companies are forced to seek profits in lower-margin, generic drugs. By making supply chains more demand driven, “big pharma” can transition from traditional make-to-stock to more efficient, make-to-demand production, Friedler said.
Some manufacturers have started applying more advanced predictive analytics to demand signals and inventory data to improve their operational decision making, according to Lora Cecere, a partner at Altimeter Group, a research company based in San Mateo, Calif.
But Cecere said that while a few demand planning software vendors provide predictive analytics, most do not. “The main issue is that the traditional applications are not set up to be demand-driven, and people don’t want to throw away what they’ve built,” she said. “There’s a lot of momentum around ‘let’s just do what we’re doing and get it over with.’ ”
Cecere said natural-language processing is another advanced technology that is helping some companies sift actionable information out of the broad stream of supply chain data. For example, Conair, a maker of personal care appliances like hair dryers, uses it to spot inventory that might be out of compliance with retailer requirements, avoiding costly penalties, she said.
Natural language processing is also proving useful in social media sentiment analysis. “It’s really a listening device,” Cecere said. Levi Strauss & Co. uses the technology to glean demand signals from comments made to customer service departments and posted on social media sites.
“They actually put new products out and they asked people what they thought, asking them questions like ‘Should these jeans be blue or black?’ ” Cecere said.
Mobile shoppers generate real-time demand signals
Increasing use of smartphones for comparison shopping and location-aware marketing is also generating new demand signals for manufacturers. “Geolocation data is starting to redefine the whole shopping experience,” Cecere said. “It’s pretty exciting.”
Smartphones’ ability to function as barcode readers has led to new Web services that can scan regular barcodes, unique identifiers and product images to send shoppers directly to manufacturers’ websites, Moore said. Marketers like the services because they keep customers focused on the individual brand instead of sending them elsewhere. “It guarantees that you get the customer query,” he said.
Moore said the services can also make it easier for manufacturers to immediately download supporting documentation the customer needs to make the purchase decision, such as allergy warnings.
Cecere said the real-time nature of social media, geolocation and POS data is a major improvement over the four-to-six-week turnaround of the syndicated market-data services that manufacturers have traditionally used. The difference can be critical in keeping stock available to customers when they are looking for it, thus avoiding lost sales.
“If there’s not a good retail system, and you don’t see that signal, you’ll probably be out of stock for a long time,” Cecere said.
Learn why demand-driven MRP is becoming more popular