To run a responsive, demand-driven supply chain, you have to get a handle on all the market data. But it can feel like drinking from a fire hose.
Some companies have built a demand signal repository (DSR), a central database that integrates, stores and normalizes terabytes of "downstream" data that, when properly analyzed, can lead to better forecasts, lower inventories and higher top-line revenue.
Because they provide daily tracking of goods at the most "granular" levels -- individual SKUs and stores -- DSRs are catching on with consumer products (CP) manufacturers in such industries as food and beverage and consumer electronics. According to AMR Research, 94% of companies with $5 billion or more in annual revenue are in some stage of implementation. In fact, DSRs are the fastest-growing category in enterprise software, with sales rising 49% in the past year, according to Steve Banker, senior director for supply chain management at ARC Advisory Group -- though admittedly they started from a small base.
Building a business case for a demand signal repository
Before technology made real-time visibility feasible, manufacturers focused on assembling the perfect order.
"Now the thinking is that the supply chain extends all the way to the retail shelf," Banker said.
Companies that have deployed a DSR know that the freshest, most detailed demand signals come from point of sale (POS) systems that record store inventory and customer buying patterns. The typical DSR also tracks inventory in other segments of the supply chain, as well as syndicated data from The Nielsen Company and Information Resources Inc. (IRI). The goal is to make better decisions in category management, vendor-managed inventory (VMI), promotions, and other key drivers of profitability.
"You get a total downstream look at inventory and sell-through," Banker said.
With sales in the billions, major CP companies need only see single-digit-percentage reductions in shelf out-of-stocks to pay for a DSR in months, he said. Other business-case pluses include better demand forecasts and retail service. A DSR can also help identify local price elasticity and avoid over- and under-pricing.
Enterprise database heavyweights Oracle and Teradata often provide the core database, but the real value comes in analytical applications that run on top. Several vendors specialize in them. For example, Terra Technology's demand-sensing software tries to identify the DSR data that is best at predicting trends in specific areas. Last summer, Procter & Gamble began using Terra's Multi-Enterprise Demand Sensing with its DSR and reported inventory savings of $100 million, Banker said.
Other vendors offering core DSR or analytics include i2 Technologies, Relational Solutions, Retail Solutions (formerly T3Ci), Shiloh Software, Vendor Managed Technologies, and Vision Chain.
You can't just throw this on a server somewhere and walk away.
Manufacturing vice president for global industry solutions. Teradata
"The analytics are really focused on looking at sales trends, out of stocks, inventory performance, and other operational issues at the store level," said Matt Johnson, senior director of consumer sector product strategy at Oracle, which also sells the analytics piece. "Over time, the architecture of a DSR becomes the architecture of the next generation of applications."
Some companies write their own analytics, often with outside help. Last year, The Clorox Company built an Oracle DSR with consultant/integrator Frontline Consulting Services (FCS). An application called Trade Answers correlates Siebel promotion information, syndicated material from IRI, store POS data, shipment data from the logistics system, and financials to measure the effectiveness of promotions, according to a presentation at Oracle OpenWorld.
In turn, FCS helped one large client that wanted a handle on the profitability of discounts that were causing significant lifts in sales, said Sumit Mehra, who is the technical director for business intelligence solutions at FCS. It was very hard to figure out which promotion worked very well, Mehra said, as it was very time-consuming to collect all the data and get it into meaningful form.
FCS wrote an Oracle DSR touch-screen application that lets executives analyze promotions by region. Besides providing comparisons of past sales, the software lets them see current shipments by warehouse and even drill down to individual SKUs for certain colors and sizes.
"You can predict when this particular SKU will run out in this particular warehouse," Mehra said. "This is all real-time."
In turn, by identifying relationships between promotional tactics and sales, managers can plan future promotions.
A number of vendors, including Oracle, sell Software as a Service (SaaS) versions that can make it easier to prove a DSR's value before investing in behind-the-firewall infrastructure. For example, One Network specializes in SaaS supply chain solutions, including DSRs, and FCS makes its DSR available online at RetailSignal.com.
Common challenges in DSR projects
Terra CEO Rob Byrne likens DSRs' birth pains to the 1990s heyday of ERP projects, when companies made huge investments just to get a better view of their data.
"It's sort of one of those painful, necessary first steps" -- not to mention complex and expensive, he said. "There hasn't been a compelling value story. I think it will get easier, and it will get cheaper."
A common mistake was to assign DSR ownership to sales teams, who lacked the clout to effect the business-process changes that held the most promise. Maintenance hassles were another reason why first-generation DSRs disappointed some companies, Johnson said. Their desire now to build a true, enterprise-wide DSR is behind much of the recent interest, he said. The benefits could go beyond short-term tactical issues to strategic ones, such as product lifecycle management.
Clorox cited data governance, the quality and alignment of external and internal data, and the importance of using syndicated software for user interaction as its take-away lessons.
Organizations should be prepared to spend most of their time integrating, harmonizing and maintaining downstream data so it doesn't create errors in the analytical software, analysts said. MDM and data cleansing will be ongoing challenges as products, stores and manufacturers are dropped or added.
It helps that DSRs come with prebuilt interfaces to popular data sources, such as Nielsen and Wal-Mart's POS system.
"The DSR translates the customer's view to the manufacturer's view," Johnson said.
These experts recommend first rolling out a DSR to your biggest retail customers, then expanding it to smaller ones for economies of scale. Banker recommends starting with a pilot for a single product and retailer.
Above all, the support of C-level executives is crucial to drive the multi-year project and mandate the cross-functional use that will bring out the repository's real value. Shoot for the quickest-ROI gains in forecasting and inventory in the first year. Assign certified project managers to the job, construct a robust IT and analytical infrastructure, and be prepared to maintain it, said Keith Henry, Teradata's manufacturing vice president for global industry solutions. It will take three to four months to set up the first customer.
"You can't just throw this on a server somewhere and walk away," he said.
About the author: Freelancer David Essex has covered information technology for BYTE, Computerworld, PC World, and numerous other publications and websites.