Cloudy With A Chance Of Out-Of-Stocks?

To ensure sunny weather ahead — and avoid out-of-stocks and lost sales or excess inventory that must be cleared out — retailers need the right forecasting strategies and tools.

Imagine the chaos that would ensue if you were not even able to hazard an educated guess related to consumer demand for the products your stores carry. Supply and demand likely would fall dangerously out of balance, leading to out-of-stocks, spoiled surplus perishables and other costly problems.

Today's retailers, fortunately, have access to technologies and tools to assist them in the forecasting and demand management process. That said, the process remains far from perfect for most of them.

If all goes well along the way, they end up buying what's selling, says Malcolm Buxton, CEO of Newport Beach, Calif.-based JustEnough Software Corp. And if it doesn't? Well … then they end up having to sell what they've bought.

"It's a very subtle difference," Buxton explains. Brian Kilcourse, managing partner of Miami-headquartered RSR Research LLC, says today's retailers face a "basic conundrum" in the forecasting arena — that's why his firm recently launched its first-ever demand forecasting survey to study the issue in detail.

"They seek to do four things that are hard to keep in balance: localize the assortment, speed up the order-to-delivery cycle, improve service levels and decrease their overall investments in inventory, " he says. "Demand forecasting has come to the forefront of processes that have to improve to achieve these seemingly mutually exclusive goals."

Retailers — especially those in the fast-moving consumer goods segment — want to be able to respond more quickly to changes in demand signals, he adds, without having to overdo it on inventory side.

Other issues Kilcourse points to include information lags or holes within the supply chain, sales or marketing promotion areas; aggregated data that prevent a more granular forecast; inconsistent or nonexistent in-process forecast performance metrics; nonintegrated multiple demand signals in planning and logistics; poor understanding of customer behavior according to channels; difficulty in capturing cross-channel events that impact customer behavior and channel demand; a "throw-it-over-the-wall" mentality across assortment, price, promotions, space and replenishment planning; and more.

Rough winds

On the store brand side, additional challenges come into play. W. Frank Dell II, president of the Stamford, Conn.-based Dellmart & Co. management consulting firm, counts larger-than-ideal order quantities (resulting in higher inventory levels and infrequent review), as well as longer lead times than the national brands and/or inconsistent lead times (because of plant scheduling issues and long transportation distances), among private label's biggest procurement issues. And on the sales end of things, retailers have no accurate forecasting model for new store brand items.

"There are the rapid- and slow-build sales models, but determining which will work for a new item is a coin toss," Dell says.

On the promotional side, retailers also must deal with a store brand promotional matrix or factor multiplier for that is different from that for national brands, he says. For example, the national brands do not run shadow promotions.

"While more stores are leveraging computer-assisted ordering technology than ever before, many of the advances in demand planning over the last few years have been the result of collaborative planning, especially supported by national branded manufacturers," adds Jim Hertel, managing partner with the Barrington, III.-based Willard Bishop retail consultancy. "While things are getting more precise and more accurate broadly, promotional forecasting is still more difficult overall."

Many private label suppliers do not have the necessary technology, Hertel notes, and also boast less-collaborative relationships with retailers. This reality translates into less-accurate forecasts.

"What happens then can include short shipments, inadequate stock to meet promotional demand and disappointed shoppers," he maintains.

A break in the clouds

Despite all the obstacles, retailers still can take steps to improve the forecasting process. Bill Emerson, founder of Palm Beach County, Fla.-based Emerson Advisors, a retail consulting firm, contends that they first must look to a higher level than forecasting and demand management when it comes to store brand items.

"It really comes down to the strategic positioning of the store brand or the private label, in general," he says, "and the level of respect that you show the brand. You must show as much respect for your store brand as you do for a national brand."

By respect, Emerson means retailers must ensure their own-brand products have the highest levels of quality and value. After all, he says, consumers disappointed with a new store brand items won't be making a second purchase.

"In that case, you'd have real issues in trying to do demand forecasting," Emerson says. "That will be a one-off, where you'll get some erroneous read of what the level of demand is, but people will never come back and buy it again, so you'll end up overstocked on it."

With new store brand items, tradeoff analysis also is critical, Emerson adds. Even if you plan to promote it, locate it in a "strike zone," place it on an end cap or do something else to pique shopper interest, the reality is that it is still likely to replace an existing product on the shelf.

"You don't know for sure what your initial response will be," he stresses. "If that [existing product] has a predictable curve to it and you're replacing it with something else, it will be like any other product, so you're going to have to do some tradeoff analysis."

According to the November 2010 edition of Willard Bishop's Competitive Edge publication — authored by Craig Rosenblum, a partner at Willard Bishop, and Susan Boyme, vice president of marketing for Roseville, Calif.-based Revionics — the most successful forecasting techniques make use of both time-series and demand-model forecasting. Together, the authors say, they provide "a level of granularity to perform merchandising planning, and a more traditional view for financial planning."

Regardless of the data type, Dell also believes most retailers need to scrutinize their data much more often than they currently do.

"The internal review should be daily, not weekly or bi-weekly," he says, adding that today's technologies allow for such an approach.

On the promotional side, Dell recommends creating and maintaining a private label promotional matrix that includes advertising, percent of price reduction, display and display type, web page, frequent shopper e-mail, coupon and other pertinent data. And to slash order cycle time and variability, he advises retailers to communicate requirements to suppliers in advance of issuing the purchase order. In turn, private label suppliers should receive distribution center withdraws and inventory or stock takeaway information daily, he adds.

"Using this information, they should schedule the plant production runs to keep customers in stock and reduce the order cycle time," Dell says.

‘Winners’ Versus ‘Laggards’

Critical to successful forecasting on the store brand front is a well-managed private label program overall. In its 2010 Private Label Management study, Miami-based RSR Research found important differences among private label "winners," "average performers" and "laggards." According to Brian Kilcourse, an RSR Research managing partner:

Retail winners are more concerned with finding and keeping dependable supplier partners than other respondents (69 percent vs. 45 percent).

Retail winners and average performers indicate they are more concerned than laggards are about inconsistent product quality (50 percent vs. 22 percent).

Retail laggards feel most vulnerable to supply chain shocks and getting compliance to specifications than retail winners and average performers (44 percent vs. 20 percent).

Although no retail winners cited an inability to "get their arms around total landed costs," 13 percent of average performers and 11 percent of laggards did.

RSR believes the last point is a core problem for many retailers that sell store brand products, Kilcourse says."Gross margin looks good on the surface," he adds, "but the cost of shipping and handling large quantities of imported product hides in general administration expenses, unallocated to any merchandising process."

Fully Integrated Approach

Malcolm Buxton, CEO of Newport Beach, Calif.-based JustEnough Software Corp., believes retailers could address many forecasting shortcomings by employing a fully integrated approach.

"It basically means using the forecasting philosophy in your planning cycle, and then using the same forecasting philosophy in your sourcing and in your store fulfillment," he says.

He notes that his company offers a fully integrated suite — the forecasting is embedded into the planning, into the sourcing and into the replenishment process for the stores and the distribution centers.

"We're using demand shapes and curves to drive the ordering and drive the placement of products into different stores," Buxton says. "So as an example, if you look at products that have sizes/size scales, at the heart of what we do is putting that forecasting philosophy into the tools that retailers need to run their operations."Using the sales, ordering and inventory-level information, JustEnough is able to "work out exactly what needs to go where and what needs to be ordered," Buxton says. Retailers may employ the services via a subscription service or have the company install the software.

"We have a whole consulting services team whose single goal is not to over-engineer the whole process," Buxton adds, "but to get the software in, get the system running and achieve that in the shortest possible time."

Avoid ‘Forecast Disparity’

Forecast disparity" poses challenges for retailers of all sizes, geographies and demographics — many retail systems have become so complex that different forecasts serve different purposes. Although this vast information is useful, "to gain confidence, it helps to understand where there are commonalities and differences," notes a Competitive Edge report from the Barrington, III.-based Willard Bishop retail consultancy.

Purchasing systems, financial systems, and price and promotion systems all might rely on unique forecasts, the firm reports, but all are built on "the assumption that history tends to repeat itself," the report states. For this reason, new businesses and new products often present a challenge. The key for retailers, therefore, is to identify the right approach, whether focusing on historical and seasonal patterns or on a "bottoms-up approach" that starts with store-specific historical data.

Regardless of the approach, "Development of strategy and tactics dynamically affect your forecast, says Craig Rosenblum, a partner at Willard Bishop and one of the report's authors. "Those who don't realize this or make changes independent of the organization are setting themselves up for failure," he tells Progressive Grocer's Store Brands.

Rosenblum adds that it is essential that retailers understand the role of store brands in relation to the overall chain or banner strategy.

"We have seen too many retailers where they are not in line," he says.

To help guide retailers on the path to successful forecasting, Willard Bishop's report points to two common pitfalls: bad data and a bad model. Bad data, per the report, are inaccurate, inconsistent or imprecise. A bad model, meanwhile, is probably the result of missing variables — for example, ignoring product cannibalization effects or disregarding data anomalies linked to a blizzard-related shutdown.

"Retailers who perform their own forecasts should regularly review input data to ensure that it is as clean as possible," the firm advises. "Small data problems can quickly manifest themselves in the form of bad forecasts."

With data in check, the firm offers guidance on best-practice forecasting. The most successful techniques, it says, incorporate both times-series forecasting and demand-model forecasting.

To learn more, visit www.willardbishop.com.