Forecasting Profits

Oct 01, 2005 9:30 PM  By

Selecting the merchandise that customers want to buy is fundamental to your business’s success. So is ordering the correct amount of that merchandise. Buy too little, and you risk backorders, cancellations, and disappointed customers. Order too much, and you’re stuck with a warehouse full of inventory you can’t sell, representing a waste of time, money, and space.

That’s why many mailers rely on inventory forecasting software, which can provide an accurate heads-up on the coming season’s product needs. Choosing the right system may seem daunting, but it comes down to considering such factors as the capacity of the system for interpreting data, the type of merchandise you sell, how well the system will integrate with the rest of your warehouse management software, and the technical support you need.

SO SOPHISTICATED

In the three years since Salt Lake City-based Sundance Catalog Co. implemented an inventory forecasting system, its fill rate increased from around 78% to about 90%, says director of merchandising and inventory control Dievo Hagen. Its in-stock rate, which had fallen below 78%, has gone up more than 20%, and it has been able to reduce expenses tied to backorders by 60%-75%.

The cataloger of southwestern-style apparel and decor used to rely on nothing more sophisticated than an Excel spreadsheet listing product sales from past seasons to guide its purchases from vendors. But in 2002, Sundance selected Forecast*21, a system from Omaha, NE-based Direct Tech, to improve its inventory forecasting and management capabilities considerably.

The software can dissect data on past sales of merchandise in ways that an Excel spreadsheet cannot, Hagen says. The average item index, for example, ranks sales of each product in one catalog mailing against another. If the number of units sold per SKU for a catalog drop is 100, for instance, but you sold 200 of a particular SKU, then the software alerts you that you need to order twice as many units of that item next time or that you should order more as soon as possible so that you can fulfill more orders while the catalog’s sales curve is still in motion. On the other hand, if you’ve sold only 50 of another SKU from that same mailing, you’ll be warned that you should try to return a portion of the inventory to the vendor or begin liquidation efforts.

“It can also look at individual items based on revenue, units sold, units per 1,000 books mailed, or dollars per 1,000 books mailed,” Hagen explains. “There are lots of different ways to use the system to see how an item, and similar items, sold historically.”

Covington, KY-based faux-fur merchant Donna Salyers’ Fabulous-Furs implemented the Inventory Forecast Manager (IFM) module from Delray Beach, FL-based Ecometry Corp. this year. CEO Guy van Rooyen says the company, which had also purchased the Ecometry warehouse management system, hopes the software will reduce its excess inventory by 50%.

Two years ago, says van Rooyen, backorders had been an issue for Fabulous-Furs; last year, overstocks were the problem. “Our hope, based on the demonstrations we’ve seen, is that it will provide us with real-time analysis of how our products are running, the ability to manage our entire SKU data, and dictate when purchase orders need to be issued,” he says.

Fabulous-Furs, which carries 1,300 SKUs and mails 6 million catalogs between mid-July and early January, generates the bulk of its sales during fall and winter. An inventory forecasting system “makes a difference when 80% of sales are coming in a short span of time,” van Rooyen says. “We want to get a better analysis to determine what’s in the holiday and sale books based on fall demand.”

Warren, PA-based apparel and home goods mailer Blair Corp. purchased the Galvin System from Marstons Mills, MA-based Galvin Associates eight years ago following the launch of its first bound catalog, says Bob Crowley, senior vice president of menswear, home, and marketing services. Prior to 1997, Blair had used loose-sheet mailers to sell its wares. The company felt that the demand curve of catalogs necessitated a weekly monitoring of inventory needs.

“We had never done assortment planning before then, category management and the time phasing aspect of it — how many units are needed this week and the week after that,” Crowley says. “When we’re building books, we’re able to use this system in conjunction with all of our other metrics to determine if a product is delivering [on par with] the contribution of space allocated to it.”

That information comes in handy when deciding which products to reorder for the next selling season, says Crowley, given the high number of repeat merchandise in the Blair catalogs (as high as 60% in the men’s book). Blair also makes use of the system’s “exception-based analysis” reports, which tell buyers what products to devote the most inventory supply dollars to. For example, if there are 300 products in a catalog, the software might tell users to devote the majority of the inventory budget to 25 items that are exhibiting 10% or greater deviation from the original forecast.

This early warning on inventory status helped Blair avoid a potential pitfall this summer in its purchase of cotton goods from China. Due to the import quota placed on cotton goods at the beginning of the summer (see “Will Chinese quotas pinch?” in the June 2005 issue), the company needed to know by early July exactly what it had to order from China. “We knew we would be in jeopardy of losing delivery if we didn’t have everything here by July 8,” says Crowley. “The system allowed us to ID which products we needed and place our orders in time.”

Overall Blair is satisfied with the software’s ability to keep backorders to a respectable minimum, Crowley adds. The company has an average of $3 million-$5 million tied up in backordered merchandise at any given time.

SUITING YOUR COMPANY

A forecasting system’s ability to dependably analyze data depends on the number of channels you need the software to serve, says George Mollo, principal of Nanuet, NY-based GJM Associates, a consulting firm specializing in merchandising operations for direct marketers. The systems most commonly used by catalog and Internet sellers, Forecast*21, and the Galvin System, need to be tweaked to be used reliably for a company’s retail sector, he says.

“In retail, foot traffic is foot traffic, but demand in a catalog is based on circulation,” Mollo explains. “If next week you’re mailing to 1 million, you’ll have a huge increase [in demand], and then it will tail off.”

No system can accurately predict inventory needs for every channel, Mollo continues, which is why he recommends programming into the software a more consistent demand curve for stores (except, of course, when major store sales events are occurring) and overlapping bell curves to represent demand in response to catalog mailings. (Because a second catalog mailing can reach customers during the tail end of the previous mailing, the effects of both mailings would be felt simultaneously.)

You must also be able to tailor the capacity of the system to the kind of merchandise you sell, says Richard von Hirschberg, senior consultant with Columbus, OH-based retail consultancy Columbus Consulting. A company selling cleaning supplies or tools may do fine with a system that can be programmed to automatically replenish inventory when it dips below a predetermined level. (The i2 Replenishment Planning solution from Dallas-based i2 Technologies is one such system.) But a marketer selling a trend-driven product such as apparel needs a system, such as Arthur from Scottsdale, AZ-based JDA Software Group, that gives inventory managers and merchants more control in altering supply based on unexpected demand or lack thereof.

“Fashion forecasting is far more subjective,” says von Hirschberg. “Forecasting for replenishment is more automated, with forecasts automatically going through the model [inventory parameters that companies program into the software, such as lead time from vendors and seasonality]. A merchandise planning system relies on more manual input, calculating the results of subjective decisions.”

Since pricing terms between vendors and companies tend to be kept private, most catalogers will not cite specific costs. But von Herschberg says pricing for inventory forecasting software can range from $300,000 to $500,000 (including implementation and licensing) for smaller catalogers that may license only a dozen or so users to $500,000-$1 million for companies with 50- 100 or so users, and as high as $5 million-$10 million for corporations such as The Gap that might have up to 500 staffers licensed to use the software.

INTEGRATION CONSIDERATION

The cost of the software isn’t the only expense to bear in mind. You’ll want to avoid incurring additional costs as a result of complications integrating the software with your existing warehouse management and database systems. Many companies, in fact, shy away from investing in forecasting software because they fear the new software won’t be able to extract the sales information needed from the company’s database, says Sunita Gupta, executive vice president of Cleveland-based retail management consultancy LakeWest Group. Integration problems sometimes occur, for instance, when new forecasting software needs sales data on a weekly, rather than monthly, basis or by product SKU rather than product category.

A chief consideration for Blair was the ability of the new software to integrate with its warehouse management system, which had been written in Assembler, an old computer language. “We started with 19 different systems,” Crowley says, “and took a project management approach, narrowing it down to six or so, and from there we narrowed it down to two. The primary reason we chose Galvin was we were able to use it with our legacy system.”

The ability of the software to work with its existing software was key for Sundance as well. “We wanted the best structure we would be able to integrate with,” Hagen says. “The software is only as good as its ability to get data out of your current ordering system.”

As well as accounting for the cost of integrating a new solution with your existing systems, you should consider how easily the forecasting software will integrate with future add-ons. Sundance, for instance, knew that Forecast*21 has, in addition to the inventory management tool it purchased, merchandise assortment planning software that it could integrate at a later date.

“The cost is the integration,” Hagen says. “The product might cost $10,000, but it could cost $40,000 to write the integration language to communicate with the other products you own.”