Rock Around the Stock

Taking its cue from the Q system, Guitar Center’s forecasts and inventory allocation now make beautiful music together.

Blues, rock, country, hip-hop, salsa – American tastes in popular music run the gamut. And the sounds that are big in El Paso this year might be totally different from the sounds that are hot in Brooklyn, or Nashville, or Spokane.

So when your business is selling musical equipment, imagine how hard it is to keep each of 200 stores across the country stocked with the mix that’s in tune with the local music scene.

That’s what Guitar Center was wrestling with three years ago. Part of its challenge stemmed from the fact that its stores differ greatly in size, ranging from 5,000-square-foot shops to 30,000-square-foot big box locations.


“Also, the types of customers we deal with vary widely depending on demographic and geographical regions,” says Bret Hayden, director of business process design at Guitar Center, Westlake Village, Calif.

The products that Guitar Center carries – guitars, amplifiers, percussion instruments, keyboards, and professional audio and recording equipment – amount to 7,000 SKUs.

To serve customers and keep profits high, the company must understand how each SKU performs in each store. A homegrown forecasting system, developed in Microsoft Excel and Access, wasn’t hitting the right notes.

“The forecasting system operated at the chain level, but we really needed to be looking at inventory at the store level,” Hayden says. “We needed the ability to look at each one of our SKUs, and each one of our stores, and understand how they perform differently from one another.”

In addition to a system that provided insufficient detail, Guitar Center faced another challenge when trying to understand the store/SKU relationship.

The company’s forecasting team used one set of business rules to determine the volume and mix of products to send to its distribution centers, while the allocation team used a different set to create the product mix for stores.

“We would end up with a serious disconnect between what forecasting thought was needed and what allocation thought was needed,” says Steve Johnson, Guitar Center’s director of forecast, allocation and replenishment.

Today, however, Guitar Center integrates forecasting and allocation in a single process, and is much better able to tailor each store’s product mix to local demand. These changes came about through the company’s work with Quantum Retail Technologies.

Guitar Center has served as a beta customer for Quantum, helping the Carlsbad, Calif., software firm develop its inventory optimization solution, Q.

The retailer ran a pilot implementation of Q in late 2005 and early 2006; then entered a detailed design and implementation phase to address its long-range forecasting and product allocation needs.

That version went live in the third quarter of 2006. A third phase of the implementation—adding commodity products such as guitar strings and drumsticks—was due to go live in late June 2007.

Too Much Data

Quantum developed Q to meet the needs of retailers who, over the last few decades, have increasingly moved decision-making responsibilities from store managers to home-office executives. Those executives base many decisions on sales data pulled from the stores.

But their enterprise resource planning (ERP) systems can’t analyze such a vast volume of information in great detail, says Mike Hrabe, Quantum’s vice president of sales and marketing. Instead, they aggregate the data and look at average performance for categories of stores and items.

“Through that smoothing, averaging, and aggregating process, retailers have effectively eliminated much of the detail associated with how items behave at the store level,” Hrabe says.

Ignoring the store-by-store detail obscures important information, such as whether a store is stocking the right product quantity, notes Chris Allan, Quantum’s founder and head of product strategy.

“A 98-percent in-stock of a certain product across the chain doesn’t really show a complete picture,” he says. “Some locations may be out of stock for several weeks; others may be overstocked.”

Q uses data from point-of-sale systems, ERP systems, and warehouse management systems to track exactly how much inventory each store has, how fast it’s selling, and how much new stock is flowing through the pipeline. In making forecast and allocation decisions, Q also considers the role each product plays in the company’s merchandising strategy.

A popular product at a marked-down price plays the role of traffic driver, Hrabe explains. The margin is low, but it draws in shoppers who might make other purchases while they’re in the store.

Another product, with a higher profit margin, is a money-maker.

Still another serves as an image item, bolstering the store’s prestige by its presence even though few people actually buy it. Think of a giant screen TV in a consumer electronics store, he says.

Products play different roles in different stores. “An image item at the Best Buy in suburban Minneapolis might be a money- maker at the Beverly Hills Best Buy,” Hrabe says.

Demand for products also changes over time. As Q recommends inventory allocations for different stores, it considers the roles the company assigns to different products at those stores; then it tracks the products’ behavior to see how well they play their parts.

More precise information about product demand and performance creates greater efficiency. “Retailers hold too much inventory for fear of losing sales, but over-inventory means lost profits,” Hrabe says.

“Retailers have unbalanced inventory because they use grade group averages and lose much of the detail. They end up with too much inventory at one store, too little at another. Q directly addresses these issues,” he adds.

At Guitar Center, the point-of-sale system feeds data into a JDA Software ERP system, which passes it along to Q. Then, Q’s recommendations and alerts pass back to JDA and to the company’s Arthur Allocation system.

“As part of Phase 3, we will integrate Q with our warehouse management system, so we’ll have information regarding shipment delivery times,” Hayden says.

Each time Guitar Center adds a new product to its assortment, the buyer and planner assign it a role and a strategy.

“They can also set up other types of parameters,” Hayden explains. “For example, they can plan for a display in the store for that product, or set a ‘max stock’ if the item is big and bulky.”

The company could assign those rules to each product on a store-by-store basis, but executives have decided to move one level higher, dividing stores into several “grades” based on their characteristics. Stores get different grades for different product categories.

“One store could be an ‘A’ store for drums, but a ‘C’ store for guitars,” Hayden says. “We have the ability to manage inventory using those grades.”

Besides helping Guitar Center planners determine what stock to order and how to allocate it to stores, Q monitors product performance in real time and tells planners when product performance doesn’t match the forecast.

For example, Q notifies planners if an item is selling better than expected. The planners can then arrange to order larger quantities in the future.

The Missing Piece

Company officials are contemplating a possible fourth phase to the Q implementation, which would focus on assortment planning.

“That’s the piece we’re currently missing in our suite of applications,” Hayden says. “We’re able to create strategies for these items, but we don’t have good visibility to how that item fits in the whole assortment.”

Quantum representatives also have been talking to executives in Guitar Center’s Music and Arts Center division, which serves the school band market through more than 90 stores.

Since Guitar Center started using Q, service levels and in-stock rates have increased, with a better inventory balance across the chain. “We don’t have as many over- and under-stocks as we had in the past,” Hayden says.

Also, now that it’s monitoring performance at the store and SKU levels, the company can generate more exception reports, and can measure forecast error. Those exception reports are important because they alert planners to problems or anomalies in parts of the operation that weren’t receiving enough attention.

“Q helps maximize users’ time and makes sure they spend their work hours where they can add the most value,” Allan says.

And that’s music to Guitar Center’s ears.

Leave a Reply

Your email address will not be published. Required fields are marked *