Reprinted with kind permission from Jan '99 Research Conference Report, Skokie, IL
Tel: (847) 673-6284 E-Mail: rbrrfl@flash.net

SPEAKERS: Don Patchell, Director, Marketing Productivity, Sears Roebuck & Co. (Hoffman Estates, IL), and Ross Link, President, Marketing Analytics, Inc. (Evanston, IL)

PRIMARY POINT: Sears is similar to many companies in its need to understand and optimize results from short- and long-term brand spending.

BACKGROUND: The number and variety of products sold by Sears presents a significant product-spending evaluation challenge.

Patchell explained how Sears Roebuck & Co.'s extensive marketing strategies in numerous media vehicles (TV, radio, newspapers, magazines, direct mail and sponsorships) presents a challenge to the company in determining the effectiveness of its advertising. It is a problem that is exacerbated by the quantity of advertising, competitive pricing, timing and even weather. Discerning those media that are most effective when they all run at the same time was an issue Sears felt it needed to answer.

Building a market response measurement capability required development in data infrastructure, analysis methodology and business implementation. Sears focused on aligning marketing support factors and resulting sales from its own and external sources. Marketing Analytics' analysis methodology used four basic modeling approaches. The first establishes promotional baselines. This is accomplished by adaptations from ACNielsen and IRI time series methods. In calculating a baseline for sales, promoted and holiday weeks are either removed or "fixed." An estimate of base sales during non-promoted weeks is created. The baseline equals the sum of all nonpromoted items in a product line. This "cross-sectional adjustment" is applied to items in the line. After a baseline is established, the actual sales due to promotions can be compared to the baseline in order to estimate the effectiveness of a sale or advertising campaign. "Automated promotional baselines allow us to calculate a return on marketing investment," said Link, "by separating incremental sales due to short-term promotions from base units."

Marketing Analytics' second step involves lift models that allocate sales to their drivers using regression-based lift models. These break promotional incremental volume into its causes (e.g., print, promotional TV, discounts, etc.). The lift model can optimize marketing spending by showing the most effective advertising vehicles at different times during a season. The company can shift to the most efficient media, maximize during its peak season, avoid oversaturation and optimize pricing.

Among such elements, pricing plays an especially dramatic role for Sears products. The study of price elasticity and margins has led to a pricing rule designed to optimize short-term profit. "We attempt to help Sears maximize short-term profit by adjusting price to raise thin margins and cut fat ones until margin equals one divided by elasticity," Link explained. "Pricing decisions should be driven by balancing price sensitivity and margins, with consideration given to whether a forecasted volume loss is acceptable. Volume may be a legitimate goal, but using the wealth of price-sensitivity modeling information can ensure the most profit is derived from any given volume and vice versa."

The third analysis methodology, focusing on volume generated by long-term sales drivers (e.g., brand advertising), is estimated through complex marketing mix regression. A fourth analysis compiles data through a formal nonlinear optimization using results from the first three methodologies. The progression to this "full mix" model allows Sears to break base volume into its causes (for example, image and branding TV, as opposed to seasonality). Print ads, a powerful Sears selling tool, can be applied to best effect for each product line. Each promotional offer in an ad can be measured for its ROI. In the end, marketing response findings can steer Sears towards easy-to-implement marketing spending that uses most efficient media, spends at proper times of the year, avoids oversaturation and optimizes pricing while maintaining image.

Patchell noted that Sears is educating its entire organization, including upper management, regarding these prized Marketing Analytics' methodologies (which have apparently outperformed other tested methodologies), so that Sears business teams can recognize and appreciate their applicability and availability.