Markdown Optimization: A Test & Learn Case Study
Anthony Bruce and Patrick O’Reilly
Applied Predictive Technologies
January 2009
www.predictivetechnologies.com
Contents
Abstract
The Challenge
The Solution
The Results
Test & Learn and APT 6
In today’s fast-paced retail environment, keeping merchandise fresh is important to maximizing sales and profit. In particular, identifying and appropriately discounting slower-turning products allows stores to sell more high-margin merchandise. Many retailers mark down products that have been lingering in the store, reducing their profit on those products to increase future profit when new merchandise sells faster at full price. The key challenges in the markdown process involve identifying which merchandise should be discounted and determining the appropriate discount price points. At what point does the loss from selling at a lower price outweigh the benefits of selling items faster?
A Fortune 500 retailer recently struggled with setting the right amount to mark down prices for distinct types of products. The retailer was concerned that the company’s discounts were too aggressive but did not know whether a smaller price reduction would slow sales of those products and ultimately hurt profitability. The retailer decided to test less aggressive price cuts for a subset of the products in the store.
The impact on profitability was still unclear. In need of a precise understanding of the costs and benefits of different markdown strategies, the retailer turned to APT for guidance.
The retailer had reduced the markdown discount for tens of thousands of products in a subset of stores to measure if this would be profitable. In other words, the retailer had tried increasing the price point charged at markdown. However, management was unable to determine the profit impact from the price increase. While the goal was clear, the analytical path was not. The retailer’s analytical team was attempting to measure the difference in product sales between test stores that employed the reduced discount strategy and control stores that did not. However, the analysts were struggling to make this work. Identifying which control stores to use was a daunting challenge.
More challenging still was that the products for which the retailer tested reduced markdown discounts were not carried in the control stores. So for each product for which the analysts tested the higher markdown price, the team wanted to compare markdown performance to a set of “like” products carried in the control stores. Ideally, the team wanted to find a set of comparable products that were similar in nature, had similar sales velocity before being marked down, and had similar amounts of inventory still on hand when the markdown period started. Furthermore, since the team had tested the reduced markdown price on thousands of items, they wanted to repeat this process of matching reduced markdown discount products to “like” products thousands of times. With traditional tools, the analytic team was overwhelmed, and management was frustrated by its inability to learn from the test.
The complex, interrelated effects of the new markdown strategy could not be appropriately analyzed or understood with the limited conventional techniques. The retailer needed to synthesize vast amounts of data, adjust for variance in prior price points and sell through rates, and identify patterns in the responses of sales and inventory levels.
To maximize markdown profit, the retailer needed to:
- Measure the gain in margin generated by charging a higher price at markdown for the items sold.
- Determine the impact of slowing unit sales by charging a higher price at markdown. This included both quantifying the impact on sales velocity from charging a higher price and accounting for the fact that any slowdown in velocity clogs shelf space and delays the introduction of new products.
- Distinguish products for which less price discount is appropriate from items that require a deeper discount.
The retailer chose APT’s enterprise Test & Learn™ software to determine the optimal markdown level. APT’s powerful analytical toolkit analyzed the response in sales, margin, and inventory levels for tens of thousands of products to identify the most profitable way forward for the company.
APT’s sophisticated software provided a precise evaluation of the new markdown structure by comparing products with a smaller price discount to similar products with the original discount. Using this software, the retailer determined that reducing the markdown discount did cause a quantifiable decrease in the sales of those products. However, even after incorporating the costs from slowing sales and delaying shipment of new merchandise, the strategy of smaller markdown discounts was successful on average and would generate $600K in pre-tax profits.
The retailer was also able to identify a more targeted markdown strategy using APT. The software examined different types of products, fashion types, price points, pre-markdown sales patterns and other characteristics to build the profile of product types for which a higher price at markdown was optimal. In doing so, APT’s software identified the types of products for which sales slowed less when the markdown price was reduced. Conversely, the software identified the types of products for which aggressive markdown discounting was necessary.
By targeting the reduced markdown discount program only to those product types for which the program was well-suited, and not introducing unprofitable markdowns, the retailer was able to reap significantly higher profits. Using this highly tailored markdown pricing strategy, the retailer is now generating $5M per year in incremental annual pre-tax profit.
The APT software answered the retailer’s central analytical questions in a timely and cost-effective manner, delivering a solution that inspired management confidence and ongoing bottom-line benefits:
- APT used test vs. control methodology to identify products with similar attributes and price levels and sell through rates prior to markdown. Comparing the sets of products showed that sales observed a statistically significant slow-down for products with higher markdown prices.
- Rigorous statistical analysis coupled with time period customization demonstrated that increasing the markdown price was profitable even after accounting for fewer units being sold at the higher price and fewer units being shipped and sold in the future at full price.
- Segmentation analysis in APT identified product characteristics that drove higher and lower profit impacts. By expanding the markdown optimization to products with high-profit characteristics, the retailer increased profit by over $5M per year.
Using APT, the retailer was able to quickly move from frustration and guesswork to a highly targeted markdown pricing strategy that drove a $5M increase in annual pre-tax profits.
Encouraged by these results, the retailer has begun planning a broader markdown testing agenda, such as determining when to take markdowns, and when and how to aggressively move to sub-markdown prices. In applying these analytical techniques more broadly across the store, the key open questions continue to involve identifying the right product and the right discounted price point at the right time. Without APT, it was not possible to provide an integrated view to determine the best markdown price and to identify products for which markdown prices should increase. APT identified significant profit opportunities that could be realized by modifications to the retailer’s current markdown strategy. Its data-driven approach defined a clear, profitable path forward and offered a platform for further testing.
In the business case discussed, a disciplined Test & Learn™ process was employed using the APT 6 software suite to uncover the most profitable path forward.
APT’s mission is to help leading retailers, banks, and consumer goods companies to institutionalize a world-best Test & Learn™ process. In support of this mission, APT addresses the expense of testing and the complexities of analysis and reporting with cutting-edge technology designed to support testing in the retail channels, and a rich library of best practices developed to consistently and rapidly propel the creation of important business insights that drive significant profit improvements.
APT 6 is the enterprise information system designed by APT to support best-practice Test & Learn™. Leading companies around the world use APT 6 to improve the effectiveness of their investments in capital expenditures, marketing, operations, merchandising and site selection.
APT 6 automates complex data collection, provides sophisticated tools required to design and interpret tests correctly, generates consistent and clear outputs for each test, and supports a disciplined and repeatable process for business users and senior decision-makers across the organization. That is why more than 40 of the Fortune 500 use APT 6 to drive dozens of high-impact tests in their organizations every year and to support business actions that create tens of millions of dollars in annual profit.