Markdown Modeling

A markdown is a permanent form of price adjustment used by the retailer to “liquidate” the inventory of a product or category of product. There are many different approaches to how retailers manage this form of price adjustment. Some retailers have an internal policy that establishes how much the first markdown must be. Depending on the retailer or the product category, first markdowns can range from 20% off to 50% or even 60% off the original price.

Also, retailers may have a set schedule of when subsequent markdowns are taken to ensure that the inventory will be eliminated by a target date. For example, an item marked down to 40% off could then see an additional price adjustment in four weeks, with an additional 30% off being applied to its price. Two weeks later a retailer could slash further and take another 20% off that price. The markdown process can be manually managed (monitor sales based on the latest price adjustment and calculate weeks of supply) or automated using a variety of software applications available to retailers.

Markdown modeling evaluates the past behavior of products within a single merchandise category.

Markdown models have a core set of variables designed to capture what matters most in markdown:

• Seasonality
• Lifecycle (Distinct from Seasonality)
• Price

Retail Buying and Markdowns

Retail buying is less about fashion and more about math. It is essentially about determining how to sell merchandise before you pay for it. If you do not understand the math and that you will have to mark down a percent of the items that you purchased, you will lose money.

We tend to fall in love with our buying decisions, sometimes to the detriment of our store. Markdowns are inevitable. But markdowns done right can be healthy; they keep a store fresh and inviting. Many retail store owners believe their items will sell for the original price “if they just give it a little more time.” Customers vote early on your merchandise, so it’s a smart strategy to mark down items sooner rather than later.

Timing Markdowns

Use a calendar to plan the timing of your markdowns. They should occur after the items have been in the store for a certain period of time. For example, some stores automatically mark down an item after 60 days, especially if it is not a regular, replaceable item. Then, the next markdown may occur at 90 days, with the final markdown planned at 120 days. This process can pressure an owner to use an open-to-buy system (It’s the amount of merchandise your retail store can buy during a certain time period. In other words, it helps determine the amount of inventory that you will need to purchase to meet customer demand while maintaining a positive cash flow.) that will keep inventory current and salable.

Managing Inventory

As a general rule, use the following “Rule of One-Thirds” model to determine how much inventory to buy. For example, if you buy 10 radios, you will sell one-third at full price, one-third at a 25 percent discount, and one-third at a 60 percent discount. Doing the math, you may be inclined to buy 7 radios instead of the initial 10. In other words, the more backstock you have of an item, the more likely it is to be marked down at some point.

Another item to consider when analyzing your markdowns is inventory turnover. There is a direct correlation between inventory turnover and markdowns. A high turnover usually means fewer markdowns, while a lower inventory turnover usually means higher or more markdowns.

Dating Purchases

Use dating on your purchases from vendors to help your cash flow and offset the timing of markdowns. Dating is when the vendor gives you extra time to pay for merchandise after it is delivered. For example, most invoices are due within 15-30 days after the merchandise is shipped; however, if the vendor gives you more time to pay, you may actually be able to sell it before payment is due.

Markdown Plan and Scenario

A markdown plan consists of a collection of scenarios that are optimized to return recommended markdowns for products within a time period and for specified store locations.

A scenario consists of objectives, a schedule assignment, and default constraints that you can customize. You set a number of rules and objectives with information such as the dates of markdown occasions, markdown percentages, and markdown price points. When the scenario is optimized using these rules, schedules are generated with a series of markdown occasions that contain markdown percentages (or price points) over a time period. For example, an item may have a series of occasions, such as 25% for the first markdown, 50% for the second markdown, and 75% for the third markdown.

Markdown Analytics

Markdowns benefit the consumer, in terms of providing a choice between paying the full price for a product with a long shelf life or a reduced price for a short shelf life — this choice tends to be welcomed by many consumers. At the same time, store managers generally prefer to sell those products reaching their ‘sell-by’ dates, rather than have to remove them from the shelves and dispose of them in the waste bin. Therefore, the store manager should support the introduction of an intelligent markdown pricing strategy.

The analytics challenge is to find the pricing ‘sweet spot’

What different industries have in common is the analytical challenge of setting a price that is neither too high nor too low. If the chosen price point is too high, stock remains on the shelves, and does not get sold before it has to be written off. If the chosen price point is too low, the product will sell out too fast, even before the end of the trading cycle, potentially resulting in stock-outs. On top of this, excessive discounts lead to lower revenues than could be achieved from this product.

Markdown pricing strategies across disparate industries pertain to widely different products, with their own life cycle lengths. Although these markets and products are quite different, they have an overarching dynamic in common: How does one find the ‘price reduction sweet spot’ that will maximize product sales, yet at the same time minimize the discount the seller has to offer, in order to optimize profits.

Intelligent markdown pricing must integrate seamlessly with existing business processes. Intelligent markdown pricing is not solely about analytics, because it has to work within a business operation that is based on people, process and technology:

People — In the traditional ‘bricks and mortar’ retailer, markdown pricing usually relies on store staff to identify items that are nearing the end of their selling life, check their stock levels and apply the appropriate process to reprice those items. Furthermore, subsequent stock checks and further price reductions may need to be made, requiring additional staff time. The store manager has the authority to override the system-generated markdown prices — for example, if the manager believes that a deeper price reduction is needed in order to clear all the stock that has reached its sell-by date.

Process — Any new markdown pricing strategy will have to be consistent with the existing process for pricing products. And every business will have such a process, whether crude or sophisticated.

Technology — The technical infrastructure that supports the pricing process is another ‘given’, although we comment below on the data infrastructure that is likely to be most suitable for this application.

The interplay between these components can be complex — in a supermarket, for example, store staff may need to make multiple checks (or scans) of perishable products reaching their sell-by dates, either to capture data or reprice items:

• An initial check on the day before, in order to obtain up-to-date stock levels for the items that will need to be cleared.
• Another scan on the last day of sale, in order to generate reduced-price stickers and reprice the items.
• A further scan later in the day, in case a second markdown is required.

All three components, making up the existing pricing/markdown operation, will need to be fully understood and allowed for, as part of any approach to ‘intelligent markdown pricing’.

Value of timely and relevant information

In order to respond accurately to the current stock level, one needs to confront a forecast1, 2 on expected sales (assuming no intervention by the retailer) with current stock levels. If the forecast expects (considerably) lower sales than available in stock, discounting is one of the options to consider in order to avoid waste.

Timely availability of data is crucial

Since the forecast is based on the last known level of inventory, a timely feed of input data is crucial in order to decide on any discount recommendation. In many cases, these data feeds are passed along into a data warehouse (DWH),3 and therefore the speed with which they can be updated is crucial. The meaning of ‘timely’ depends on the volatility of sales, the particular industry and stock levels (and promotions) of complementary and substitute products. Data that arrive late can be highly accurate and properly cleansed, but are utterly useless if they come too late for making markdown recommendations.

Value of historic information

In order to identify the ‘pricing sweet spot’ for a product, one needs historic sales and price data to calculate price elasticity (PE). At the very least, the analyst would want to have weekly data for at least one complete year, to support PE modeling.

Having several years of history allows seasonal patterns to be identified

Analysts always seem to want more data, of better quality. However, in the end, they will simply need to do the best they can, with whatever data they have available. Since one needs to consider cyclical (seasonal) patterns in sales, at least one entire cycle is needed, but preferably several. So when looking to discount summer clothes, at least a full year is required, but if at all possible one should obtain 2–3 years or more. Analysts attempt to determine seasonal patterns; therefore, it is preferable to have multiple seasonal recordings in order to separate ‘signal’ from ‘noise’, that is, to discern ‘true’ seasonal variation from erratic fluctuations. Global retailers can sometimes benefit from the fact that summer is only a half year later on the southern hemisphere, but even then, having more data is always better.

Value of granular data

Granular data, at basket level, are of great value in two ways — at the first and last stages of the analytics process.

First, they enable the retailer to examine current patterns of markdown sales and answer business questions such as:

• How do marked-down items sell across the trading day? When do the first reduced-price transactions take place?
• How many price reductions are required in order to clear the entire stock of a product that is reaching the end of its selling period? And how deep are the markdowns that are being made?
• How successful is the retailer in clearing marked-down products, and how does the success rate vary between stores, product types and price points?
• What are the characteristics of customers who purchase marked-down products at different times of the day, week, month or year (depending on the cycles that are typical of this particular industry)?

At the last stage of the analytics process, which we outline later in this paper, the granular data set can be used as a ‘test bed’ for simulating the effects of repricing the product. It allows one to predict the markdown sales and calculate the reduction in waste loss. In other words, one can estimate the return on investment that could be achieved through intelligent markdown pricing, which invariably is the most compelling way to convince the retailer’s management team to move forward with this analytic approach.

Developing  markdown pricing

The following steps create a markdown pricing solution

• Understand the existing markdown pricing process.
• Carry out preliminary analysis to examine levels and patterns of waste losses currently being incurred.
• Calculate PEs for products being marked down.
• Create an alternative markdown pricing strategy designed to reduce the losses.
• Simulate the impact of the pricing strategy and measure its benefits.
• Carry out a live test to see how customers react to the change in markdown pricing.

Retail Markdown Calculation

Whether it’s called a markdowns sale, promotion or clearance event, a retailer is going to receive less money for the product than they originally wanted.

Retailers calculate markdown percentage with the following calculation:

Markdown percentage = Amount of reduction / Original selling price

Here’s an example:

An entire rack of designer shirts priced at Rs. 100 is being discounted to Rs. 60 for the holiday weekend. The manager wants to know what the markdown percentage is. To calculate it, use this equation:

Markdown percentage = Rs. 40 (Amount of reduction from original selling price) / Rs. 100 (Original selling price)

The markdown percentage, then, would be 40 percent.

The formulas for calculating markdown price is as

Markdown Rs.  = Original Retail Rs.  × Reduction Percent

New Retail Price Rs.  = Original Retail Price Rs.  – Markdown Rs.

Example:

The merchandise manager of the store requested that the buyer prepare for a weekend special promotional sale. After looking at sales figures and product classifications the buyer decided to run a 25% off sale on some trendy cotton blouses and cotton blend shirts. The buyer submitted the following plan to the merchandise manager:

• In the Missy Sportswear Department, he would advertise a 25% off weekend sale on a select group of trendy cotton blend shirts and cotton blouses.
• The buyer reduced 24 cotton blouses originally retailing at Rs. 56.00 and 18 cotton blend shirts originally priced at Rs. 48.00 for the 25% off sale.

The retail sales promotion was very successful.

• The buyer sold 18 of the 24 blouses at the 25% off retail price for Rs. 42.00 each and 10 of the 18 shirts at the 25% off retail price for Rs. 36.00 each.
• All of the remaining blouses and shirts sold at a second reduction of 33 & 1/3% off.

Using the above steps to calculate the markdown price as

Markdown Rs.  = Original Retail Rs.  × Reduction Percent

Markdown Rs.  Blouses = Rs. 56.00 × 25% (.25)

Markdown Rs.  Blouses = Rs. 14.00

Markdown Rs.  Shirts = Rs. 48.00 × 25% (.25)

Markdown Rs.  Shirts = Rs. 12.00

New retail price –

New Retail Price Rs.  = Original Retail Price Rs.  – Markdown Rs.

New Retail Price Blouse = Rs. 56.00 – Rs. 14.00

New Retail Price Blouse = Rs. 42.00

New Retail Price Shirt = Rs. 48.00 – Rs. 12.00

New Retail Price Shirt = Rs. 36.00