Pricing optimization is one of the best tools you have at your disposal to increase your profit. Studies have shown base price optimization can yield an increase of 2% – 5% in margin, promotional optimization can yield 5% – 20% and mark down can lead to a 6% – 10% improvement. That is too great of an opportunity to ignore. If you are not using science and have a good amount of transaction data, then you could almost certainly benefit from using optimization. If you are considering optimization, you can take steps to make sure you are fully prepared to take advantage of the solutions.
First, a brief explanation. Products go through different lifecycles which closely tracks with what types of algorithms you can use to optimize prices. Many products adhere to a lifecycle where the product is introduced, then sales increase, eventually even out, and finally decreases at the end of life as inventory is sold through. Each stage in the product lifecycle requires different optimization techniques. The initial and day-to-day price is established at introduction and you monitor performance for a period of time before using promotions to increase sales and profit. The initial price can be optimized but is typically bounded by constraints and business rules you have which limits optimization. Promotions allow more freedom in using elasticity to understand what the best price is and mark down optimizes your sell through.
The general barometer mentioned above is valid in most cases and you can do some high level analysis to determine what benefits you can achieve, but truthfully if there is an opportunity you won’t be able to realize the benefit unless you can actually do the optimization. So, instead of discussing a process for estimating the opportunity, we’ll discuss how you can figure out if you can unlock the opportunity.
For optimization to work, you must have enough data and price variation. The data elements needed depends on the type of optimization you are doing but you always need base data like products, location, and sales history. You might also need things like inventory positions, marketing instruments, cost, and promotions. Below we explain what each element is and how it relates to the specific optimization:
Sale price. It is important to have the price the customer sees when they make a purchase. As simple as this sounds, its sometimes difficult for companies to get this price. For example, if you’re a manufacturer, distributor or any other entity that does not have control over the final price the customer sees, it may be difficult to get it. Retailers have the transaction data, but in many cases the data isn’t clean and needs to be fixed.
Number of units sold. The transaction data will also include the number of units sold per location. Number of units sold and the sale price are the foundation of your historical data which is used in the forecast. If you can’t get the number of units sold, you can possibly get the number of units shipped to a given location. This isn’t ideal, but it’s better than nothing.
Price variation. Sometimes it is difficult to get enough data to build an accurate demand curve. But you can get it good enough then use basic analysis to set your price. Price variation can come from many different sources such as discounts, coupons, and price errors. It is essential to know the regular price, the promotional price, and the date range when the price was in effect.
Cost. When optimizing for profit, you’ll need to know how much you paid for it.
Marketing instruments. The marketing instrument used can influence the effectiveness of the promotion significantly. When capturing the price variation, it is important to know exactly what instrument was used because not all instruments are created equal.
Competitive prices. If you are in competitive markets, you’ll need the prices these competitors and the proximity to your stores. The same is true for your eCommerce channel. This data is tied to business rules which drive day to day pricing.
Inventory. If you’re trying to do mark down optimization, you’ll need inventory positions at each location including stores or distribution centers. Inventory would also include any future buys that have been made already.
Not all of this data is necessary to get started with optimization and you can add new data streams after your initial dip into the optimization pool. The basis for optimization is the forecast. If you don’t have enough price variation or data, you may need to substitute similar products, aggregate at a higher level, or use other forecasting techniques to get an accurate picture of demand. When you are trying to evaluate whether or not you can do optimization, the data is analyzed to see if there is enough to feel statistically confident.
When you’ve verified you can do optimization, what is all this data used for? For day-to-day pricing, a lot of the prices are dictated by business rules. These typically restrict the prices in a narrow band based on competitive products, target price points, and other factors. After that there is a small amount of room to maneuver using price elasticity. Promotions have more latitude in using price elasticity and also consider cannibalization and halo effect from other products. Finally, mark downs are constrained by available inventory and try to maximize your sell throughs based on your business goals.
If you pass the litmus test for having the data, you have an opportunity. The next step is to go through the process of collecting, cleansing, and preparing the data for an optimization tool so that you can unlock the potential benefits. In a future article, we will discuss how you use this data in each of the different types of optimization.