Like everything in existence, products too have a lifecycle. However, the constant innovation and technological development that create new products every day, are making product lifecycles shorter and more unpredictable. In a glimpse, shelves and warehouses can be full of unsold inventory.
Many retailers resort to price markdowns to increase demand as products approach the end of their lifecycle to clear inventory. However, most of markdown pricing decisions are manual and lack an analytical base. Pricing decisions are based on a set of empirically defined rules, with no actual analytical evidence and no distinction between products and stores. The process becomes inefficient and time-consuming, making it difficult to define a clear strategy.
Also, it is challenging to balance the risk of out-of-date stock at the end of the cycle with the loss of margin due to pricing markdowns.
It becomes clear that understanding how a price change affects demand is crucial when making markdown decisions, specially when we are talking about a retailer with thousands of stores, each with different behaviour.
The solution to this complex challenge is to implement an advanced analytical model to define optimal prices of end-of-life items at each store.
Using machine learning algorithms, it’s possible to estimate products’ expected sales during their lifecycle and price elasticity and, thus, support markdown decisions by determining, for each product at each store, the best markdown pricing strategy.
An effective analytical approach to a markdown optimization system should include:
Applying this approach to an omnichannel retailer, LTPlabs developed a Decision Support System (DSS) to optimize clearance pricing which allows a faster actuation on the end-of-life items.
This user-friendly life-cycle monitoring and price markdown tool allows for aggregated category views and product deep dives, which results in improvement of markdown timing and depth to achieve inventory, profit and revenue objectives.
Product managers can select their goals (margin, sellouts, etc.), monitor product lifecycle and receive alerts when inconsistencies are detected or there’s a product in need for intervention. The model also gives recommendations of the moment in time and price for product markdown and expected outcome, including what-if analysis of alternative markdown strategies.
With all this information in one place, it’s easier to develop customized markdown plans for individual stores based on store-specific customer demand. The sellout predictions turn to be more accurate and the stock clearance increased, when compared to previous seasons.
By: Nuno Pedro , João Pedro Moucho