Overview (Optimization - Markdown)
Purpose
Markdown Optimization Accelerator intends to provide recommendations of discounts for products (near expiry, end of life/season, etc.) whose inventory needs to be cleared out in a given time frame, typically of a B2C company, but not only.
This accelerator enables price optimization based on:
Stock level target (relative or absolute quantity or stock value)
Setting price limits (price change, minimum margin, position to recommended retail price)
Maximization of the revenue and/or profit
Aligning rules with competitor prices (optional)
To model the market demand, an advanced elasticity model is used and sold quantity is adjusted depending on price changes, so typically a large price decrease will end up in more sales and decreasing stock in order to reach the target. These models are intended to be updated periodically (typically weekly) to follow market response and furthermore adjust the prices.
Outputs
The outputs of the optimization model are markdown discount and prices by product, store and time period. These markdown discount and prices are displayed in the interface and can be fetched in any part of Pricefx solution and also queried by external systems through APIs.
A set of dashboards is also provided to allow you to review and assess the outputs with charts and tables with the recommended prices, markdown discount, forecasted revenue and margin, stock coverage etc.
Approach
The optimization relies on two parts:
Powerful elasticity model, built with Multifactor Elasticity Optimization, deployed at the same time as the Shelf Price Optimization
Markdown Optimization Accelerator itself
Here is an overview of the two models working together with inputs and outputs:
Â
With this data mapping in place and the model parameters properly set, most of the work is done by the Optimization Engine, which positions the prices based on the price boundaries, competition prices, stock coverage etc. in order to optimize revenue and/or margin.
For details see Optimization Engine.
One strong assumption about computing the stock is that no more delivery will happen in the upcoming time frame – this is consistent with products that have a low turn over and stock should be reduced. That also enables computation of remaining stock from last stock - sold quantity. This calculation is then used for the following period to determine the final stock target.
For products with low sales or inconsistent price elasticity, a fallback approach is implemented, using a simple elasticity aggregated at product group level. In this case, forecasted quantity is based on a reference period.
Limitations
Selection of products to markdown – This accelerator does not provide recommendations or guidelines of products to markdown. Those definitions or rules are business related. It is suggested to create a specific logic with specific rules, which stores products to markdown in a Data Source. This Data Source is then used to define the product scope.
Product scope – Relies on past transactions, so a product with no sales will not be part of the optimization. A longer time period might be defined to counter this aspect.
No predefined extension point – There is no out-of-the-box extension point defined for now. If you intend to add specific features, custom code should be written. (But then the accelerator becomes specific and it cannot be updated without extra effort to port those modifications.) Do not hesitate to report specific requirements and possible extension points to Pricefx.
Data requirements – See Data Requirements (Optimization - Markdown).
Out of scope:
Product cannibalization
Highly seasonal sensitive products (requires specific predictive models)
Promotion types recommendation (only depth)
Handling damaged, repackaged and refurbished products