Dynamic Replenishment Model: Key Calculations & Formulas

Center of Excellence

1. Abstract

Retail replenishment is the process of counting the inventory you have left in stock and reordering the right amount of stock at the right time.

In other words, it's the process of ensuring that you always have sufficient inventory in store to meet customer demand.

2. Content

As part of the standard Board Replenishment engine, Board will calculate the Target Stock Quantity (IAQ), required in each store for each SKU (Product at size level).

The standard input parameters, technical requirements, and calculations are defined below.

3. Context

3.1 Target Stock Level

The target stock level is the ideal stock required in each store to ensure the business can meet expected consumer demand and ensure they have enough stock in store to react to peaks in demand.

Data Source	Calculation
Dimensions	SKU, Site
Board Technical Translation	Standard Algorithm: ROUND(MIN(MAX(ROS*Weeks Cover), Min Units), Max Units) * Ranging

3.2 Input Parameters

3.2.1 Rate of Sale (ROS)

Data Source	Historic Sales Actuals
Dimensions	SKU, Site, Week
Board Technical Translation	Required: Aggregate to Week Filter On ‘Sales Weeks Included’ Optional: Data Cleanse (Smooth Sales Peaks and Troughs) Translate to Forecast (BEAM)

Rate of Sale (ROS) is the average number of units per week sold over a selected period of time.

The time period can include both historic actuals and forecast, or just historic actuals, and is controlled using the weeks included parameter (see the section below).

These sales form a baseline demand, which can be surfaced to the user in a review screen, where they can adjust and assess the baseline demand (see screenshot above).

The baseline demand can also be cleansed through the Board Beam engine, accounting for out-of-stock weeks and forecasted to account for expected uplift/events.

Once a baseline demand has been set at SKU/Store/Week for the weeks included, Board will then aggregate those sales by SKU/Store and divide by Weeks in Stock to calculate the average ROS to utilized in the Target Stock Quantity calculation.

Weeks in Stock is the count of weeks where the product was available to be sold in store within the weeks included.

3.2.2 Min & Max Stock Level

Data Source	User Input
Dimensions	Option/Category/Department/Division, Grade/Region, Size
Board Technical Translation	Required: Push Down/Convert Input to SKU/Store Level Optional: Input by Size or Apply Size Curve after pushing down to option

Min & Max quantity is utilized to control the stock levels in store.

The max quantity is often utilized to control capacity and ensure stores do not receive more stock than they can hold.

The minimum quantity is used to ensure coverage and that peripheral stores and sizes are allocated and stock is not wholly consumed by large flagship stores.

In Board these quantities are usually managed at a high level, allowing the user to push down through a procedure to lower levels of the product hierarchy and manage or overwrite by exception.

Sales Weeks Included

Data Source	User Input
Dimensions	Division/Department, Replen Week (Non-Time Entity), Week (Time Entity), Region.
Board Technical Translation	Required: Convert Replen Week to Actual Weeks on time entity

Sales Weeks included is the parameter to define the time width of sales you want to include in your Rate of Sales (ROS) calculation.

This is typically defined at a high level (Division/Category, Region/Global).

The “Weeks Included” can include both historical weeks and forecast weeks.

The technical task for this parameter is convert the user input; the non-time entity is translated dynamically to the correct weeks before each replenishment batch run.

For example, a user might define for Ladieswear they want to include the last 3 weeks of history and 2 weeks of forecast, the actual weeks included will need to be reset at the start of each week.

The standard board functionality to achieve this in a dataflow is to use the Time Offset feature, with a relative offset based on the number of weeks included.

3.2.3 Target Weeks Cover

Data Source	User Input
Dimensions	Division/Department, Store Grade/Region.
Board Technical Translation	Required: Push Down/Convert Input to Store Level

The target weeks cover is utilized to allow the user to define how much sales cover they want in store. Sales cover is the number of weeks a store can meet customer demand with the available stock in store.

A store will require sufficient cover to ensure that sales will not be impacted by supply chain timelines (time it takes to send stock from warehouse to store) or unforeseen peaks in demand.

Weeks cover is normally maintained at a high level of aggregation on both the product and site tree.

3.2.4 Store Ranging

Data Source	User Input
Dimensions	Option, Store
Board Technical Translation	Required: NA

Ranging is a user-controlled matrix between Store and Option, where the user is defining which stores they want the product to be sold in.

We utilize the store ranging in the replenishment process to ensure we only calculate the target stock level (IAQ) for the relevant stores per option.

Center of Excellence

Thanks @Guy Pearson for the insights on this topic!