What is demand planning?

• Record stock transfer, stock take and stock wastage to identify stock availability
• Calculate item quantity sold and modifier quantity sold
• Map stock to item based on UOM and stock to item name
• Calculate theoretical vs actual stock usage
• Calculate reorder point and quantity to reduce the need for manual determination
• Calculate stock turnover rate to determine effective stock holding
• Identify over and under stock that have adverse impact on cost or revenue
• Plan for how much is needed in coming week and months to determine how much to produce or purchase

Steps of demand planning in VAL

Single-day-stock model

Suitable for stock made for the day and sell for the day scenario such as pastry, bread, etc.

1. Load and collect data
2. Calculate individual sku sales by day of the week and allow for manual visualisation and manual planning
3. Calculate wastage and identify over and uner stock that have adverse impact on cost or revenue

Carry-stock model

Suitable for stock that is used over multiple days period such as ice cream, salad, etc

1. Load and collect data
2. Calculate stock turn over to track how fast stocks need replenishment
3. Calculate reorder point and quantity to determine when and how much to reorder
4. Calculate theoretical vs actual stock to determine variance of planning

Forecasting

Forecasting the sales of individual SKU using forecasting model as below

ES (Exponential Smoothing):

What it does: This model predicts future sales by averaging past sales data, giving more weight to recent data points. It adjusts quickly to changes in sales trends.

Used for: Items with stable sales patterns and moderate fluctuations, like everyday grocery items or popular beverages.

TBATS (Trigonometric, Box-Cox Transformation, ARMA Errors, Trend, and Seasonal Components):

What it does: This complex model handles multiple seasonal patterns and trends, making it good for items with various periodic sales fluctuations, such as weekly, monthly, or yearly.

Used for: Seasonal items with complex demand patterns, like holiday-specific products or fashion items that change with seasons.

ETS (Error, Trend, Seasonal):

What it does: This model breaks down sales data into error, trend, and seasonal components to make predictions. It’s good for items where these components are prominent.

Used for: Items with clear trends and seasonal effects, like ice cream in the summer or hot drinks in the winter.

STL (Seasonal-Trend Decomposition Procedure Based on Loess):

What it does: This model decomposes sales data into seasonal, trend, and residual (random) components using local regression. It’s flexible and adapts to various seasonal patterns.

Used for: Items with strong and consistent seasonal patterns, such as fresh produce that’s seasonally available.

ARIMA (AutoRegressive Integrated Moving Average):

What it does: This model uses past sales data and its relationship over time to make predictions. It’s good for items with clear historical patterns without strong seasonality.

Used for: Non-seasonal items with consistent demand over time, like household essentials or staple foods.

SARIMA (Seasonal ARIMA):

What it does: This extends ARIMA by adding seasonal components, making it suitable for items with repeating seasonal patterns in their sales data.

Used for: Seasonal items with predictable sales cycles, such as holiday decorations or seasonal clothing.

1. Load and collect data

In VAL, we have both direct connectivity as well as pre-built report reader for the commonly used POS and payment methods used by our customers.

For direct connectivity, users do not need to perform any task as VAL will automatically pull data from source system.

For pre-built report reader, users will have to download the report from source system and drop the files in VAL Drive. VAL Drive will read the files and use the appropriate report reader to read the data and upload data into VAL.

Single-day-stock Model

2. Calculate individual SKU sales

We provide information to users to allow them to analyse their sales so they are able to take on manual demand planning in a much more efficient and eliminate manual aggregation and cleaning of data.

3. Calculate individual SKU wastage and identify over and under selling

We provide information to users to allow them to analyse their wastage so they are able to monitor the effectiveness of their demand planning and identify scenario of understocking based on the last sales timing.

Carry-stock Model

2. Stock turn-over

The weekly stock turnover rate tells users how often their inventory is sold and replaced within a week. It helps businesses understand how quickly their products are moving. A high turnover rate means items are selling fast, indicating strong demand or effective sales strategies. Conversely, a low turnover rate suggests that products are sitting on shelves for too long, which could mean overstocking, low demand, or pricing issues. By monitoring this rate, businesses can make informed decisions about purchasing, pricing, and inventory management. It helps them avoid stockouts (running out of stock) or overstock (having too much inventory), both of which can impact profits. In essence, knowing the weekly stock turnover rate helps businesses optimize their inventory levels, ensuring they have the right products available at the right time to meet customer demand.

3. Reorder point

Reorder Point: This is the specific inventory level at which new stock should be ordered to avoid running out. It considers the time it takes to receive new stock (lead time) and the average sales during that period. When inventory hits this level, it signals the need to place an order.

Reorder Quantity: This is the amount of stock to order once the reorder point is reached. It ensures that the new stock will meet demand until the next reorder, minimizing the risk of stockouts and overstocking.

Together, these metrics help businesses maintain optimal inventory levels, ensuring products are available when customers need them without tying up too much money in excess stock. This balance improves customer satisfaction by preventing stockouts and increases efficiency by reducing storage costs and waste from unsold items.

4. Actual vs Theoretical

Actual vs. Theoretical Stock Usage compares the amount of stock actually used to the amount that should have been used based on recipes or production plans.

Actual Usage: This is the real amount of stock consumed in a given period. It includes all losses, waste, and usage not accounted for by planned recipes or production processes.

Theoretical Usage: This is the expected amount of stock that should have been used based on sales, recipes, or production plans. It assumes everything is used efficiently and as planned, with no waste.

Comparing the two helps users identify discrepancies and inefficiencies. If actual usage is higher than theoretical, it indicates issues like wastage, theft, or over-portioning. If actual usage is lower, it may indicate under-portioning or errors in tracking.

This comparison helps businesses improve accuracy in inventory management, reduce costs by minimizing waste, and enhance overall efficiency in stock usage. It provides insights into operational performance and helps in making informed decisions to optimize stock control.

Forecasting

With multuple forecasting models available us, we take into consideration seasonality and market factors and the nature of your business to evaluate what is appropriate for each of your SKU, at the same time allowing you to evaluate it against your manual forecasting.