About replenishment parameters based on demand forecast

Businesses strive to have enough stock for potential sales. Replenishing the stock in time and in the proper quantities helps businesses to retain customers while reducing storage costs.
For an overview of the automated replenishment functionality implemented in Visma Net, seeAbout automated replenishment.
To predict the demand, that is, the stock level you may need in the future for each item, you can use demand forecast methods.

Demand forecast method

The current version of Visma Net uses the moving average method, one of the simplest methods.

In Visma Net, you can use the predicted average daily demand, calculated with the help of demand forecasting models, to periodically update the values of replenishment parameters (such as maximum quantity, reorder point, and safety stock) in order to more precisely calculate the quantities required for replenishment.

With the moving average method, you calculate the demand for a specific future period based on historical data for multiple consecutive periods in the immediate past.

This method works well for items with specific trends in sales or with stable sales not subjected to random significant fluctuations.

To apply the moving average method to forecasting demand, you choose

  • the specific time period (a week, a month, or a quarter), and
  • the number of these periods to analyse the data, depending on how sales of the particular product or group of products (implemented in Visma Net as an item class) change over time.

This forecast model uses the data available for the specified number of past periods to forecast demand for the period immediately following the last period whose data is used in forecasting.

Data from full periods

If you select a month as the forecast period and want to perform a forecast on the 15th day of the current month, 15 days' worth of data will not be used for forecasting; only data from full periods is used. The system calculates the daily demand for the nearest future period based on daily sales in all days of the specified number of prior periods.

As the actual data for the last forecast period becomes available, you can forecast demand for the next period by using the data of the previous periods shifted by one period.

This method smooths out the data; the more periods you use to calculate the average demand, the greater the smoothing effect.

Average lead time calculation

The average lead time is calculated for each stock item‒preferred supplier pair based on all purchase orders in the system history.

The average lead time is computed as average difference (on all orders) between the date when

  • an order is created (the Requested on date) and
  • the date of the receipt created for the goods listed on the order.

Sales of some products follow certain cycles or patterns. For example, the peak of ski sales is in the winter.
Such patterns in sales can be described as seasonality.

A set of seasonality settings is a list of low and high seasons with appropriate factors (coefficients) that show how sales decline or increase in each of these specific seasons, as compared to average sales volumes calculated over all time. The seasons for particular products or groups of products with similar sales behaviour and corresponding factors can be determined by analysing sales data spanning several financial years.

Defining seasonality settings

You define seasonality settings by using theReplenishment seasonality (IN206600) window.
The seasons for one seasonality should not intersect. For date ranges not included in any season, the factor is 1.0 by default.
A low season for groups of products may span multiple financial periods or be contained in a single financial period.

The seasonality settings are used in the following procedure:

  1. The system normalises the historical data available for the specified seasons.

    The sales volumes for each day of a season are deleted by the appropriate factor to calculate the sales as though it were a normal season.
  2. The system calculates the average daily sales amount based on the historical data of the specified number of periods.

    For seasons within the specified periods, the system uses the normalised data.

    The standard deviation for the average daily demand is calculated for actual sales data, not for normalised data.

  3. The average daily sales amount is adjusted for seasons if seasons are expected within the forecast period.
    The average daily demand is multiplied by the average seasonality factor (calculated over all days in the forecast period—normal days and days of seasons).

The system uses:

  • the calculated average daily demand,
  • average lead time, and
  • their standard deviation values,

to compute the parameters below that are used in automated replenishment: Reorder point and Safety stock.

The reorder point calculation

Reorder point = (Average daily demand)*(Average lead time) + (Safety stock)

The safety stock calculation

The safety stock is calculated as follows to address possible fluctuations in demand:

Safety stock = serviceFactor * sqrt (aLeadTimeMSE * aDemandPerDay)^2 + (aLeadTime * aDemandMSE)^2

Where MSE is calculated as statistics method: Mean squared error

The Maximum stock is set equal to Reorder point; if needed, you can manually increase the value of Maximum stock.

Service level setting

The service level is used to optimise the safety stock level. Generally, it is set on the item-class level, but you can specify it for each item in each warehouse if needed.
The service level represents the probability of not reaching stockout. The Normsinv (service level) function provides the factor that adjusts the safety stock required for the specified service level. With the service level selected at 50%, the safety stock is zero (that is, not required).

Default value for service level

The factor reaches 1.0 with the service level approximately 84%, which is the default value for service level in theItem classes (IN201000) window.
To maintain a higher service level, you will need a safety stock that exceeds the calculated quantity.

You specify the forecast method and its parameters for each item class by using the Item classes (IN201000) window, and then the settings are used as default values for items of the class. You can select values that differ from the default ones for particular items of the class by using the Stock items (IN202500) window.
Moreover, you can adjust the replenishment settings for each warehouse in which the items are stocked.

You use the following Visma Net windows to update replenishment parameters and then prepare the replenishment:

  • Calculate replenishment parameters (IN508500):
    You use this window to compute the average daily demand and average lead time, and then to calculate the following parameters used in automated replenishment: Maximum quantity, Reorder point, and Safety stock.
  • Apply replenishment parameters (IN509500):
    In this window, you review the parameter values suggested by the forecast and replace the old values of replenishment parameters.
    You can manually adjust the suggested parameters if needed.
  • Prepare replenishment (IN508000):
    Use this window to find the items that need replenishment and to calculate replenishment quantities for these items.
  • Create purchase orders (PO505000):
    To generate purchase orders for items requiring replenishment, you can use this window.

Related concepts

Manage replenishment - overview

Related windows

Calculate replenishment parameters (IN508500)

Apply replenishment parameters (IN509500)

Prepare replenishment (IN508000)