BI Forecasted Inventory - Order Method
1. BI Forecasted Inventory - Order Method Overview
The BI Forecasted Inventory order method is a more advanced way of predicting inventory levels. It combines the "Days of Inventory" order method with additional rules that help make the forecast more accurate and reliable. The main purpose of additional rules is to prevent the sales figures from becoming skewed by identifying and adjusting/replacing the bad days.
2. Understanding the Prediction Process
The prediction process includes the key points outlined below:
- When calculating the predictions, the system starts by looking at the first day (tomorrow) and predicting each day until 30 days ahead.
- The predictions are calculated up to 30 days ahead to cover the various types of ordering our system produces. This can handle weekly, biweekly, and monthly orders using the 30 days we’ve predicted for each item.
- Each day gets evaluated using the last 45 days of sales history from the date being evaluated.
- When calculating predictions, the parent and child are combined in a rolled-up format. The child's details are factored into the parent's prediction outcome.
- Each day is calculated separately, allowing you to choose the days of inventory you wish to have in your order. As we get closer to a future predicted date, the data becomes more accurate. The data will regenerate the predictions with more fine-tuned trend data.
3. Identifying Bad Days
When we calculate the predictions, in each of the last 45 days prior to the date being evaluated, we determine if that day's sales will be included as is in the average sales figure. The dates are classified into normal and bad days.
• A normal day: We use the actual quantity sold without any adjustments.• A bad day: We replace the expected sales figures with a more realistic estimate.
Identifying "Bad Days" involves detecting instances when data points deviate from normal. The bad days can be identified when one of the criteria below is fulfilled:
1. Out-Of-Stock:
These are days when the product is out of stock, leading to a false representation of the demand for that period.
2. Abnormal Sales / Anomaly Days
- Low Sales: These are days when product sales are abnormally low. This is often due to factors like bad weather, holidays, or external events.
- High Sales: These are days when sales are abnormally high. This is due to situations such as a customer purchasing a large volume of the product randomly or potentially from an in-store deal or promotion.
4. Adjusting Sales Figures for Each Bad Day
Each Bad Day is replaced with a more realistic sales figure using localized sales occurring around the date being evaluated.
- Out-of-Stock: If an item is Out-of-Stock on a given day, we generally want to remove this from our sample data or replace it with another sales figure. This is because without stock, sales cannot happen, causing our figures to be skewed. We input another figure, assuming this day had stock to help create better numbers.
- Anomaly Day: Like Out-of-Stock days, a day where the sales are abnormally low or high can impact forecasting. If the sales do not match what a regular day would look like, we go ahead and replace the sales figures for those days.
4.1. Example for Adjusting Sales Figures for the Out-of-Stock and Anomaly Days
If we do not find 5 recent days, then we use as many days
as we find within the 30-day period and average out those days. This could be anywhere
from 0-5 days in
total getting averaged.
See the example below, using the date 01/31/2025 as a Bad Day being replaced:
See the example below, using the date 01/31/2025 as a Bad Day being replaced:
No – We already picked the 5 most recent days from 01/05 – 01/30.
Yes – Used in Average Calculation
- In the above example, we found 5 sale days in the last 30-day period (01/01/2025–01/31/2025).
- Let’s say sales didn’t occur on 01/02, 01/05, and 01/10; then we have the sales only on 01/15, 01/21, and 01/25. In this case, we will consider only these 3 days.
- In the above example, we will use the
quantities marked in bold for the calculation. The average is found to be 3.4. This value will be used as the projected sales figure for the day.
5. Adjusting Fluctuations in Forecast vs Demand
We will slightly increase or decrease the number
predicted if we detect we have over- or under-ordered the previous day. This
usually helps pick up local sales trends around the date (short-term).
Similarly, if we predicted 8 quantities
but the next day, we calculated 10 were sold. In this case, we underpredicted
the value by a slight amount. Hence, we will adjust the upcoming days to
account for this figure.
6. Creating an Order with BI Forecasted Inventory Method
The process of placing an order with BI Forecasting is very straightforward. There is only one area that is unique to the BI setup, and the other is choosing item filters.
6.1. BI Forecast Settings
Configuring the order involves a few options, nothing overly complex compared to other ordering methods in the system. The workflow is straightforward, and most of the features work out of the box.
Figure 6.1.1
See Figure 6.1.1 above for reference on how to define some basic configurations. The options
for the BI Forecasted Inventory Order Method can be filled out as defined below:
- Days of Inventory: This is the field that lets you define how many days you want to keep on hand in the system.
It's recommended to have at least 2-3 days of excess stock for unexpected sales spikes, ensuring the buffer of days aligns with the next shipment time (for example, 1 week plus a few extra days).
- Use Item Min Qty: This field being toggled on will ensure an item’s minimum quantity set will be kept on hand, regardless of sales.
- Require Multiples of: This field
will force the order to round up to the specified quantity in the field.
For Example: If the order requests 1 to be ordered, it will round up to 2 for the order.
6.2. Order Filters
Other than the BI Setting, the other important
part to select is Order Filters.
Figure 6.2.1
When the "Custom" button is selected, the
screen highlighted in Figure 6.2.2 is displayed.
Figure
6.2.2
From here, you can choose All Items or
set up filters to limit specific "Departments, Categories, Manufacturers, Product Tags, or Department
Tags."
Once you’ve selected the
necessary filters and adjusted the settings, the order can be created now, and
it should pick up prediction-based quantities for all your suggested values.
6.3. Useful BI Forecast Fields on the "Edit Purchase Order" Screen
When an order
is created, all the order items are logged onto the order, and they can be seen
on the "Edit Purchase
Order" screen.
Each item row
in the order includes fields that allow you to refine or adjust the order. You
can also review the values, accept them as they are, optionally add new items
from custom requests, or update the item values based on the metrics shown in
the grid.
See Figure
6.3.1 below for reference.
Figure 6.3.1
Highlighted Fields’
Description:
- 7d Pred: It shows the sum of the next 7 days of predictions from the order date.
- 14d Pred: It shows the sum of the
next 14 days of predictions from the order date.
- 30d Pred: It shows the sum of the
next 30 days of predictions from the order date.
A preview of
these columns can be seen in Figure 6.3.2 below:
Figure 6.3.2
6.4. Showing "30 Days Forecast"
Graph
- On the same Edit Purchase Order screen, on the order grid, you can right-click any item to expose the context menu.
- In the context menu, there is an option to Show 30-Day Forecast, as highlighted and
displayed in Figure 6.4.1.
Figure 6.4.1
- If you click on this option, you will see a screen highlighted in Figure 6.4.2 below:
Figure 6.4.2
Here are some highlights of the graph:
See Figure 6.4.2 above.
- You will see the 7-day,
14-day, and 30-day totals for the highlighted product, as well as a
visual graph (blue line) and trend line (red
line) to see more useful
information about the forecasted sales.
- 30 dots in the graph represent 30 days.
If you click on the dot on the graph, it shows the currently opened item’s
prediction for that date along with the predicted quantities. See Figure 6.4.3 below.
Figure 6.4.3
- It can be seen for each day in the forecasted
30-day period counting from the order’s start date.
You may also see this same information on
the Products page (explained below in section 7: Viewing Forecast from the "Edit Product" page), which lets you look at this information across all stores or by individual
store.
6.5. Understanding Order Logs
On the order logs, which can be accessed via the Context Menu -> View Logs, you can see different details that help analyze the decisions made by the predictions.
Figure 6.5.1
The
fields are described
as follows:
Figure 6.5.2
7. Viewing
Forecast from the “Edit
Product" Page
- On the Edit Product page, we can see the “View Forecast” button on the "Sales Statistics" section on the right sidebar. This lets you access the multi-store or single-store view, like what is shown on the order edit page.
Figure 7.1
- On the product page, by default, the information is for All Locations, which is sometimes useful for seeing forecasts across every store in the account. The forecast for that product for each day is displayed as highlighted in Figure 7.2.
Figure 7.2
-
You can change the
location directly from the Choose Location field to browse that store’s
information.
Related Articles
Cycle Count Coverage Report
Cycle Count Coverage Report Overview Reports -> Inventory Reports -> "Cycle Count Coverage" The store admin/executive manages the inventory of the store using a POS system handheld device to count the inventory periodically. At the time of counting ...Order Requirements
What's the Purpose of Order Requirements? Typically, the order requirements of a retail business are based on customers’ demand, with objectives to manage its inventory of goods, reduce costs, and increase overall retail efficiency and profit. But ...Inventory Definitions
Inventory Definitions Overview "Inventory Definitions" lets you create definitions for stock verification processes such as cycle counts and spot checks. "Inventory Definitions" is available to corporate or store employees with relevant rights. They ...Inventory Adjustments
Inventory Adjustments Reports Overview The "Inventory Adjustments" report type will display the list of products for which adjustments were made during a specifically selected date range (time period). Various filters can also be configured, ...Inventory Settings
Inventory Settings Overview "Inventory Settings" lets you configure settings that are needed to manage any inventory-related process. You can configure settings related to invoicing, receiving, and alerts regarding product stock and inventory. 1. ...