System updates

We’re experiencing issues that may affect your Square services. We’ll continue to update our status page with more information.

Back to Home

Stock Forecast with Square for Retail

Overview

With Inventory Management on Square for Retail, you have access to automated inventory updates via Stock Forecasting. A dismissible banner in your online Square Dashboard will appear and provide you with insights into what stock is running low (based on your sales), as well as give you the ability to easily generate purchase orders when reviewing your inventory.

Getting Started

Along with the dismissible banner, you’ll also be able to view your Stock Forecast within your Item Library. Based on data associated with your Square for Retail account, this section will provide you with a list of items, grouped by vendor, that are likely to sell out in the near future.

Note: While this doesn’t provide the ability to auto generate purchase orders for all of your items, you can still select any of the items listed in Stock Forecast to generate an automated purchase order.

Stock Forecasting with Square for Retail is automated through your online Square Dashboard. At this time, there is no way to disable or turn off Stock Forecasting, but you can dismiss the feature at any time by clicking Not now on the banner.

Learn more about auto generating purchase orders with Square for Retail.

FAQs

How does Stock Forecast work?

At a high level, Square makes decisions based on two pieces of information

  • The current inventory (in stock) amount of an item
  • A sales prediction for that item

Our sales predictions forecast how many units of an item a business will sell in the next two weeks.

We compare our sales prediction to the current inventory count, and show the item on the Stock Forecast page if our sales prediction for the next two weeks exceeds the inventory for that item.

For example, you may have an item ‘XL Slim Fit Jeans’ with a current inventory of 10 items. Our sales prediction for this item in the next two weeks is 14 (1 per day). Since we forecast you will sell more than your current inventory in the next two weeks, we mark this item as running low, and display it on the Stock Forecast page and in Auto POs.

When creating Auto POs, we’ll automatically include an order quantity for each low stock item in the Auto PO. If the item has been ordered from the same vendor & location in the past, we’ll use the quantity from that PO. Otherwise we’ll fill in our sales prediction for the next two weeks.

How are sales predictions made?

We make sales predictions in two ways.

30-Day Average

The majority of items look at the average number of sales from the last 30 days, and assume the next 14 days will be similar.

Suppose you sold 40 pairs of the ‘XL Slim Fit Jeans’ in the last 30 days, and you have 10 in inventory. This means, on average you sell 40/30 = 1.33 pairs per day. To calculate your expected sales for the next 14 days, we multiply this value by 14. Our prediction in this case would be 1.33 pairs per day x 14 days = 18.66.

For this item, if you have 10 pairs in stock, we’d mark this as running low since for the next 14 days, we expect you to sell 8.66 more pairs than you have on hand.

Smarter Model

The 30-day average is very accurate for a majority of items – especially if the item is low volume (<5 sales per month) and has consistent sales. However, higher volume and less regular items are harder to predict. They may have seasonality, trends over time as your business grows or large one-time sales that will not happen in the future.

For these items, an AI based model is used that accounts for these factors. Some of the data we consider is:

  • The sales history of the item over the last two years
  • Related items (for example, hamburger sales could be similar to hotdog sales)
  • Merchant category (for example, coffee shops have similar sales patterns)

All of this information is fed into the model which outputs a more accurate sales forecast. Higher accuracy is possible due to the additional data sources, more complex model, and the fact that we can leverage sales knowledge across all of our sellers. Large retailers have plenty of data they use to forecast their own sales. Our model helps level the playing field by giving less-massive retailers the same data-advantage.

Can't find what you need?