Inventory forecasting estimates how much of each product will sell in each store and channel, so a brand can order the right quantity before stock runs out. PACX forecasts demand by product, size, store and channel, and turns the forecast into what to order, how much and by when.
Last reviewed
How does a forecast become an order?
A forecast by itself is a chart. The order is the forecast, less the stock you have, less the stock on the way.
An example with sample data, for one kurta in all sizes and channels. The need for the next 4 weeks is 4,500 pieces. There are 860 in stock and 1,200 on the way. The order is 2,440 pieces, to be placed by 12 November.
What does the forecast learn from?
PACX uses a foundation forecasting model trained on retail demand patterns across many products and seasons, then tuned to your own sales.
Sales by product and size
Stores and channels
The festival calendar
Offers and price changes
Days when a product was sold out
How accurate is it?
Accuracy depends on the category, the channel and the data. PACX publishes one measured result: on one live deployment in footwear and apparel, the method reached 82 percent forecast accuracy at the item × store level on a monthly horizon. Accuracy here is one minus the weighted absolute percentage error, measured on a period the model had not seen.
Other categories, channels and data will give different numbers. The white paper explains the method and its limits.
What can you do after the forecast?
PACX has working screens for forecasting, allocation and ordering. Your team reviews each recommendation and approves it.
Reorder a product before it runs out
Send the right quantity to each store
Move stock from a store where it does not sell to a store or channel where it does
See money stuck in stock by age: 90 to 120 days, 120 to 180 days, and more than 180 days
Who is it for?
PACX is for retail and e-commerce brands that sell on many channels at once: their own website, marketplaces, quick commerce apps and stores.
What is the difference between demand forecasting and inventory forecasting?
Demand forecasting estimates what customers will buy. Inventory forecasting takes that estimate, takes off the stock in hand and the stock on the way, and gives the quantity to order and the date to order by.
How does PACX forecast demand?
PACX uses a foundation forecasting model trained on retail demand patterns across many products and seasons, then tuned to your own sales by product, size, store and channel. It allows for festivals, offers, price changes and sold-out days, and turns the forecast into what to order, how much and by when.
What data does a forecast need?
Typical inputs are dated sales, stock positions, product attributes, locations, prices, promotions, receipts and a retail calendar. The exact minimum depends on the decision the forecast supports.
Does PACX place orders by itself?
No. PACX recommends the order and your team approves it. PACX only reads your data. It never changes your orders, listings or ads.
How much does PACX cost?
PACX is quote-based. The price depends on the decisions you want help with, the systems to connect, your number of products, stores and channels, and the support you need.