Know what sells.
Before it sells out.

Prioricast tells you what to order, when to order it, and where to send it — powered by forecasts validated against your actual sales.

Retailers lose $1.7 trillion a year to running out and overstocking. Two-thirds of it is empty shelves.IHL Group, Inventory Distortion Study, 2026

Prioricast · Scented candle, Cedar
The Prioricast product page for a scented candle: order 354 by Dec 10, 305 in stock, runs out Dec 23, with the sales history and the forecast range on a chart.

The real product page, from an example store.

Every product gets the forecast its history can support.

A bestseller and a beanie you stocked last month shouldn’t be forecast the same way. Prioricast reads each product’s history and picks the method it can support.

Pick a product to see the method it gets.

Forecast method · pick a product

Tested on your past before it plans your future.

Before recommending a single order, Prioricast runs a backtest on every product’s own history.

  1. 1
    Replay the past

    It rewinds to a past date, forecasts the next six weeks using only the sales it had seen, and compares that with what really sold. Then it steps forward a month and repeats.

  2. 2
    Tune the safety stock

    It replays your sales at five safety stock levels and keeps the one with the lowest total cost: lost sales plus the cash tied up in extra stock.

  3. 3
    Score every week, live

    Each Sunday it saves next week’s forecast and a range that should catch 8 weeks in 10, then checks it against what sold. Misses included.

You see the result for your own store on the History test page.

Backtest · five replays of one product
JanFebMarAprMayJunJulAugSepOctReplay 1off by 11%Replay 2off by 8%Replay 3off by 15%Replay 4off by 9%Replay 5off by 12%Average error across the replays11%
Sales it was allowed to seeSix weeks it forecast, scored against real sales

One product, example data.

Safety stock · five levels replayed
Safety stock levelLost salesExtra stockTotal cost
Light 90%$1,840$310$2,150
Standard 95%$960$470$1,430
High 98%$420$690$1,110
Very high 99%$260$880$1,140
Maximum 99.5%$170$1,060$1,230
Profit lost to sell-outs, per yearCost of holding the extra stock, per year

Example data. Picked: High, the lowest total cost.

Weekly score · last eight weeks
20304050Sep 6Sep 13Sep 20Sep 27Oct 4Oct 11Oct 18now
Expected rangeSold within the rangeSold moreSold less

One product, week by week. Example data.

Real Prioricast screens, from an example store.

Built for how stores really sell.

Real-world messiness, already handled in every forecast.

Days it sold outDays your shelf was empty are left out, so a sold-out day doesn’t read as low demand.
A product page for Linen throw blanket, Oat. The cursor opens “Why this order”, which lists 9 out-of-stock days left out, so empty shelves don’t read as low demand.
HolidaysIf a holiday lifts a product’s sales, Prioricast plans for it on the right date each year.
The cursor opens Holidays for the same product. Thanksgiving to Cyber Monday is in this forecast, busiest on Saturday, Nov 28.
One-off rushesA one-time spike, like a viral week, won’t inflate your future orders.
Knit scarf, Navy. “Why this order” lists a one-off rush left out, so the spike won’t inflate orders.
Sales and discountsRan a sale this time last year? Prioricast asks if it’s running again and shows the order both ways. It also learns how much a discount lifts each product.
Ceramic mug, Speckled. A card asks whether last year’s Black Friday sale is running again: order 337 without it, or 370 with it. The cursor opens Price sensitivity, which shows 10% off lifts sales 41%, 20% off 106% and 30% off 217%, measured in its busy season.

From purchase order to shelf.

Everything for a supplier lands on one purchase order. When the delivery shows up in your stock, Prioricast checks it in.

Whole cases, minimums met

Rounded up to the case pack and checked against the supplier’s minimum.

Split by store

With more than one location, it shows how much goes to each.

Ready to send

Paste it into a chat, or download a PDF.

Checks itself in

Part deliveries stay owed, and every delivery sharpens the lead time.

Orders · ready to send

Harbor Textile Co.

Sea, lands around Jan 15
ProductLocationCasesUnits
Linen throw blanket, OatDue Nov 25Midtown1484
Linen throw blanket, OatDue 4 days later, added earlyWestside636
Linen throw blanket, SlateDue 6 days later, added earlyMidtown530
Waffle bath towel, RustDue 2 days later, added earlyMidtown448
Order total $3,486. Minimum of $1,500 met.
2To orderdue soon
1To checkneeds a quick answer
4On the waycounted from arrival
27Receivedchecked in on its own

Plug in your store. Then forget about it.

Connect Square or Shopify, or upload a spreadsheet. After that, it runs on its own.

SquarePaste your access token. Sales, stock and costs come in for every location, and item vendors become your suppliers.
The Prioricast connect screen: Square
ShopifyEnter your store name and an Admin API token. Orders and stock come in, and product vendors become your suppliers.
The Prioricast connect screen: Shopify
A spreadsheetNo store to connect? Drop in an Excel or CSV of your sales. Add a location column if you have several shops.
The Prioricast connect screen: A spreadsheet

Read-only. It asks only for permission to read orders, products and stock, so it can’t change anything in your store. Your token stays on the server, never in the browser.

  • Every nightPulls in yesterday’s sales and updates every forecast.
  • Every hourChecks stock, sales and prices.
  • Every SundaySaves next week’s forecast, then grades it.

Order with confidence.

Send us your sales history. We’ll show you how accurate Prioricast would have been on your own products, before you commit to anything.

We load your sales and run the history test on every product. You see how accurate it would have been, then decide.

Which systems does it work with?

Square and Shopify connect directly. Anything else works with a spreadsheet upload: .xlsx, .xls or .csv, many products in one file.

Can it change anything in my store?

No. It only asks to read orders, products and stock. It can’t edit products, prices or orders.

How much sales history does it need?

Any amount works, and more lets it learn more. Weekly patterns show up quickly, yearly patterns need about 15 months, and the history test needs about six. A new product borrows from its other locations until its own history builds.

Does it handle more than one location?

Yes. Each location gets its own forecast, and one purchase order shows how much goes to each.

What happens after I sign up?

We load your sales and run the history test on every product. You see how accurate it would have been on your own store before you commit to anything.