Version 1, October 2, 2026. Updated October 3, 2026, to describe measured lead times and when seasonality applies. This page describes exactly how Posto works out what to reorder and how much. When the math changes, we'll update this page and note what changed and when.
Why we publish this
Most inventory apps ask you to trust a number. We'd rather show you how we got it. Posto uses standard inventory math, the same formulas taught in operations courses, not an AI model. Every recommendation in the app has a "Why this quantity?" breakdown, and this page explains what sits behind it.
The settings, and Posto's defaults
- Service level: 95%. How often you want to avoid running out while waiting on an order.
- Sales history: up to 90 days.
- Days of cover: 30. How long an order should last after it arrives.
- Review period: 7 days. How often you check and place orders.
- Lead time: set per supplier, and per product if one is slower. Once 3 orders from a supplier have arrived within the past year, Posto uses the average of how long they actually took instead, and labels it "measured." A lead time set on a product still wins. If a product has several suppliers and none is chosen as primary, Posto uses 7 days until you pick one.
- Black Friday and Cyber Monday: 3 times normal demand, from Thanksgiving to Cyber Monday.
- Holiday peak: 1.6 times normal demand, from the day after Cyber Monday (or December 1) to December 23.
You can change any of these for your whole store, or for a single product.
Step 1: how much you really sell
Posto starts with your average daily sales for each product, but it's careful about which days count.
- Only days Posto can read. Shopify lets apps read the last 60 days of orders unless an app has access to your full order history. Posto never treats days it can't see as days with no sales.
- Only days the product was for sale. Counting starts when the product became available, so a new product isn't made to look slower than it is.
- Not days you were sold out. Posto checks your stock every day. Days with no sales because you had nothing to sell are left out, so a stockout doesn't shrink your next order. A day with any sales always counts.
- New products are flagged. A forecast built on fewer than 8 weeks of history is marked "Limited history," and under 2 weeks as a rough estimate.
Step 2: looking ahead
You order stock to sell over the coming weeks, not the last ones, so Posto projects demand forward.
- Seasonality. With enough history, Posto learns how each month compares with a normal month. An index of 1.3 means a month usually runs about 30% above normal. With thin history, it leans toward "no seasonality" rather than guessing, and it caps unusually big single days, like one bulk order, so they don't distort a month. Shopify lets Posto read about the last 60 days of orders, which isn't enough to learn a month-by-month pattern, so today this stays close to neutral and the events below do the seasonal work.
- Trend. If a product is growing or slowing, Posto fits a line through its recent sales and carries it forward, within sensible limits.
- Events. Black Friday, Cyber Monday, and the holiday peak are added on top. Last year's event days are taken out first, so they don't inflate your everyday average and get counted twice.
Step 3: when to reorder
Reorder point = expected sales over (lead time + review period) + safety stock
The review period matters because you don't place an order the moment stock dips. If you check weekly, your stock has to last until the next check, plus the delivery time.
Safety stock = Z × √(days covered × daily sales variability² + daily sales² × lead time variability²)
Z comes from your service level: 1.645 at 95%, and 2.326 at 99%. Higher means fewer stockouts and more cash on shelves. The reorder point is rounded up to a whole unit, so Posto errs toward ordering a touch early, not late.
Posto compares the reorder point with what you have on hand plus what's already on order, so it never asks you to reorder stock that's on the way.
Slow sellers get extra care. For products that sell a few units at a time and unevenly, the usual formula can leave too thin a buffer. Posto checks a model built for small, lumpy demand and uses it when it calls for more stock, never less.
Step 4: how much to order
Order-up-to level = expected sales over (lead time + days of cover) + safety stock
Order quantity = order-up-to level − on hand − on order
- Never zero when it matters. If a product is at or below its reorder point, Posto always recommends at least one unit.
- Oversold stock is made up. If you've sold more than you had, the shortfall is added to the order.
- Optional order sizing. If you enter what it costs to place an order with a supplier and what it costs to hold a unit for a year, Posto uses the economic order quantity as a minimum, to combine small, frequent orders into fewer, cheaper ones. It never makes an order smaller.
- Supplier rules last. The quantity is then rounded up to the supplier's minimum and case pack.
A worked example
An oat body wash sells about 6 a day, the supplier takes 7 days, and you review weekly, so the reorder point covers 14 days. Daily sales swing by about 3.4 units. You have 63 on hand and nothing on order. This example is illustrative.
- Expected sales over 14 days: 6 × 14 = 84.
- Safety stock at 95%: 1.645 × 3.4 × √14 ≈ 21.
- Reorder point: 84 + 21 = 105. With 63 on hand, it's time to order.
- Order-up-to level: 6 × (7 + 30) + 21 = 243.
- Order: 243 − 63 = 180. If the supplier sells in cases of 12, 180 is already 15 full cases.
Sales at risk
For each product, Posto estimates what you'd fail to sell if you ordered today: the units you'd run short before the delivery arrives, valued at the current price. It assumes stock already on order lands in time and uses list price, so the estimate errs on the low side. Recommendations are sorted by urgency, then by sales at risk.
Landed cost
When you add freight, duty, and fees to a purchase order, Posto spreads them across the order's lines, by value or by units, your choice, to give a landed cost per unit. While deliveries are still coming, costs are split over what you ordered, and once the order is finished, over what actually arrived. Our landed cost guide walks through an example.
Check it yourself
Pick a product you know well and compare its "Why this quantity?" breakdown with your own numbers. Our reorder point guide shows the same math step by step. If something doesn't add up, tell us through the feedback form. We read everything.