How does Nova decide what you should order?
Nova recommends how many units to order today so that, once they arrive after your lead time, you still hold your target stock. It compares your sales over that horizon with your whole pipeline: at Amazon, on the way, at your warehouse and on order. Every number has a default from day one; the ones you type replace it.
Checked against Nova's code on October 29, 2026
All docs
What do you give Nova?
On the Supply chain page (Data Inputs, Supply chain, workspace owners only) or through your AI assistant:
| Input | Meaning |
|---|---|
| Production | Days from your purchase order to the goods leaving the factory |
| Transit to warehouse | Days from the factory to your warehouse or 3PL. 0 when goods ship straight to Amazon |
| Transit to Amazon | Days from your warehouse to sellable at Amazon, receiving included |
| Target stock at Amazon | Days of supply you want to hold at Amazon |
| Target stock at warehouse / 3PL | Days of supply you want to hold at your warehouse |
| Case pack | The order is rounded up to a multiple of it |
| At warehouse / 3PL | Units you hold, with the count date |
| On order | Your open supplier orders, units and expected date |
Nova finds your warehouses itself, from the shippers of your shipment plans over the last twelve months. A factory shipping straight to Amazon gets its own line.
Where does each number come from?
For each product and stock location, each number is picked separately, and the first one found wins:
- The product's own number for that location.
- The product's own number for all locations.
- Your number for that warehouse.
- Your number for the account (where the production lead time lives).
- Nova's fleet default, shown in gold.
Fleet defaults
| Number | Default |
|---|---|
| Production | 45 days |
| Transit to Amazon | 14 days (EU, US, Mexico, across a border), 12 (Canada), 10 (UK) |
| Transit to warehouse | 40 days by sea from Asia to North America, 45 to Europe, 10 by air or truck, 5 within the territory, 30 when no factory was read |
| Target stock | 45 days at Amazon and 30 at the warehouse; straight from the factory, 60 at Amazon and 0 at the warehouse |
What is the formula?
Horizon = production + transit to warehouse + transit to Amazon + target at warehouse + target at Amazon.
Pipeline = Sellable + Reserved + Inbound units + AWD stock + Inbound to AWD + At warehouse / 3PL + On order. Unfulfillable and Researching units are left out.
Units to order = Daily units sold × horizon − pipeline, never below 0, rounded up to the case pack.
Order by = today + (pipeline ÷ Daily units sold − horizon). It reads Now today and Late N d once passed. It is the day to place the order, not Amazon's Recommended ship date.
When one warehouse serves several stock locations, Nova adds their sales and their stock and gives one recommendation for the group.
When is there no recommendation?
In this order:
| # | Situation | Units to order |
|---|---|---|
| 1 | The row is the US stock offered in Canada, Mexico or Brazil | None: ordering for the US covers it |
| 2 | No sales in the last 30 days | Blank |
| 3 | A lead time or target resolves to nothing | Blank |
| 4 | Otherwise | The formula, 0 when nothing is needed |
Worked example
Daily units sold 10. Production 45, transit to warehouse 40, transit to Amazon 14, targets 45 and 30: horizon 174 days, so 1,740 units needed. Pipeline: 540 on and toward Amazon, 960 at your warehouse, nothing on order, 1,500 in all.
Units to order = 1,740 − 1,500 = 240, rounded to a case pack of 48: 240. Order by: 150 − 174 = −24, so Late 24 d.
What can your AI assistant import?
Writes need the workspace owner and an OAuth connection, and run in two steps: a preview with a diff, then a commit after you confirm.
| Tool | What it writes |
|---|---|
import_supply_chain | Product lines: count at the warehouse, supplier orders, case pack, the five numbers. Up to 2,000 rows per call |
set_stock_defaults | Your warehouse and account numbers, and factory lines |
import_stock_params, import_external_stock, import_supplier_orders | Lead times and targets, counts, orders. Older tools, still working, replaced by import_supply_chain |
To read before writing: get_supply_chain_setup, get_supply_chain_products and get_stock_setup. The recommendation itself is in query_inventory. Data Share's supply_chain table holds what you typed, not the computed figures.
Ask your AI
Which products should I reorder this week, how many units each, and which of them rest on Nova's default lead times instead of mine?
Something missing or unclear? Write to contact@novadata.io. New to Nova? Start a free trial.