> ## Documentation Index
> Fetch the complete documentation index at: https://docs.stockful.app/llms.txt
> Use this file to discover all available pages before exploring further.

# Production

> Batches: what you set out to make, what you actually made, and the yield variance between them.

A **batch** is one production run. Open **Production** from the Raw materials hub and **Start a batch**.

Choose what you are making and how many, and Stockful walks its bill of materials to show exactly what it will take, at the quantity you asked for.

## The three states

|                 |                                                                                                  |
| --------------- | ------------------------------------------------------------------------------------------------ |
| **Draft**       | Nothing is reserved and nothing has moved. Fully editable, and deletable.                        |
| **In progress** | Material is reserved. It is still on the shelf, but it is spoken for and stops counting as free. |
| **Completed**   | Material is consumed and the finished goods are added to Shopify.                                |

Reserving rather than consuming at the start is deliberate: the material has not been used yet, so the ledger should not say it has. What it does do is stop the reorder page telling you the stock is available when a batch is already relying on it.

## Closing a batch

When you close one you enter what you **actually** made and what it **actually** took. Both can differ from the plan, and that is the point.

* Finished goods go to Shopify through the same path a stock adjustment uses, so they appear in the product's inventory history like any other change.
* Material is consumed at the actual quantities, not the planned ones.
* The gap between planned and actual is the **yield variance**.

<Note>
  Cancelling an in-progress batch releases its reservation and writes no movements. Nothing was consumed, so nothing is recorded as consumed.
</Note>

## Yield variance

A batch that planned 24 cookies and made 22 has a variance of -2. One batch tells you nothing; a pattern tells you a lot. A recipe that consistently under-yields is either wrong or the process is losing something, and both are worth knowing.

The **Yield Variance** report on the [Reports](/user-guide/reports) page sorts by the worst shortfalls.

Because a material has no Shopify quantity to reconcile against, closing batches honestly is what keeps material stock true. A batch closed at its planned numbers when it really took more will drift, and nothing can detect it.