Every sale has a plan. It is usually in one person's head, and it covers the part that is fun to plan: what is discounted, by how much, and the copy. The part that goes wrong is the part nobody wrote down.
The fields worth having
| Field | Why it matters |
|---|---|
| Exact start and end, with timezone named | Store timezone, not yours. Writing the timezone down is what prevents the most common failure. |
| Where the original prices are saved | A file path or a tool. If the answer is 'the app remembers', name the app and how to reach the list. |
| Which products are in scope, as a number | A count you verified, not a collection name. 1,240 variants is checkable; 'the summer collection' is not. |
| What happens if it half-runs | Finish forward or roll back, decided now rather than at 2am. |
| Who ends it if you are unavailable | A name, and confirmation that they have access. |
| What the storefront should look like after | Strikethrough gone, badges gone, banner gone. Someone has to check, and they need to know what they are checking for. |
The one line most plans are missing
Where the original prices are saved. If a sale goes wrong, that single answer decides whether recovery is five minutes or a weekend.
Keep it to one page
A long runbook does not get read during an incident. One page, printed or pinned, with the six fields above filled in, covers the situations that actually occur. Everything else can live in a longer document nobody opens under pressure.
Test the handover, not the document
Give the page to whoever would end the sale in your absence and ask them to talk you through ending it. The gaps show up in thirty seconds and they are never the ones you expected.