Importing a paper job book
Two ways in: type the open jobs at the counter over an afternoon, or upload a spreadsheet. Most shops with fewer than forty open jobs type them, because building the spreadsheet takes about as long.
What to bring across, and what not to
Bring the open jobs. Anything currently on a bench, waiting on a part, or sitting on the shelf uncollected. That is what you need to be able to see tomorrow morning.
Leave everything closed behind. This is the part people get wrong. Importing three years of finished work does not give you history, it gives you a search box that returns forty results for every customer name and buries the one you wanted. Your old pads are still your records, in the box, where the accountant can find them.
If you genuinely need a closed job later, type that one in when you need it.
The CSV columns
| Column | Required | Format | If empty |
|---|---|---|---|
| `summary` | Yes | Free text | Row rejected |
| `item` | Yes | Free text | Row rejected |
| `customer_name` | Yes | Free text | Row rejected |
| `customer_mobile` | No | Digits, spaces and `+` | No messages send for that job |
| `status` | No | One of `in`, `bench`, `waiting`, `done` | Defaults to `in` |
| `opened_at` | No | `YYYY-MM-DD` | Defaults to today |
| `promised_at` | No | `YYYY-MM-DD` | No promised date |
| `waiting_for` | No | Free text | — |
| `supplier` | No | Free text | No chase email for that job |
Header row required, UTF-8, comma delimited.
A sample
```
summary,item,customer_name,customer_mobile,status,opened_at,promised_at,waiting_for,supplier
Creaking at the bottom bracket,"Kona Rove, black",M. Ferreira,+1 555 0164,bench,2026-08-24,2026-08-29,,
Freehub body replacement,"Genesis Croix, grey",R. Iyer,+1 555 0142,waiting,2026-08-14,2026-09-02,Freehub body 11-speed,Calder Distribution
```
Dates are the thing that goes wrong
Use `YYYY-MM-DD` and nothing else. A spreadsheet exporting `08/09/2026` is ambiguous between August 9th and September 8th, and we will not guess — those rows are rejected with the row number rather than imported at a date that might be four weeks out. If your export insists, format the column as text first.
Questions
Can I import twice?
Yes. Rows are matched on customer mobile plus summary, so re-uploading a corrected file updates rather than duplicates. Rows that match nothing are created.
What if two jobs have the same customer?
That is normal and fine. Customers are matched on the mobile number, so one person with three things in the shop is one customer with three jobs, and their page shows all three.
Can I undo an import?
Within an hour, yes, in one click — every import is a batch and the batch can be rolled back. After that you delete jobs individually, which is deliberate: an undo that works a week later would take out everything anybody had done to those jobs since.