How to record actual values#
For: Manager | Admin You’ll need: At least one driver — see How to create a demand driver.
The forecast learns from what actually happened. There are three ways to get those numbers in: type them, import a file, or have another system push them.
Type one value#
Go to Planning → Demand → Actuals.
Click ‘Create’.
Pick the ‘Driver’, the ‘Location’, and optionally the ‘Area’.
Set the ‘Date’, and the ‘Time’ if the driver is measured more finely than daily.
Enter the ‘Value’. Money drivers take the amount as you’d write it, for example 4000.50.
Click ‘Save’.
Recording the same driver, place, and time bucket twice updates the existing value rather than adding a second one.
Import a CSV#
Best for backfilling history from a spreadsheet or a point-of-sale export.
Go to Planning → Demand → Actuals.
Click ‘Import’.
Choose your ‘CSV file’ and upload it.
The columns are date[,time],location[,area],driver,value. The driver column takes the driver’s code or its name.
Warning
The import is all-or-nothing. One bad row and nothing is imported — you get the reason (with the line number for a bad date, time, or value), fix it, and upload again. That is deliberate: a half-imported week teaches the forecast something false.
Push them over the API#
If your point-of-sale, booking, or ERP system can post to a URL, it can keep the numbers current without anyone touching a spreadsheet. POST /api/public/v1/demand/observations accepts up to 1000 rows at a time and is safe to retry — re-sending the same bucket updates it rather than duplicating it. See the Developer Portal.
How far back is worth recording#
More history means a better sense of your seasons. A year lets Shiftavo compare a week with the same week last year; a few weeks is enough to learn your weekday pattern. Anything is better than nothing.
Sub-daily values older than about 13 months are automatically rolled up into daily totals to keep the history manageable. The daily total is preserved exactly.
Verify it worked#
The rows appear in the list under Planning → Demand → Actuals, and the ‘Actual’ column on How to read the demand plan fills in for those days.