Almost every business has one
A report assembled every week or month by pulling numbers from two or three systems into a spreadsheet, formatting it, and sending it to whoever needs it.
It takes two hours, it happens fifty times a year, and the person doing it is usually senior enough that those hours are expensive.
Why it automates well
- The steps are the same every time
- The sources are known
- The output format is defined
- It happens on a schedule
- Errors in it are consequential and hard to spot
What the automation does
- Pull from each source on a schedule
- Validate that each source returned plausible data
- Transform and combine according to the agreed definitions
- Produce the output in the required format
- Send it to whoever needs it, before they need it
Define the numbers precisely first
Automating a report requires stating exactly what each figure means — which date it is counted on, how cancellations are treated, what is included and excluded.
That conversation frequently reveals that two people have been calculating the same number differently, which is valuable to discover regardless.
Validate before sending
| Check | Prevents |
|---|---|
| Source returned data | An empty report |
| Totals within expected range | A wildly wrong figure being circulated |
| Comparison with last period | Silent source failure |
| All expected sections present | A partial report |
| Nothing null that should not be | Formatting errors |
An automatically generated report with a wrong number is worse than a manual one, because nobody was looking at it while it was assembled.