There is a category of work in every small business that has no owner and no name. Somebody does it, every month, because it needs doing. It is not in anybody's job description. It never comes up in a meeting. And it takes an evening.
A wall calendar rebuilt from scratch. A price sheet retyped into a new document. A report assembled by copying numbers out of one system into another. A list of this month's jobs pasted into a spreadsheet so somebody else can read it.
Nobody was hired to do any of it. It accumulated.
Why it is invisible
Three reasons, and all three are about how the work is shaped rather than how big it is.
It happens monthly, so it never feels urgent. A daily annoyance gets fixed because it annoys you daily. A monthly one is over before you have finished being irritated by it.
It is nobody's specialty, so nobody complains about it. The person who does it is usually doing it because they were the one holding the file. They do not think of it as a problem; they think of it as Tuesday.
It looks small in isolation. Two hours is not worth a meeting. Two hours a month, times however many of these you have, is a working week or more a year, and nobody adds it up because nobody sees them as the same category of thing.
How to find yours
Ask a different question than the one people usually ask. Not "what takes the longest", which surfaces the big obvious jobs that are actually somebody's actual work.
Ask: what do we rebuild?
Rebuild is the word that finds them. Anything that gets made from scratch on a schedule, out of information that already exists somewhere else, is on the list. Anything typed twice is on the list. Anything where somebody opens last month's version and edits it into this month's version is definitely on the list.
Ask the people who do them, not the people who own the business. The owner usually does not know these exist, because they stopped being the person doing them years ago.
Then do the arithmetic, out loud
Write each one down with two numbers next to it: how long it takes, and how often it happens. Multiply out for a year.
The first one on the list is rarely impressive on its own. The list is. Most businesses that do this exercise properly find between five and fifteen of them, and the total surprises everyone in the room.
That total is the actual number. Not the biggest job on the list, the total.
What to do with the list
Sort it by a simple test: does it happen the same way every time, from information that already exists?
The ones that pass are the candidates. Something rebuilt monthly from a file somebody already maintains can be generated instead of rebuilt, and after that it takes nobody any time at all.
The ones that fail usually fail because there is judgment in the middle. Those are not lost causes, but they are a different job: the machine does the parts either side and hands a person the decision, which is still most of the saving.
The part that makes it stick
The saving decays if the outputs come from different places. If the calendar on the wall and the listing on the website are generated from two separate files, somebody has to keep them agreeing, and that somebody is doing a new monthly job you just invented.
One source, many outputs. When the source changes, everything downstream changes with it, and there is nothing left to keep in sync.
That is the property worth insisting on, and it is the difference between automating a task and actually getting the evening back.
We have a worked example of exactly this on the site, with a sample you can run in your browser: the calendar automation. It started as one job at one studio, which is how they all start.