This blog could publish five posts today. It never will, on purpose.

Every business that publishes something on a regular schedule, a blog, a newsletter, a string of customer emails, eventually builds up a backlog of stuff that is ready to go. The temptation once several pieces are ready is to let all of them out at once and clear the queue. My own publishing system could do exactly that today. It never will, and that is on purpose, not a limit I ran into by accident.

The reason is the same reason you do not send a customer four texts in one afternoon about the same appointment. Getting something ready is not the same decision as when to actually put it in front of someone. A backlog of approved work is a good problem to have. Dumping all of it on one day is not a benefit. It is just a decision made without thinking about it.

The job that releases posts checks the date on each one, picks at most one that is actually due, and stops there. If I run it again by accident, or it runs twice because something hiccuped, nothing extra goes out. It looks at what already published today and does nothing the second time through. That part matters more than the daily cap itself. An automation that behaves differently depending on whether it happens to run once or twice is not safe to leave alone, because sooner or later it will run twice.

Whatever you automate that reaches a customer, ask both questions before you trust it. How often should this actually go out. And what happens if it runs by accident a second time. If the honest answer to that second one is “it sends again,” that is the part to fix first, before the part that decides what to send.