We were sure, and we were wrong — five pieces of evidence, one bad conclusion
Our buy buttons pointed at a shop that had been deleted. We asked an assistant to find a shop that exists and repoint them. It did. We shipped it.
The next day we checked again, and found five pieces of evidence that the shop was not ours. We declared it, in writing, to the whole team.
Confident, documented, and wrong
Then we asked the owner
It was their own shop, on a second account.All five pieces of evidence were correct. Only the conclusion was wrong.
Why the evidence could not settle it
Every one of those five proves the same single thing: “this browser session cannot manage that shop.” A shop owned by the same person on a different account produces exactly the same five results.
Where the reasoning broke
The rule we now use
Before writing “X is not the case”, ask:
can this evidence distinguish “not the case”
from “true, but invisible from where I am standing”?
If it cannot, write “I do not know” and ask someone who can.
Absence of access is not absence of ownership
The cost of being confidently wrong is higher than being unsure. Two other people changed their own links because of our message. We had to send a correction to all of them.
What did not change
The original mistake still stands: we asked “does it exist?” when the question that mattered was “where does the money go?” That one is still worth asking first — we just cannot answer it from the outside.
