Name the result first

Two sisters in Manama sell abayas online. One of them puts a small amount each month into paid posts. The other is sure it is money thrown away.

They have the same argument at the end of every month. One points to the orders that came in the week the posts ran. The other points to the regulars who would have ordered anyway. Each finds the number that suits her, and each leaves the conversation exactly where she started.

The trouble is not the data. They have plenty. The trouble is that neither has said what would move her.

So the argument has no finish line. Any result can be read both ways, and it will be, because each sister is defending a position rather than testing one.

There is a simple fix, and it comes before anybody opens a spreadsheet. Each of them says, out loud, what she would need to see to change her mind. Perhaps it is first-time buyers rising over two months while the posts run, and falling back when they pause. Perhaps it is something else. The point is that they agree on it now, while neither knows which way the answer will fall.

Then they run the test and look.

It sounds obvious, and it is surprisingly rare. Most arguments about whether something works are not really arguments. They are two people taking turns to restate a view that no evidence was ever going to reach.

If one sister cannot name any result that would change her mind, that is worth knowing too. It means she is not weighing the question at all. She has already decided, and she is only there to watch.

Agree on what counts as proof. Only then is it worth looking at the numbers.