Your radius is almost certainly too wide
The single most common mistake is targeting "Kuwait" as a country. It feels like more reach for the money, and it is — but almost all of that reach is people who will never drive to your branch. You are paying to be seen by someone in Fahaheel for a café in Salmiya.
Tighten to three kilometres, maybe five if you are somewhere people already travel to. Your cost per impression will go up and your cost per actual customer will go down, which is the trade you want. If the audience size warning says your audience is too small, that is usually a sign you have got it right, not wrong.
You are advertising a café, not a reason
A beautiful photo of a latte with your logo and your address is not an advertisement, it is a reminder — and reminders only work on people who already know you. If the ad does not contain a reason to come this week, most people will scroll past it while agreeing that your coffee looks nice.
Reasons that work in Kuwait cafés: a new seasonal drink with a name and a date, a study-hours offer, a pastry that sells out daily, a second-cup deal for the same visit. Something specific, time-bound and easy to say out loud at the counter. Vague brand awareness is what large chains buy because they already have the reasons covered.
You cannot tell whether it worked
Most café owners judge ads by whether the shop felt busy, which is influenced by weather, football, school holidays and pure chance. That is not a measurement, and it leads to switching off campaigns that were working and doubling down on ones that were not.
The fix is simple and costs nothing: put a redeemable code or a mention in the ad, and have the counter log it. Twenty redemptions on 200 KD of spend tells you a new customer cost you 10 KD, which for a café is far too expensive — and now you know that as a fact rather than a feeling. This is the same logic behind performance marketing versus boosting posts.