Check these before you touch anything
Did something break? A rejected ad, a card that failed, a page that went down, a product that went out of stock, a form or a WhatsApp link that stopped working. Check this first every time. It is boring and it is often the answer.
Did you change something? Open the change history. A budget edit, a new creative added into a working ad, a switched objective, a paused ad set. Most sudden drops have a human cause with a timestamp.
Did the season change? Ramadan, back to school, the days around payday, the weeks after a price rise. Egypt has strong rhythms and they move the whole market, not just you.
What people blame that is usually not it
The algorithm. Platforms do change, but a change that broke your account would show in every account. If your drop has a date and a shape, it has a cause you can find.
The audience being "burned out". In a market this size, you are rarely out of people. You are usually out of a hook that still surprises them.
Competitors. Worth knowing about, not worth reacting to. You cannot see their numbers, so you would be copying something you cannot verify.
Diagnose in this order
- Look at the funnel stage by stage
Impressions, clicks, adds to cart or messages, confirmed orders, delivered orders. Find the first step where the number dropped. Everything after that is a symptom.
- If clicks dropped, the creative is tired
Same audience, fewer people stopping. That is the video or the image, not the targeting. Write new hooks for the same offer.
- If clicks held but orders dropped, look at the page
Something changed between the click and the order: speed, price, stock, delivery time, a broken button. Open it on a phone on mobile data and go through it yourself.
- If orders held but delivered orders dropped, look at operations
Confirmation is slower, the courier is slower, or a price change is meeting people at the door. No creative fixes this.
- Compare like with like
Seven days against the previous seven, not against your best week ever. Short windows on small budgets swing for no reason.
- Change one thing at a time
If you change the creative, the budget and the audience in the same week, you will not know what worked, and you will be here again next month.
- Write down what you changed and when
A three-line log per week is the difference between learning and guessing. It costs nothing and it pays every time this happens again.
The question to ask before scaling anything back up
Once you have found the step that broke and fixed it, do not jump straight back to the old budget. Go back in steps and watch the same funnel, because the thing that broke can break again, and a big budget makes a small problem expensive fast.
If you cannot say in one sentence why performance dropped, you have not finished diagnosing. Spending more on top of an unknown problem is the most expensive way to find out what it was.
The short version
Find the first step in the funnel that dropped, fix that step, change one thing, and compare against a like-for-like window. Targeting is the last thing to touch, not the first.