The math nobody does at the start

Meta's learning-phase guidance points to about 50 optimization events, like purchases, in the week after a significant edit before delivery settles. That is guidance for getting stable results, not a guarantee.

Put your own numbers in:

1
What can you afford to pay for one sale? That is your break-even cost per sale.
2
Multiply it by 50. That is roughly what one week of a fair test costs.
3
Multiply by about 4.3 for a month.

If your product can only afford $40 for a sale, that is about $2,000 a week, or $9,000 a month. At $90 it is about $4,500 a week.

If your monthly budget is $1,500 to $2,000, you are not funding a test. You are paying Meta to guess.

What that looks like in real life

A small budget spread across a small warm audience gives you very few clicks and very few purchases. A couple of weeks in, the report says cheap traffic, healthy add-to-carts, leads coming in. None of that pays the bills. The owner, who wanted sales, quits right before the data would have meant anything.

What to do when the budget is too small for the test

1
Fix the offer and checkout first. A checkout that leaks eats any budget. It is the cheapest improvement.
2
Build the email list and the follow-up before paying to send people to a list you cannot use yet.
3
Retarget the people who already visit instead of buying cold traffic you cannot afford to learn on.
4
Run a small, clearly labeled traffic or engagement test if you want creative data, and agree up front that it is not a sales test.
5
Wait until you can fund the real test for at least four weeks, untouched.

Before you hire anyone for ads, ask for these in writing

Your average order value and what you can afford to pay per sale. Your real monthly ad budget. What event the campaign is optimizing for, such as a purchase or a lead. A minimum run time with no pausing or edits during it. And who owns your website, pixel, and tracking.

If your ad person will not talk about these before building, that is your answer.

Ads can work. They just cannot rescue an offer, a checkout, or a budget that is not ready. I would rather tell you that before you spend than after. Results depend on your offer, audience, budget, creative, and landing page, and no one can guarantee them.