Micro-Budget Ad Testing

Here is the arithmetic that governs everything. At ten dollars a day with a twenty dollar cost per acquisition, you generate roughly one conversion every two days. Reaching the thirty conversions a smart bidding algorithm needs would take two months and six hundred dollars, and by then the creative is stale and the test is contaminated. The conversion event is simply not available to you as a testing metric. Anyone telling a beginner to test until they have significance has never run a ten dollar budget. So you test on the signals that arrive in hours: stop rate, hold rate, click, cost per click. THE THREE GATES: does it stop the scroll (hook rate, threshold ~25%), does it hold (hold rate at 15 seconds, threshold ~15-20%), does it move them (click-through, threshold ~1% on paid social). The diagnostic value is in which gate fails, because an ad does not simply not work: it fails at exactly one of three places, and the fix is completely different at each. THE FIVE-RUNG LADDER, run in order and never two at once: message test at $5-10/day across five different arguments; hook test isolating the opening; audience test, which is only worth paying for once the creative already passes gates one and two; offer test, which beginners wrongly run first; then careful scaling at no more than 20% every two to three days. And when the budget is genuinely too small for the goal, the skill says so plainly and gives the honest alternative rather than building a plan it knows cannot work.