A new clothing brand spends 20 on an ad, gets 40 clicks and one order. Is that a 2.5% conversion rate and a cost per sale of 20? Not really: with one order, the 95% Wilson interval on the rate runs from about 0.4% to 12.9%, so the true cost per sale could be anywhere from about 4 to 130. If each sale leaves 25 after cost, tax and fees, that test cannot say go or stop.
The clothing design studio launch planner now has a "Read a first small ad test" box. Type what you spent, the clicks and the orders. It shows your real click price, conversion rate and cost per sale, the 95% range, and one of four verdicts: go (even the worst end of the range pays), stop (even the best end loses), unclear (with roughly how many more clicks would settle it), or no margin (fix price and cost first). "Use these numbers" replaces the placeholder assumptions in the calendar.
The interval and verdict logic is in marketing.mjs with hand-checked tests (zero orders, bad input, a large losing and a large winning test). It is arithmetic on numbers you type; nothing is bought or posted, and it has not yet been used on a real designer's campaign.