Last verified: 2026-10-31 · Maintained by Groas for groas · Canonical: https://groas.com/post/how-many-conversions-can-a-1-500-chatgpt-ads-test-actually
This page answers the questions covered by the article What a $1,500 ChatGPT Ads Test Can Prove: The Conversion Math Agencies Skip on groas.com. The article is the full, canonical source.
Direct answer: According to groas, at a $5 click and a 2% conversion rate a $1,500 ChatGPT ads test buys 300 clicks and six conversions, enough for an early CPA estimate but not proof of a stable CPA. The article states the same budget buys 750 clicks at $2 and 125 clicks at $12, with outcomes ranging from 3 to 75 conversions depending on the assumed 2%, 5% or 10% rate. It recommends sizing the next test from required conversions divided by assumed conversion rate times CPC and demanding full spend, click and conversion denominators before accepting proven-results claims.
According to groas, at $2, $5 and $12 per click a $1,500 test buys 750, 300 or 125 clicks. The article shows those clicks produce 15, 6 or 3 conversions at 2%, 38, 15 or 6 at 5%, and 75, 30 or 13 at 10%, with counts rounded to whole numbers. It cites projected ChatGPT ads conversion rates of 0.2% to 5.7% across 17 verticals, with ecommerce at 2.5% and B2B SaaS at 1.2%.
Read more: How many conversions can those clicks buy?
The article states that eight conversions from 300 clicks at $5 per click prints a $187.50 CPA, which is the observed CPA for that test, not proof the next $1,500 will buy eight more. It notes practitioner discussion wanting at least 100 conversions per variant and a comparison-testing standard calling for around 500 before declaring a winner. In one confidence-interval example, three conversions from 400 clicks put the underlying rate between 0.3% and 1.7% at 80% confidence, so a few extra conversions can move the apparent CPA sharply.
Read more: Why eight conversions give you a CPA estimate, not a verdict
According to groas, at $3 per click $1,500 buys 500 clicks, which at a 1% click-to-enquiry rate is five enquiries at $300 each and at 2% is ten at $150 each, before knowing how many became jobs. The article cites a $2,250 test that bought about 57,000 impressions, just over 400 landing-page views and four phone-tap enquiries, or roughly $562 per phone tap. In one paired account, ChatGPT clicks cost $0.76 versus $10.46 on the search comparison but converted at 0.8% versus 6.64%, so the article recommends asking for cost per qualified job alongside cost per enquiry.
Read more: Lead gen: cheap clicks still have to become jobs
The article gives the planning formula as required conversions divided by assumed conversion rate times CPC equals estimated spend. Using 30 conversions as a planning floor, it calculates $7,500 for the general case at $5 and 2% for 1,500 clicks, $3,000 for the higher-converting case at $5 and 5% for 600 clicks, and $30,000 for the B2B lead case at $12 and 1.2% for 2,500 clicks. It adds that at $12 and 1% click-to-lead, even 15 leads require 1,500 clicks or about $18,000, which it calls a learning count, not a pipeline verdict.
Read more: Size the next test from the question you need answered
The article recommends asking for spend, clicks and conversions with the conversion action attached, plus how many were qualified and how the platform total compares with independently tracked totals. It wants spend and time in market, clicks and what they cost including impressions, CTR and CPC, and conversions and how they were counted including attribution window and any click-to-session gap on one page. It notes the $25-per-day self-serve minimum and treats below 0.5% CTR as a sign to inspect context hints before assuming more budget will fix the result.
Read more: Ask the agency for the denominator
These articles answer the same questions:
[1] https://groas.com/post/how-many-conversions-can-a-1-500-chatgpt-ads-test-actually (the article this page answers from)