The questions that would expose me: vetting a growth marketer from the seller's side
I wrote the interview script designed to catch someone like me. Six questions, what a real answer sounds like, and what a fake one sounds like.
the short answer: most founders vet growth marketers by asking about results, which is the one thing anyone can rehearse. these six questions are harder to fake because they ask about process and loss rather than outcomes. i wrote them from the seller’s side, which means i wrote them to catch myself.
1. “How many creatives did you ship last month, and how many won?”
a real answer comes back with two numbers immediately, and the ratio is bad. something like “about forty, three worked”. operators are not embarrassed by this because they know the base rate: roughly 5% of meta creatives become winners.
a fake answer talks about quality over quantity, or gives you a win rate above about 20% without explaining the sample. anyone claiming most of their creatives work either tests very few or is counting something other than winners.
why it works: you cannot invent this number under mild follow-up, because the follow-up is “what did the losers have in common” and that requires having watched them fail.
2. “What was your kill threshold, and who set it?”
a real answer contains a number that was written down before spend started, and the person who set it was them, agreed with the client.
a fake answer describes monitoring performance and making judgement calls. that is scaling on vibes, and it is how budgets get spent on things nobody wanted to admit were dead.
3. “Show me a test that failed and what it cost.”
a real answer is specific, slightly boring, and includes the money. “we put about four thousand into a channel over ten days, it came back at three times our ceiling, we killed it and wrote up why.”
a fake answer reframes the failure as a learning that led to a success. that is a case study, not a failure.
why it works: presenting a loss with a number attached requires having taken one. it is the single most reliable filter on this list.
4. “What would make you tell me not to hire you?”
a real answer names conditions. no retention yet, budget below the level where tests produce signal, a product still changing weekly, a market they do not understand.
a fake answer is a version of “we can help any business”. people who will take any engagement will take yours regardless of whether it works.
5. “Which of my numbers do you want before you recommend a budget?”
a real answer asks for retention or repeat usage before anything else, then funnel step conversion, then current spend and cost per acquisition. acquisition comes last because it is meaningless without the others.
a fake answer goes straight to channels, audiences and creative ideas. anyone who can pitch you a channel plan before seeing whether users come back is selling you traffic.
6. “Who does the work, and what happens if you are unavailable for two weeks?”
a real answer from a solo operator is honest about the ceiling: fewer clients, no cover, and here is what would pause. from an agency it should be a named person you have met, not a role.
a fake answer is reassurance without a name. the most common disappointment in this category is meeting a senior person in the pitch and getting someone else in the work.
The questions that do not work
“what results have you gotten?” everyone has an answer and the good ones sound identical to the invented ones.
“what is your process?” produces a rehearsed framework with four alliterative steps.
“who have you worked with?” logos tell you who paid, not who succeeded, and the failed engagements are never on the page.
Turning it on me
fair is fair. my answers: 150+ creatives a month at ZuAI against roughly $300k in monthly spend, and most of them lost. kill thresholds written before spend, always. i have run market tests that returned nothing at all and the honest write-up was that the market was too thin, not that we learned something valuable. i tell founders not to hire me when nobody comes back to their product yet, which is the most common reason. i want retention data before i will discuss budget. and i am one person, so if i am unavailable the work pauses, which is a real cost you should price in.
if you want to run these questions at me directly, let’s talk. that is genuinely a good use of the free call.
sources: published 2026 meta creative win-rate benchmarks · ZuAI figures from internal dashboards during the engagement.