Indictment: Two hundred words of biography, zero words of an actual question — 'what to do next' is not a question, it's a shrug.
see the specimen they pasted
So I'm building this app, it's kind of like Uber but for dog walkers, and my co-founder thinks we should focus on cities first but I'm not sure, anyway we have about 200 users and we're pre-seed and based in Austin and I used to work at a bank — can you help me figure out what to do next?
“it's kind of like Uber but for dog walkers”
'Kind of like Uber but for X' is the startup equivalent of a name tag — it tells the model the pitch, not the problem. The model now knows your comp, not your constraint.
cut: the Uber analogy — replace with the specific decision or bottleneck you're stuck on
Background context only earns its place when it changes the answer — drop anything the model doesn't need to act.
“my co-founder thinks we should focus on cities first but I'm not sure”
This is the only real tension in the prompt, and you buried it in a subordinate clause like it's a footnote. That's your question — ask it directly.
swap: the buried aside → 'Should we concentrate growth in one city before expanding, or grow diffusely across markets? What criteria should drive that call?'
The actual question is the thing you're least sure about — lead with it, not the backstory.
“I used to work at a bank”
Your banking background is doing absolutely nothing here unless you're asking about financial modeling or regulatory risk — right now it's just noise the model will try to awkwardly honor.
cut: 'I used to work at a bank' — or explain how it's relevant to the decision you're making
Every detail you give the model is a weight it carries — only pack what pulls in the direction you want to go.
“can you help me figure out what to do next?”
'What to do next' is a question so wide the model will answer with a generic startup playbook that fits ten thousand companies — not yours.
swap: 'what to do next' → a specific decision with named trade-offs, e.g. 'Should we double down on Austin before entering Dallas and Houston, or run lean experiments in all three simultaneously?'
Swap every 'what should I do' for the specific fork in the road — name the two options you're actually weighing.
You are an early-stage startup advisor with experience in marketplace and on-demand consumer apps. Context (use only what's relevant): - Product: a two-sided marketplace connecting dog owners with vetted dog walkers - Stage: pre-seed, ~200 users, based in Austin, TX - Core tension: my co-founder wants to concentrate growth in one city before expanding; I'm unsure whether that's the right call Task: Give me a structured recommendation on the city-concentration vs. multi-market question. Specifically: 1. What are the strongest arguments FOR locking into Austin first? 2. What are the strongest arguments AGAINST — i.e., when does early geographic concentration backfire for marketplace startups? 3. What 2–3 metrics or signals at our stage should actually drive this decision? 4. What would you recommend given 200 users and pre-seed constraints? Format: Use numbered sections matching the four questions above. One concrete recommendation at the end, max two sentences. Skip generic startup advice that applies to any company — stay specific to two-sided local-service marketplaces. Total response under 400 words.
/roast/example/context-dump