YC-style, written-first, decision-oriented. Start · Validate · Position · Build · Launch · Sell · Operate · Grow · Ownership. Every card gives a short answer, a heuristic, and an Aury seed.
Before writing a line of product, get honest about the problem, the person, and whether this is worth your next two years.
Two filters: does the problem come up unprompted in real conversations, and would you personally use the solution today. If both are no, walk away — no amount of execution fixes a problem nobody has.
No. Do the first 20 customer conversations and the first paying pilot on nights and weekends. Only quit when the work is bottlenecked by your hours, not by uncertainty.
Solo is fine — but be honest about the skill you're missing. If you're solo and non-technical building software, budget for a builder or no-code path from day one.
Write the one-sentence idea, name the exact first user, list the painful moment, and book five conversations. Do not buy a domain, design a logo, or build a prototype until you can describe who hurts and why now.
Twenty focused conversations tell you more than any survey. You're a journalist, not a salesperson.
Ask about the last time they hit the problem, step by step. No hypotheticals. No pitching. Follow curiosity, not your script.
Strong: they've tried to solve it (paid tools, spreadsheets, hires), they'd pre-pay, they introduce you to peers. Weak: "cool", "interesting", "let me know when it's ready".
Around 15–20 in the same customer segment. If you're hearing the same three pains by call 12, you have a signal. If every call is a different problem, your segment is too broad.
Useful for headline / audience testing, not for validating demand. An email signup is not a customer. Getting someone to pre-pay $50 is.
Treat it as a weak signal, not a moral failure. Either the pain is not urgent, the buyer is wrong, or the promise is too vague. Change one variable at a time before you declare the idea dead.
Positioning is a decision, not a paragraph. Nail it before you spend a euro on brand.
"For [audience] who [pain], [Product] is a [category] that [outcome], unlike [alternative]." If you can't finish it cleanly, you're not ready to build a homepage.
Usually no. Sit inside a category customers already reach for ("CRM", "invoicing", "the design tool") and be "the X for Y". Category creation is a 3-year, capital-heavy game.
Outcome first. In the customer's language. No jargon, no product name, no "AI-powered". Cut every word that could describe a competitor.
If five different customer types can read it and all think it might be for them, it is too broad. Strong positioning makes the wrong people self-select out and the right people feel uncomfortably seen.
One job, one price, one paying user. Everything else is Phase 2.
Small enough that one person could use it end-to-end this week and pay you for it. If v1 needs settings, teams and an admin, it's not v1.
Whichever gets a paying user first. No-code is not a weakness — it's a constraint that forces you to ship. Rewrite only when a real customer complaint demands it.
The shortest credible one. What you (or your builder) can ship in weeks. What has an escape hatch when it caps out. Fashion is expensive — pick fit.
Objective, one core flow, screens (name + purpose), functional requirements (verifiable), tech constraints, acceptance criteria. Aurygine's Build Package produces this for you.
V1 contains only what proves the core promise and lets one customer complete the job. Later contains collaboration, settings, polish, automation, analytics, and edge cases. If removing it does not break the first paid outcome, it is later.
Launch is not one big day. It's a series of small ones, each aimed at the smallest audience that would actually care.
When v1 does the one job end-to-end, you can charge for it, and you have a list of 30-100 people who might care. Not before.
Only if your audience is on Product Hunt (mostly developers, indie hackers, product people). For niches (dentists, real-estate agents, non-profits), a Product Hunt launch is noise.
Positioning, pricing, landing page, one core flow works, payments live, terms/privacy, support email, first 3 case studies (real or friendly), and a 10-day communication plan.
By hand. Every one. From your list, warm intros, one narrow community, one narrow content bet. No scaled channels until you've earned 10 by hand.
Measure conversations started, qualified replies, demos booked, payment attempts, and objections repeated more than twice. Vanity traffic is useful only if it changes the next message or the next audience.
Selling as a founder is not "sales". It's making it easy for the right person to say yes.
Anchor to the value you deliver, not your costs. Start higher than feels comfortable. Discount for design partners, not for cold prospects. Never charge in a currency your founder-brain can't survive on.
Two-thirds them talking, one-third you. Understand what they've tried, what they'd pay to solve, and what would make them a no. Send a written follow-up in 24h with the exact price and next step.
Lead = shown interest. Prospect = has the pain and the budget. Pipeline deal = has a decision-maker, a compelling event, and a next step on the calendar. Only the third one belongs on your dashboard.
When you've sold enough deals yourself to know the exact playbook — script, objections, close rate, deal size. Never hire sales to figure out sales.
Once real revenue exists, the job shifts from building the thing to running the company. Small systems, held weekly, beat heroics.
Cash, MRR, pipeline, one customer conversation, and one thing that's slipping. Keep it under 15 minutes so you actually do it.
Between 3 and 6. Any more and you're not running the business, you're doing analytics theatre. Pick the ones a decision would actually turn on.
At least 12 months at current burn. Under 6 months you're making desperation decisions. Over 24 and you're probably not shipping urgently enough.
When there's a role you've done well enough to describe in writing, that's clearly bottlenecking growth, and where a good hire pays for itself inside 6 months.
Churn is a leading indicator of a positioning or product problem. Fix it upstream: who you sell to, what you promise, whether the product delivers in week one. Retention campaigns don't fix a bad fit.
Growth is not a hack. It's finding the one loop that works, then feeding it discipline and time.
The one closest to your customer's actual behaviour today. Own one channel before you touch a second. Two mediocre channels is worse than one great one.
Content compounds slowly. Ads scale money if unit economics work. Outbound is fastest at low volume but hardest to sustain solo. Pick the one your founder-brain can hold consistently for 6 months.
When more capital would clearly accelerate a working loop — not when it would help you find one. Bootstrapping longer keeps ownership; raising earlier trades ownership for time.
Revenue is repeatable, churn is under control, CAC pays back within 12 months, and you're the bottleneck — not the product. If any of those isn't true, don't scale.
Ownership questions get expensive when left late. Get them right early, then forget about them.
Before you charge a customer or sign a co-founder. Not before. "Which entity" depends on where you live and where you sell — get one hour with a local specialist.
Rarely 50/50 unless roles, commitment and risk are truly equal. Always vest over 4 years with a 1-year cliff — no exceptions, even for the person you love working with today.
Customer terms of service + privacy, a supplier / builder contract with IP assignment, and a co-founder / operating agreement if there's more than one of you. Template first, lawyer later.
Clean cap table, all IP in the entity, real customer contracts, monthly bookkeeping. A messy company is a discounted company.
Aury answers as an operator who has already built and sold companies — with a position, tradeoffs, and one next action.