1. Before building

Exit when: 3 people describe the same pain unprompted and 2 already pay for a bad workaround

  1. Pick the niche before the idea
  2. Write a detailed one-paragraph hypothesis: who they are, what’s broken, and what it costs them
  3. Build a list of 100 contacts in one sitting. Name, company, title, and one observable signal (hiring, posts, company news)
  4. Sort by who said something you can quote back, then by who’s been active in the last 60 days
  5. Reach out where they’re active. For B2B, LinkedIn usually beats cold email
  6. Contact 10 people a day. Ask one question about how they do the work today. Earn the call from the reply
  7. Listen on the calls. Ask what they did the last time the problem hit. Save their exact words
  8. Get at least 8 calls before jumping to conclusions
  9. No repeated pain after 8 calls? Change one thing:
    • No replies → change the message
    • Calls but no shared pain → change the hypothesis. The next one is in your transcripts
    • Pain confirmed, no budget → ask who signs. If nobody does, rebuild the list around someone who can
    • Tried all three → change the niche
  10. Sell the outcome before you build it. Use three slides: problem, solution, outcome. Use their words, not yours
  11. Offer to deliver it by hand, name a price, and take payment

2. First customers

Exit when: 10 customers pay real money and 3+ would be annoyed if you shut down

  • Keep doing the work by hand until the patterns tell you exactly what to build
  • Watch customers complete real work over screen share. When they get stuck, ask what they expected
  • Ask what they would do if the product disappeared. Listen for a specific workaround, not praise
  • Solve one problem. Keep the pitch to one sentence
  • Launch the 90/10 version while it still embarrasses you: 90% of the outcome for 10% of the work
  • Keep deals small and fast. Take the ones that close in a week
  • Fire the wrong customers. The loud one dragging your roadmap costs more than they pay

3. Systematize

Exit when: revenue grows without a manual scramble each week

  • Fix retention before ramping up marketing. Growth follows a product people stick with
  • Make one channel work, then add another
  • Track activation at the account level: first output, time to it, and repeat output
  • Log failed attempts and technical errors. Fix them, then follow up with the affected customer
  • Turn repeated manual work into a checklist. Automate it only after you’ve done it 20 times

4. Scaling

  • Add features only when the same request keeps recurring
  • Grow the team only when demand exceeds your capacity
  • Add funnel tracking and session replay

Always

  • Sleep and exercise
  • Keep your composure with people
  • Keep two or three founder friends. Talk to them weekly
  • When giving up starts to sound good, take care of yourself and keep going

Final words

Building moves in uneven rhythms. One quiet week can give insights that completely changes the following ones. Each stage has its own pace and asks for different kind of work.

Keep listening to customers. Their words, payments, completed outcomes and repeated use show you what deserves attention.

The early stages reward curiosity and speed. Growth rewards focus, consistency, and repetition of the messages and channels that work.

Take the smallest useful action. The result will reveal the next step.

Keep listening, keep learning, and give the work time to compound.