Y Combinator - How to Get Your First 10 Customers

Short Summary

Main recommendation: Get your first 10 customers by working your warm network first, then manually showing up where your buyers are, starting with people who already trust you and using in-person and direct, personalized outreach before scaling with tools. The video stresses that your first customers come from you personally (the founder doing unscalable work), not from automation.

Key Points

• Before any outreach, determine where your target buyer actually spends time, since cold email and LinkedIn only suit some buyers (e.g., sales leaders), while others (school administrators, property managers, truck dispatchers) may need in-person or phone contact.
• Customers 1 to 3 almost always come from your warm network (friends, former colleagues, classmates, and people one introduction away), because early buyers trust the founder.
• Showing up in person, at small industry conferences, and at small founder dinners or happy hours converts better than cold outreach for early customers.
• Customers 4 to 10 typically come from unscalable, manual tactics such as Reddit outreach, personalized DMs, and free consulting sessions; prospecting tools and higher-volume automation make sense only after about 10 to 20 quality customers and a refined pitch.

Details Worth Noting

• Concrete instructions:
  • Spend about an hour answering concrete questions about your buyer: their average day, how often they check email, which conferences they attend, whether they use Reddit or LinkedIn, whether they pick up the phone, whether they rely on friends' recommendations, and which industry newsletters they read. If you can't answer most of these, talk to more real customers first.
  • Work your sources in order: (1) your personal network (former colleagues, classmates, industry friends); (2) your second-degree LinkedIn connections, asking for warm intros; (3) AI network search tools such as Happenstance, which lets you search your extended network in natural language (e.g., "find people who work on notification systems at big companies").
  • When asking for an intro, say who you want to meet, why they specifically would care, and what to say in the forwarded email.
  • For conferences: set up a Calendly with back-to-back 15-minute slots across every day of the event; email the attendee list a couple of times beforehand to fill slots; email again during the event to catch people who didn't open earlier emails.
  • For micro events: host a founder dinner or happy hour for about 6 to 10 people of your ideal customer profile, with roughly $50 to $100 per head.
  • For Reddit: find old threads where people complain about your problem and DM each commenter individually; one healthcare founder did two to five posts a day for a couple of months.
  • For outbound tools: Apollo (lead database with email finding and basic sequencer; generous free tier) is enough for most founders at this stage; Clay adds AI research and enrichment workflows when you need to qualify leads on specifics such as software used, hiring, or recent LinkedIn posts; LinkedIn Premium helps with fresh professional data, and a good sequence is a connection request without a message followed by a short DM after acceptance.
  • Frame outreach as a genuine request for advice, mentorship, product review, or whiteboard session, but only if you are actually open to learning (do not disguise a sales call).
  • Keep outreach under 75 words, include one clear call to action (reply, call, or 15-minute demo), and read the draft aloud to a friend to catch lines that sound unnatural or AI-written.
  • Give the prospect something first, such as a quick vulnerability scan of their public website, an app walkthrough with specific suggestions, or a short tailored audit note, then ask for a meeting.
  • Follow up three or four times over a couple of weeks.
• A founder who got little traction from months of email and LinkedIn outreach to a legacy industry closed more in 3 days at an industry trade show than in 3 months of cold emails.
• One founder flew to a single executive buyer four weeks in a row; the buyer kept rescheduling but eventually closed. Another was kicked out of a customer's office after 8 minutes but later won that customer as one of his biggest accounts.
• One founder offered lawyers $100 to $200 per hour of product feedback; about 30% of those contacted accepted, and despite the expensive-sounding math, the conversion rate made his customer acquisition cost reasonable, which fits markets with high contract values.
• One founder talked to 200 salespeople before her team built the product, testing one hypothesis per week through LinkedIn connections; about 50% of requests were accepted and 20% of those converted to calls.
• One DevTools founder offered free whiteboarding sessions on agent architecture and then a shared Slack channel, where the architecture required his product.
• Reddit threads persist and are pulled into Google search for years, so work done today keeps paying off.

Bottom Line

Your first 10 customers depend less on tools and more on your own manual effort: lean on your network, show up in person, and do the unscalable work that automation cannot replicate, then scale only once your pitch is proven.