← Writing

Where to Find Your Ten

A step-by-step operational playbook for finding ten real people to have discovery conversations with. Week by week, channel by channel, with anti-thrashing rules built in.

The Ten Conversations piece makes the case for why you need to talk to real people before you build. This piece is the operational companion. How to actually find those people, what to say when you reach out, and how to keep yourself from thrashing between channels without making progress.

I needed this playbook myself. After years of building in isolation, the hardest part wasn't accepting that I needed to talk to people. It was figuring out where those people were, what to say to get them on a call, and how to stop switching tactics every three days when the first one felt slow.

So here's what actually works. Week by week. With specific time commitments and decision points built in so you don't mistake activity for progress.

Before you start: the who-where pair

You need two things nailed down before you do anything else. Skip this and every tactic in this guide becomes noise.

Who: A specific type of person, described precisely enough that you could recognize them in a room. Not "small business owners." Not "people who use AI." Something like: "freelance designers who manage more than five clients and are drowning in project handoffs." The narrower you go, the easier they are to find and the more useful the conversations will be.

Where: A specific place, online or offline, where at least ten of those people spend time. A subreddit. A Slack community. A LinkedIn search filter. A local meetup. If you can't name the place, your "who" is still too broad. Go back and slice again.

Checkpoint: Can you complete this sentence? "I'm looking for [specific person] and I can find them at [specific place]." If not, you're not ready for outreach. Spend 30 minutes narrowing until you can.

This is worth getting right. I wasted weeks messaging vaguely-defined people on channels where they didn't hang out. The problem wasn't the message. It was the aim.

The three-week playbook

This is designed around 30-45 minutes a day. You don't need to quit your job for this. You need a focused half hour and the discipline to do the same thing every day instead of switching tactics when it feels slow.

Week 1: Warm network + LinkedIn (conversations 1-4)

Goal: 3-4 conversations booked. Time: 30 min/day.

Templates for Week 1

Warm intro request (to your connector, not the target):

Hey [name], I'm researching how [type of person] handles [specific problem]. Do you know anyone who deals with this? I'd love a quick 15-minute conversation with them. Not selling anything, just trying to understand the problem. Happy for you to check with them first before connecting us.

LinkedIn DM (after engaging with their content for a few days):

Hi [name], I've been following your posts about [specific topic]. I'm researching how [people in their role] deal with [specific problem]. Would you be open to a 15-minute call? No pitch. Just trying to understand the problem before building anything. Happy to share what I'm learning from others.

Two things about these templates. First, use them as a starting point, not a script. The moment a message sounds templated, it gets ignored. Second, "happy to share what I'm learning" is doing real work. You're offering something back. That changes the dynamic from "give me your time" to "let's exchange."

Anti-thrash rule: Do not add a new channel this week. LinkedIn and warm intros only. The urge to also try Twitter, Reddit, cold email, and a Slack community is the build trap applied to outreach. Pick two. Work them for seven days. Add channels later if you need to.

Week 2: Follow the thread + add one channel (conversations 5-7)

Goal: 3 more conversations booked. Time: 30 min/day.

End-of-Week-2 checkpoint: You should have completed 4-6 conversations and have 3-5 more booked. If you have fewer than 3 completed, your "who" might be too hard to reach. Revisit your who-where pair before adding more channels.
Anti-thrash rule: If a channel isn't producing responses after 15-20 genuine attempts, drop it and try a different one. But give it 15-20 attempts first. Three messages and "this isn't working" is impatience, not data.

Week 3: Fill the gaps + decide (conversations 8-10)

Goal: Complete all 10 conversations. Make your Go/Pivot/Kill decision. Time: 30 min/day.

The anti-thrashing rules

I built these because I know what I'd do without them: spend three days on LinkedIn, get impatient, switch to Reddit, spend two days on Reddit, discover a Slack community, spend a day there, decide cold email is probably better, draft a template, revise it four times, and end the week with zero conversations and five half-started channels.

That's the build trap applied to outreach. Same avoidance pattern, different material. You feel busy. You're not producing.

The rules:

  1. Two channels max in Week 1. Warm intros and LinkedIn. That's it. You don't need more. You need depth on two, not surface on five.
  2. Add one channel per week. Only after the current channels are actually being worked at volume. "At volume" means 15+ outreach attempts per channel, not three messages and a pivot.
  3. 15-20 attempts before you call a channel dead. Three unanswered DMs is not evidence that LinkedIn doesn't work. It's evidence that you sent three DMs. Work the channel before you judge it.
  4. Referrals are not a channel. They're the engine. Every conversation should end with "who else do you know who deals with this?" This is not optional. It is the single thing that turns a slow pipeline into a compounding one.
  5. 30 minutes a day, same time every day. Not two hours on Monday and nothing the rest of the week. Consistency beats intensity. Put it on your calendar like a meeting.
  6. Do not rewrite your outreach message more than once. Send the first version. If the response rate is below 10% after 20 sends, revise it once. If you're on your fourth revision and you've sent twelve messages, you're polishing. Ship the message.
  7. Track in one place. A spreadsheet. A Notion table. A notes file. One place with: name, channel, date reached out, response (yes/no/pending), conversation date, key quote, referrals given. If it's not written down, it didn't happen.

What to say on the call

This isn't a script. It's guardrails.

Open with their world, not yours. "Tell me about how you handle [problem area] right now." Then shut up. Let them talk. The first two minutes are warm-up. The real answers come at minute five, after the polished version runs out and the honest version starts.

Follow the energy. When they lean in, speed up, or get specific, you're close to something real. Ask "tell me more about that" and stay there. Don't rush to the next question.

Ask about behavior, not opinions. "What did you do last time this happened?" is ten times more useful than "Would you use a tool that does X?" People are terrible at predicting their future behavior. They're honest about what they already did.

Ask what they've tried. "Have you tried to solve this before? What happened?" This tells you what you're actually competing with. Usually it's not another product. It's a spreadsheet, a manual process, or ignoring the problem.

Ask about money. Not "would you pay for this?" Ask "what are you spending on this now?" Time is money too. "I spend three hours a week on this" is a budget, even if they don't think of it that way.

Do not pitch. If they ask what you're building, keep it to one sentence and pivot back: "I'm exploring building something in this space, but I'm still figuring out what the real problem is. That's why I'm talking to you." The moment you explain your solution, the conversation stops being discovery and starts being a demo.

Close with the referral ask. "This was really helpful. Who else do you know who deals with this?" Do not skip this. Do not feel awkward about it. They just gave you 15 minutes. Asking for a name is not a burden. It's a compliment: you're saying their world is the one you want to understand better.

What your tracker should look like

Nothing fancy. A spreadsheet with these columns:

After ten conversations, read the "one-line summary" column top to bottom. The pattern will be visible. Not the pattern you expected. The real one.

The final checkpoint: Ten rows filled. Summaries written in their words. Pattern identified. Decision made: Go, Pivot, or Kill. If you're at this point, you've done more real validation than most builders do in a year of building. The decision is the finish line. Not the eleventh conversation.

When it feels slow

It will feel slow. Especially in Week 1 when you've sent fifteen messages and gotten two responses. The build trap will whisper: you could be coding right now. You could be making progress on the product. This outreach thing is a waste of time. You should just build the thing and see what happens.

That whisper is the same voice that gave me 30 beautiful, unseen projects.

Two responses out of fifteen isn't failure. It's a 13% response rate, which is normal for cold outreach. Send fifteen more. Follow up with the thirteen who didn't respond. Ask your two respondents for referrals. The pipeline compounds. It just doesn't compound on Day 2.

The discomfort you feel reaching out to strangers is the same discomfort you'll feel launching, selling, and asking people to pay. This is practice. The product that survives is the one built by someone who learned to be uncomfortable in public, not just comfortable in the editor.

The playbook is three weeks. Not three weeks of thinking about it. Three weeks of doing it, 30 minutes a day, starting tomorrow. Put it on your calendar right now. The first message is the hardest. The second one is easier. By the fifth, it's just what you do before lunch.

Related: Ten Conversations and One Decision — why ten conversations matter and what to do with what you learn. And The Build Trap Got Cheaper — why AI makes it easier than ever to skip this step.

Builder's Path is a public lab from Sellhausen AI Systems focused on AI-native building, validation, and product judgment.

Built by Frank Sellhausen · Thinking · Privacy