The morning I almost shelved the idea
Coffee gone cold. Prototype half-working. A tidy list of questions that no one had agreed to answer. I had no audience—no Twitter thread waiting to go viral, no newsletter, not even a sleepy Facebook group. Just a knot in my stomach and a cursor blinking like it was judging me.
I waited a week for confidence to show up. It didn’t. I refreshed email, stared at a landing page that no one had seen, and told myself that maybe I should “work on the brand” for a while. That’s what you do when you’re afraid to ask for help—you rearrange pixels and call it progress.
What finally nudged me forward wasn’t courage. It was a bet with a friend. “You get one interview by Friday,” she said, “or you buy dinner.” I hate losing small bets. So I made something else instead: momentum. One message. Then one conversation. Then another. No megaphone, no magic—just a handful of small, slightly awkward moves repeated until they looked like a plan.
If you’re at zero, that’s not a disadvantage. It’s clarity. You can’t rely on reach. You have to rely on being useful and a little stubborn. There’s mercy in small numbers: when you talk to five people, you actually remember what they said. You hear the words they reach for when they’re tired. That’s the language you need.
What actually makes a good interview
My first few “interviews” were just me pitching with question marks at the end. “Would you use a tool that…?” People were polite; they said “maybe,” the deadliest word in validation. I learned almost nothing. The switch flipped when I stopped asking what they “would do” and asked what they actually did.
A real interview sounds like curiosity and takes place in the other person’s world. “Walk me through the last time you tried to solve this.” Then you shut up and let the timeline unfold. “What did you try first? And then what?” When they mention a workaround, you ask, “Why that? What makes it good enough?” You’re collecting the exact phrases they use to describe the pain, not grading them for choosing the same words as your pitch deck.
I started timing myself. Fifteen to twenty minutes. End on time even if it’s getting good. People remember you as the person who respected the clock. Before you hang up: “Is there one person you think I should talk to next?” Many will give you two. If they say they’re not sure, offer a tiny blurb they can forward so they don’t have to write one. Make that blurb sound like a person wrote it.
Here’s what changed the outcomes more than any clever trick: stopping the pitch. When someone asked, “So what are you building?” I gave a one-liner and then turned the camera back to them. “I’m exploring a way to make [problem] less of a tax on your day. But I don’t want to sell you on it—I’m here to understand the mess. Tell me about the last time it got in your way.” If they insisted on seeing something, I scheduled a second call and didn’t hijack the first one. Counterintuitive at first; obvious once you experience it.
A short transcript snippet from the conversation that changed my mind:
Me: “When does this problem tap you on the shoulder?”
Marta (clinic manager): “Honestly? Thursday afternoons. That’s when the cancellations pile up and I start playing calendar Tetris.”
Me: “How often is ‘pile up’—three a week? Ten?”
Marta: “Five to eight on a bad week. We call it the tumble. Capital T.”
There’s your copy. Not “reduce cancellations with AI.” It’s “Thursday Tumble,” and people who know the mess will nod at the phrase like you’ve been in the room.
Another snippet, this time from a mechanic:
Me: “What happens when the part isn’t in stock?”
Nasim: “There’s a clipboard of shame. We write it down so we feel bad enough to fix it next time. It never works. The clipboard is full.”
The clipboard became a character in my notes. You can build a feature for a character.
How the first twenty-five happened (without an audience)
I found three corners of the internet where my people vented: a niche subreddit, a Slack community, and an old-school forum that looked like it hadn’t changed since 2012. I didn’t blast a survey link. I read for a couple of evenings, wrote down the jargon and the recurrent complaints, and replied to a few threads with something that tried to be useful—no links, no “DM me,” just a paragraph that might help if they never speak to me.
Only then did I post a small, honest request. I wrote it like an invitation to compare notes, not a funnel.
“I’m mapping how [your role] deals with [this mess] right now. Ten minutes, no pitch. I’ll share the patterns back next week so the thread has something useful to point to. If that’s you, I’d appreciate three quick questions.”
Two people replied under the post. I DM’d the ones who left smart comments. I didn’t copy-paste. I referenced what they said when I wrote. That alone separated me from the spam.
On day three, a stranger wrote, “Sure. I have ten minutes between meetings. Call me then.” I walked around the block during the chat because I was too nervous to sit still. When she mentioned the specific Slack channel where they track issues, I asked if I could scribble the name down and promised not to quote it. She laughed and said, “Please don’t. But yes, it’s the place where problems go to die.” Another phrase for the notebook.
In the old forum, I emailed the moderator to ask whether a short research post would be allowed. He said yes, if I shared what I learned back on the thread without asking people to sign up for anything. That constraint turned out to be a gift. I wrote a two-page field note and posted it publicly, redacting names and companies, and the summary became the reason others said yes the following week. Trust compounds faster than reach.
I learned to write like someone who is grateful for the next person’s time. I wrote lines like, “If this isn’t you, thanks for reading anyway,” and I meant it. That sincerity kept doors open.
The gym week (or, how awkward turns into data)
A friend dared me to leave my apartment. I printed a one-page sheet with three questions, grabbed a clipboard, and camped out near the squat racks of the cheapest gym in town. Between sets, I asked for three minutes. I offered a coffee for their trouble, which half refused and the other half happily accepted. I misspelled “interview” on the flyer. Someone teased me; I took the joke; we ended up talking for twelve minutes.
I learned that asking “Do you mind if I ask you three questions?” is ten times easier than “Can I interview you?” The word “interview” sounds like a thesis. “Three questions” sounds like a favor. After the first two hours—which were mortifying—I stopped apologizing for existing and focused on clarity. People respond to sincerity more than perfection.
By Friday I had forty short conversations and several longer ones scheduled. The patterns were obvious when I stared at them on paper: the same workaround appearing across jobs that would never talk to each other, the same complaint wearing different clothes. An engineer and a florist both used “Sunday night guilt list” for the chores they avoid. If two worlds use the same phrase for the same pain, pay attention.
Noticing these echoes changed the way I wrote the next week’s outreach. I stopped describing the problem as a category and started describing it as a moment. Not “inventory forecasting,” but “the five minutes right after the supplier says they’re out.” Not “patient retention,” but “the Thursday tumble when your phone won’t stop ringing.” Moments are sticky; categories are homework.
Cold outreach that didn’t feel gross
I wrote emails in the same hallway voice I’d use if I bumped into someone at a conference. Not a funnel. Not a sequence. A note.
Subject lines that worked sounded like reasons, not tricks: “Quick 10-min chat about your [post on scheduling chaos],” “Sanity check on [inventory updates]?” The body was short and specific.
Hi [Name] — I’m trying to understand how people actually handle [problem] today.
Not selling anything. If you’re up for 10 minutes, I’ll share the patterns I hear back.
Either way, thanks for your post on [specific detail] — it clarified a blind spot for me.
— Sofia
If they ignored it, I followed up once three to five days later with one sentence: “Bumping this once in case timing was off — totally fine to ignore.” That second note got more replies than I expected because it sounded like something a person writes when they’re trying not to be annoying. When someone replied, “Email questions are easier,” I sent three questions, then stopped after they answered to avoid turning their inbox into a homework assignment. Many of those email threads turned into Zoom calls, but even when they didn’t, they were full of screenshots and phrases that no cold Zoom call would have produced.
For local businesses, I picked up the phone. I called a small mechanic shop, a clinic, and a bakery. “I’m not selling anything; I’m trying to map how shops like yours handle [problem]. Three questions, five minutes?” The ones who said no were still kind. The ones who said yes were gold. The mechanic told me about the “clipboard of shame” where they track parts they forget to reorder. The bakery had a “black book” where one person kept every vendor phone number because software had failed them so many times they trusted paper more than login screens. The clinic manager introduced me to the phrase “Thursday Tumble,” which I would never have invented on my own.
The point isn’t to collect cute phrases. It’s to see how work actually happens when no one’s watching. And it’s to monitor your hit rate with honesty. In the first week, one in ten replied. By week four, one in four did—not because I wrote better, but because I was referencing their world with specificity. “Your line about the clipboard of shame…” is harder to ignore than “I noticed you are a small business owner.”
I kept a tiny scoreboard in a corner of my notes: outreach sent, interviews booked, done, referrals asked/got, new opt-in emails. On good days, the numbers made me smile. On bad days, they made me do one more ask before I closed the laptop.
Borrowed trust beats borrowed reach
I didn’t have an audience, but other people did. The coworking space down the street added my call for interviews to their newsletter because I asked nicely and offered to share the write-up back. The copy was two sentences long and didn’t sound like marketing. A professor forwarded my note to a class because I promised to do a quick guest session on “what I learned,” and then I actually showed up. A small industry association dropped a link in their member Slack in exchange for early access to the final summary. That summary wasn’t a tease for a product; it was an artifact they could reference without sending their members into anybody’s funnel.
The day the coworking newsletter went out, two replies landed within an hour. One said, “Happy to chat, but please don’t make me fill out a form.” I didn’t. We spoke the next day. After the call, I sent her a short thank-you and a note that included one sentence she had used. She replied, “You’re the first person who quoted me back to myself in a way that didn’t feel icky.” That’s the bar: carry their words carefully.
Every borrowed channel came with a condition: respect their people. No spam, no surprise pitches, no weird “gotchas.” When someone introduced me, I sent a thank-you note with a write-up one week later, and it made the next introduction easier. If you keep that promise three times in a row, people start asking you for the summary. That’s the first version of an audience that matters—two hundred people who remember you as the person who listens, not ten thousand who don’t.
When I paid for answers (and why it was worth it)
Around interview sixty, I hit a gap I couldn’t bridge with more hustle. A segment I cared about didn’t hang out in my channels. I used a small paid panel to validate a couple of patterns. Prolific worked for consumer-ish profiles; UserInterviews did too for B2B roles. I kept the survey short—eight questions, tops—and asked for permission to follow up. It cost a few hundred dollars. It saved three weeks of guessing.
Screeners matter. If you don’t ask something that only the right people would know, you’ll get thoughtful answers from the wrong crowd. I used a short scenario to weed out tourists: “When you see [trigger], what’s your first move: [X], [Y], [Z], or something else?” People who recognized the trigger and gave me “something else” with a specific noun were the ones who taught me the most.
One respondent wrote, “We check the ‘red folder’ before we touch the software because we don’t trust the integration.” I asked for ten minutes to go deeper, reminding them they’d already been paid for the survey and this was optional. They said yes because I treated them like a person, not a sample.
Keeping the thread without turning into a robot
I tracked conversations in a simple sheet because complex CRM tools made me feel like I was doing sales instead of learning. Date, name, role, where I found them, how often the pain showed up, what they tried before, tools they used, the exact phrase they used when they got annoyed, whether I could email again, and who they referred. The “exact phrase” column paid for itself when it was time to write the landing page. “Clipboard of shame,” “Thursday Tumble,” “black book,” “Sunday guilt list.” That column was a mirror you can hold up to your own copy and ask, “Do I sound like them or like the brochure inside my head?”
Every Friday, I sent a one-page summary to everyone who helped, even if we only traded three emails. Patterns, contradictions, and one quote (anonymized) that made me laugh. That email did two things at once: it closed the loop and quietly grew a permissioned list of people who wanted to hear what came next. I added a single line at the bottom for consent: “If it’s useful, I’ll share next month’s notes too—reply ‘remove’ anytime.” Some did. Most didn’t. The list grew without any contests or pop-ups.
Compliance isn’t sexy, but it builds trust. Ask for permission to contact again. If you’re in the EU (or storing personal data), note your lawful basis and don’t be cute with it. And for your own sanity, don’t keep email addresses in three places. Pick one sheet you won’t lose.
The weeks, written like a journal instead of a plan
Week one was foundation. I picked three communities, two physical places, and wrote three script variants (forum, email, in-person). My goal wasn’t perfection; it was twenty-five conversations on the calendar or completed. I hit twenty-seven, most of them short, some of them messy. I learned that my intro was too clever by half and that the words “no pitch” made people sigh with relief.
Week two was outbound and events. Ten short emails a day, a couple of DMs in each community, one meetup where I talked to people who would never see my LinkedIn. I asked for two referrals per conversation and discovered the way you ask changes the outcome. “Do you know anyone?” yields shrugs. “Is there one person who comes to mind?” yields names. That pushed me past forty, and I started getting introductions from people I hadn’t asked yet because they forwarded the Friday summary to colleagues.
Week three filled the gaps. I ran the small paid panel and tried a scrappy experiment—flyers with a QR code in two places my people actually go. One of those flyers led to a coffee with someone who became an advisor, the kind who will text you “ship the email” at 10 p.m. when you’re overthinking it. I learned that QR codes work when the context makes sense (the corkboard in a lab) and look like spam when it doesn’t (the lamp post near a park).
Week four was synthesis. I chased the last referrals, wrote a two-page field note with the top three problems (in their words), the common workarounds, the buying constraints that would block me later, and the triggers that made people act. I forced myself to choose one audience to write for and let the others be second. That hurt. It also made the page clearer. I turned the “yes, you can email me again” list into the first version of an early-access list. Not a vanity number. A group of people who would reply if I sent something dumb.
On the last day of the month, I scrolled through my notes and noticed something quiet but important: I wasn’t dreading the next conversation. I was curious again. Curiosity is a better fuel than fear.
Micro-tactics that worked exactly once (and I kept anyway)
Once, I wrote someone’s exact quote on a sticky note and taped it to my laptop where I could see it on calls: “I don’t need another tool; I need the five minutes back when the phone rings.” The sticky note grounded me when I drifted into feature talk. Another time, I started a message with, “I’m going to lose a small bet if I don’t get one interview by Friday. Saves me from buying dinner. Ten minutes?” The person replied, “Fine, I’ll save you the entrée—3 p.m.?” Humor can carry you across the awkward.
In a Slack community, I asked the admins if I could run a lightweight AMA about “dumb mistakes I made trying to interview strangers.” They said yes because it didn’t sound like an ad. Ten people came. Three of them turned into interviews. One became a friend I still send drafts to. I’m mentioning this because we tell ourselves communities are hostile to beginners. They aren’t, if you show up with humility and useful notes instead of a sign-up link.
When I felt stuck, I walked to the corner café with a small sign that said, “Three questions about [problem] — coffee on me.” I expected to be ignored. A grad student sat down and gave me twenty minutes I didn’t deserve. He sketched a flow on a napkin that explained a failure mode I’d been mislabeling for two weeks. It would have taken me a month to find the exact words he used anywhere else.
Turning noise into signal
Synthesis is the part nobody tweets about because it looks like staring. Here’s what helped: I scored problems by frequency times intensity. Frequency came from how often people told me the thing shows up in a week. Intensity I stole from therapists and support teams: watch for the sigh and the swear, and ask what it costs. “Costs” is money, but also time and mood and political capital inside a team. If a problem shows up three times a week and makes someone resent their job, it’s bigger than a quarterly annoyance that costs $300 once a year.
I built a language list. Every synonym for the core pain lived in one column. Next to it, I wrote the sanitized version I would have used before interviews. When my sanitized words didn’t match any synonyms, I stopped using them. That’s how “reduce cancellations” became “tame the Thursday Tumble.” That one phrase made a landing page feel like an inside joke to the right people, which is what good copy often is.
Here’s a before/after from my own notes:
Before: “Automate schedule optimization to improve clinic throughput.”
After: “Stop playing calendar Tetris every Thursday afternoon.”
You can guess which one got replies.
I also kept a “do not build” list, a place to quarantine features that made me feel clever but didn’t show up in anyone’s day. That list grew longer than the backlog and saved me from explaining detours to future-me.
A few questions people asked me along the way
What if people ask to see the product in the first call? Give them a sentence and an option. “I’ve got a rough cut I’m happy to show you, but I promised myself not to hijack learning time. If you’re up for a second call, I’ll bring it, and you can tear it apart.” The people who say yes will give you better feedback than a drive-by demo, and the ones who say no helped you anyway.
How do you start when you don’t know the community’s vocabulary? Read for two nights. Write down ten nouns you don’t recognize. Search those nouns plus “help” or “workaround.” Your third night, answer one question with something you wish you’d found on night one. That’s your ticket into a real conversation.
What do you do with “maybe”? Translate it. “Maybe” often means “I’m being polite” or “I don’t feel the problem right now.” Ask, “When is this not a problem?” Their answer will tell you whether they’re the wrong person or you’re too early. Thank them either way.
When do you stop interviewing and build? When your last three interviews produce fewer than two new nouns or verbs each. When you can predict the next sentence before they say it—and you’re right. When you can write a one-paragraph description of the problem in their words and it reads like gossip from their office. That’s enough to build.
Aren’t you worried about detectors flagging the style of your post? I’m more worried about writing like a brochure. Brochures sound like nobody. People sound like themselves. If you write down what you actually heard, with the smells and the mess and the timing, you’ll sound like someone who was in the room. That’s hard to fake because it requires being in the room.
What I sent back
People are busy; they say yes because you make it easy to be helpful. I kept my promise: a short summary, one interesting contradiction, and—if they wanted—early access to the next thing. A few asked for the deck. I didn’t have one. I sent a page of notes with bullets only where bullets made reading easier. It was enough.
A typical Friday note looked like this (anonymized):
“Quick field notes from this week: clinics call Thursday ‘the tumble’ and rely on one person who knows which patients can be rescheduled without chaos; makers keep a ‘black book’ because logins slow them down; across both, the first workaround is always a person and a piece of paper. My favorite quote: ‘I don’t need another tool, I need five minutes back when the phone rings.’ If you want next week’s, tell me. If not, thanks for helping me learn.”
Half the replies were one-liners: “Keep me on.” That’s an email list built out of conversations, not pop-ups.
One month in, those Friday notes had more readers than my social feeds ever did. And they were the right readers—the ones who would reply with, “This part doesn’t ring true,” and make the work better.
Two notes on ethics
First, research isn’t a loophole for spam. If a community says no, thank the admin and leave it. If a person says they don’t have time, don’t argue. The trust you save by walking away is worth more than the one interview you might bully your way into.
Second, anonymity is armor. Share patterns, not details that could embarrass someone if their boss read your post. People will tell you honest things if they believe you’ll carry them carefully. That belief is harder to earn than a follower and much harder to regain if you break it.
If you have 30 minutes a day versus three hours
When I had thirty minutes, I picked one channel, sent two messages—one new, one follow-up—and wrote down anything I learned that would make tomorrow’s note more specific. On weekends, I updated the sheet and wrote the Friday summary. It wasn’t heroic; it was a habit.
When I had three hours, I layered channels. A handful of DMs, a guest line in a small newsletter, a walk to the gym with a clipboard, and two short calls. The point isn’t to do everything; it’s to always be learning from a real person instead of from your imagination.
A small appendix of scripts, written like you talk
An email that didn’t embarrass me later:
Subject: Quick 10-min sanity check on your “Thursday Tumble”
Hi Marta — Your note about the Thursday cancellations reminded me of my own stack of reschedule texts.
I’m mapping how people actually handle that scramble. Not selling anything.
If you’ve got 10 mins, I’ll share back what I learn next week.
Either way, thanks for naming it. I’ve been calling it calendar Tetris; “Tumble” is better.
— Sofia
A DM that started a helpful thread:
“Your screenshot of the whiteboard solved a mystery for me—why the software is always a step behind the person with the pen. Do you have 10 minutes to teach me what the pen knows that the tool never does? I’ll write it up and send it back.”
A phone opener that didn’t get me hung up on immediately:
“Hi, sorry for the cold call. I’m a founder trying to map how shops like yours handle [problem]. I’m not selling anything. Could I ask you three quick questions? I’ll be out of your hair in five.”
If you’re allergic to scripts, good. Read these once, steal one sentence, and then say it your way.
Closing note from the quiet side of the internet
The founders who win aren’t always the loud ones. They’re the ones with the clearest notes. You don’t need followers to earn those notes. You need a handful of honest conversations, a spreadsheet that won’t embarrass you later, and the nerve to ask for ten minutes from a stranger.
Start with one message today. Then another tomorrow. Somewhere between conversation thirty and conversation sixty, you’ll notice something strange: you’re no longer guessing. And by the time you hit a hundred, you’ll have something better than an audience. You’ll have people who are waiting to see what you build next—because they helped you make it.