The standard advice on customer interviews is some version of just talk to people in your network. If you read enough product blogs you'll see this restated a dozen ways. Reach out to your existing contacts. Ask for warm intros. Post on LinkedIn. Lean on your professional connections.
For a meaningful slice of founders, this advice is useless. If you spent ten years as a software engineer and you're now building a tool for veterinary clinics, your network has zero veterinarians in it. The network advice assumes founder-market fit at exactly the point you're trying to figure out whether you have it. If you're a solo founder building for a market you don't already live inside, the recruiting problem is the first real barrier between you and validation. It's also the barrier that stops most people from doing customer interviews at all. What follows is what actually works when you're starting from zero, with no customer panel, no email list, and no warm intros.
Why this is harder than people make it sound
In 2014, an entrepreneur named Justin Mares wanted to start a bone broth company. He didn't have a food industry network. What he did have was a conviction that there was demand for bone broth in specific health communities, and the patience to spend weeks reading Reddit threads, Mark's Daily Apple comments, and PaleoHacks discussions before he ever posted anything. He didn't survey people. He didn't pitch a product. He read what people were complaining about — the difficulty of sourcing high-quality bone broth, the price gouging on the few brands that existed, the time it took to make at home — and listened for patterns. That listening was the validation. The company he started, Kettle & Fire, has since done over $100M in revenue.
The thing worth noticing about that story isn't that Reddit is magic. It's that Mares spent more time reading than asking. Most founders skip the listening phase entirely. They write a LinkedIn post titled "Hey founders, I'm researching a new tool — would you spare 15 minutes?" and wonder why nobody replies.
The real reason cold outreach fails for early-stage founders isn't the channel. It's that the message goes out before the founder has done any homework on what their target market actually says, where they hang out, or what would make a stranger consider their fifteen-minute ask worth taking. The recruiting problem is downstream of the homework problem.
Five paths that actually work
In practice, there are five ways solo founders without a network find people to interview. Most founders end up using a mix of three or four.
Cold outreach on LinkedIn or email is the obvious first move. Done well, cold LinkedIn messages get reply rates in the 7 to 15 percent range; with real personalization, top-quartile messages clear 25 percent. To get 10 actual interviews from cold outreach, you'll need to send maybe 80 to 150 targeted messages, of which 20 to 40 will reply, of which 10 to 15 will agree to a call, of which 10 will actually show up. There isn't a way around that math; the volume is the price of admission.
The script that works is short (under five sentences), grounded in their world before your ask, and ends with something so small that saying yes costs almost nothing. A version that's worked for me and the founders I've watched do this well:
Hi [name], I'm researching how [specific role] handles [specific problem you've already done homework on]. Not selling anything; I'm building a tool and want to understand the workflow first. Would you have 15 minutes this week or next? Happy to send my questions in advance so you can decide whether it's worth your time.
The happy to send my questions in advance line is more powerful than it looks. It moves the ask from can I have your time to here's evidence I won't waste it, and it lets the recipient self-select before they commit. Most won't ask for the questions. They'll just say yes.
Community immersion is the slower play, and it's the Justin Mares move. Find the three or four subreddits, Discord servers, Slack groups, or niche forums where your target customer is already complaining about the problem in their own words. Spend two weeks reading before you post a single thing. Answer questions when you can answer them honestly. When you do post a research request, do it with full disclosure: who you are, what you're working on, why you're asking, and what's in it for the person on the other end. A short Loom of you on camera goes a long way. Anonymity reads as a sales pitch in disguise; faces don't.
The communities to target are smaller than founders expect. A subreddit with 8,000 active members where everyone recognizes everyone is more useful than one with 800,000 where you'll get drowned out in twenty minutes. The signal that you've found a good one is that the same usernames keep appearing in the threads relevant to your problem.
Paid recruitment exists for a reason. When you've genuinely tried the above and still can't reach the person you need, pay for it. UserInterviews and Respondent are the two main panels. As of early 2026, UserInterviews charges $49 per recruited B2C participant and $98 for B2B on their pay-as-you-go plan, before incentives. Add typical incentives of $100 to $200 per hour and your fully-loaded cost lands somewhere between $150 and $300 per completed interview. Respondent takes 50 percent of the participant incentive with a $40 minimum and tends to be stronger for hard-to-reach professionals: surgeons, ops leaders at specific company sizes, people in regulated industries.
The argument for paying is straightforward. If you're trying to validate a B2B tool for dental practice managers, you can easily spend three weeks on LinkedIn for fewer than five interviews. The same three weeks plus around $1,500 on UserInterviews gets you 10 dental practice managers on the calendar. Time is the scarce resource for a solo founder, not money. Pay for it when the math works.
There's a real catch, though. Incentivized participants are slightly different from real prospective customers. They're showing up partly for the money, and a chunk of them are professional survey-takers (the public-facing acceptance rate on UserInterviews is around 3.4 percent, which tells you how many people are gunning for these). The signal is still usable, but you should treat what they say more skeptically than what you'd hear from someone who agreed to talk to you for free, and you should lean harder on questions about past behavior than questions about hypothetical interest.
The most underused path is proxy experts. If you can't easily reach your target customer, find the people who serve your target customer professionally: accountants who specialize in restaurants, consultants who work with veterinary clinics, recruiters who place data engineers, and so on. These people talk to dozens of your target customers every month and have already aggregated patterns across them. They're also far easier to find on LinkedIn than the customers themselves, because they advertise their specialization right in their profile.
One conversation with a proxy expert can save you ten interviews. You won't replace direct customer conversations with this, but you'll learn what to ask in them.
The slowest path is to trade access for value. Publish a useful resource that your target market actually wants. Sponsor a small meetup. Do unpaid work for one person in the space in exchange for an honest conversation later. Write three thoughtful answers a day in the niche subreddit. None of these converts in a week. All of them, over a month or two, generate inbound interest from exactly the people you've been trying to reach. The trade-access route produces the warmest contacts of any of the five, and it builds something that compounds.
What to do about no-shows
Even with the perfect script, 20 to 30 percent of confirmed interviews will ghost. Founders take this personally. They shouldn't. It's the baseline cost of doing this at all.
A few small things move the number, none of them complicated. Send a calendar invite immediately after they agree, with the call link already in it. If they don't accept the invite within a day, they probably aren't showing up and you should re-confirm. A one-line reminder the day before also helps. And if you've offered any incentive, mention it casually in the reminder, something like looking forward to the chat tomorrow, will send the $20 Amazon gift card right after we wrap.
For everyone else, accept the loss and move on. The fastest way to never recover from a ghost is to spend mental energy on it. Send the next outreach message. The funnel is the funnel.
One last thing
Bandos exists because doing this in the right order is hard, and most founders skip the homework that makes recruiting actually work. The product walks you through it: idea, customer definition, need mapping, and then it generates the Mom-Test-disciplined interview script you'd run with the people you recruit through the methods above. When responses come back, they're scored against the criteria you set ahead of time, so you're not relying on memory of the conversation to draw conclusions. You can see how the whole flow works or start a project for free.
The hardest part of recruiting interviews isn't actually picking the right channel; it's getting yourself to do it at all when nobody is making you. Most founders skip customer interviews not because they don't believe in them, but because the recruiting feels too hard to start and the homework phase feels too slow to be productive.
Send ten messages this week. Not perfect ones, just ten. Some will reply.