Beta testers
How to Find Beta Testers for a Startup
The prototype works on your machine. Friends and family have said it looks promising. The question every pre-launch founder hits is the one the publication has not yet answered: where do I find the first 10–50 people who will use a half-built product, tell me what is broken, and not pretend to like it because they want to be polite?
Last updated · September 11, 2026
Quick answer
How do you find beta testers for a startup?
A beta tester is a person who uses the unfinished product, runs the workflow it is supposed to support, and reports what broke, what worked, and what they would change. They are not focus-group respondents and not friends-and-family approvers. The fastest way to find them is the Mom Test rule: ask about commitments and specifics, not opinions and hypotheticals. The seven primary channels — direct outreach to your own network, niche communities where the buyer already congregates, the launch platforms Product Hunt and BetaList, conferences and meetups where the persona attends, existing users of a competing product, paid acquisition at small budget, and existing customers from a prior venture — each produce a different shape of tester. The doctrine for the first ten is Paul Graham's 'Do Things That Don't Scale' (2013): recruit them by hand, one at a time, until the cohort is large enough that a real acquisition channel starts to work. The disqualifying filters are people who will not use it in their own workflow, people whose commitment is verbal only, and people who have not paid for or worked around a similar solution.
Key takeaways
The playbook in 30 seconds
- A beta tester is a person who will actually use the unfinished product in their own workflow, not a friend who is being polite.
- Apply the Mom Test (Rob Fitzpatrick, 2013) to recruitment: ask about commitments and specifics, not opinions and hypotheticals.
- Seven primary channels work — your own network, niche communities, Product Hunt, BetaList, conferences, existing users of a competitor, paid acquisition — and each produces a different shape of tester.
- For the first ten, do not scale: recruit by hand, one at a time. The doctrine is Paul Graham's 'Do Things That Don't Scale' (2013).
- Three filters disqualify the wrong tester: they will not use it in their own workflow, their commitment is verbal only, or they have not paid for or worked around a similar problem.
- Five recurring failure modes: friends-and-family as data, asking strangers who have no reason to care, free access as the only incentive, no follow-up after signup, and treating the cohort as a launch announcement instead of a working group.
Definition
What a beta tester actually is
A beta tester is a person who will use the unfinished product in the same workflow the production version is meant to support, who will report what is broken, what is confusing, and what is missing, and who is not invested in the founder's success. The 'use it in their own workflow' clause is the disqualifier that filters out most of the wrong candidates. A tester who will not run the product as part of their own day is a tester who cannot tell you whether the product solves the problem, only whether the product opens. The Mom Test framing — Rob Fitzpatrick's 2013 book on how to talk to customers — describes a tester as someone whose feedback you can trust precisely because they have nothing to gain by giving it. A tester who is also an investor, a co-founder's relative, or a founder who is building a similar product is a tester whose signal is corrupted by an incentive other than the truth.
What a beta tester is not
A beta tester is not a focus-group respondent, not a Twitter follower, not a friend or family member being supportive, not a person who has signed up but never opened the link, and not a person who has installed the app but never completed the workflow it was built for. A list of email addresses from a giveaway is not a beta. A waitlist of curious strangers is not a beta.
What a beta tester is
A beta tester is a person who runs the product in the workflow they would use the production version in, who reports what they observed, and who can be re-contacted. The number of testers you need depends on the breadth of the workflow — one tester per distinct workflow the product covers is a rough floor.
The Mom Test rule
Apply the Mom Test to recruitment, not only to interviews
The Mom Test (Rob Fitzpatrick, 2013) was written for customer interviews, but the same three rules govern how you ask a stranger to be your beta tester. The shift is from asking 'would you use this?' to asking about a commitment the stranger has already made or is about to make.
- 1
Ask about their last similar commitment
Instead of 'would you try this?', ask 'when was the last time you tried to solve this problem, and what did you use?' A real answer — a specific tool, a specific date, a specific frustration — is a tester who has lived the problem. A vague answer is a tester who has not.
- 2
Ask about specifics, not opinions
Instead of 'what do you think of this idea?', ask 'walk me through what you would do today, step by step, when this problem comes up.' Specifics tell you whether the tester is the right persona; opinions tell you only whether the tester is being polite.
- 3
Make a small commitment before the test starts
Before sending the beta link, ask the tester to commit to a 15-minute call in the next seven days where they will show you what they did with the product. Verbal 'sure, I'll try it' is not a commitment. A booked calendar slot is a commitment. The commitment is the filter that separates beta testers from free signups.
Rob Fitzpatrick's companion essay on this topic is published at momtestbook.com/how-to-find-beta-testers/. The essay and the book together are the canonical primary source for the discipline; the channels below should be read as the place to apply the rule, not as a substitute for it.
Channels
Seven channels for finding beta testers, with primary-source notes
Each channel produces a different shape of tester. None of them is sufficient on its own for the first 10; the right combination depends on the buyer and the workflow. The channels are ordered by cost: cheapest (your own network) first, most expensive (paid acquisition) last.
Your own network — first-degree contacts and second-degree intros
Start by listing every first-degree contact who fits the buyer persona, plus the second-degree contacts who could introduce you to one. Send each of them the recruitment message in the next section. Expect a 10–30% positive response rate; the rest either do not fit the persona or are too busy. This is the cheapest channel and the one that produces the highest signal for the first three to five testers, because you can run a Mom Test conversation before sending the beta link.
Primary source: Paul Graham, 'Do Things That Don't Scale' (2013), paulgraham.com/ds.html — the essay argues that the first ten users of any startup come from manual outreach.
Niche communities — forums, subreddits, Slack groups, Discord servers
Find the specific community where the buyer already congregates. For B2B buyers this is usually a niche Slack or Discord; for B2C it is usually a subreddit or a forum. Read the rules of the community before posting — most communities prohibit direct promotion, and breaking the rules produces a ban. The right move is to participate in conversations for two to four weeks before recruiting, and to ask permission to share the beta in the context of a thread the community already cares about.
No single platform publishes the rules across all communities; verify each community's sidebar, rules page, or moderator guidance before posting.
Existing users of a competing product
If a competitor exists, their user base is the most qualified tester pool you will find. The recruiting move is direct: find the competitor's user-facing surface (a forum, a feature-request board, a community), identify the users who publicly describe the problem your product solves, and send them a one-paragraph message that names the problem in their words. The advantage is that the tester has already voted with their time that the problem is real; the cost is that the message has to be specific enough that the tester does not file it as spam.
Primary source: Fitzpatrick, The Mom Test (2013) — the rule 'ask about their last commitment' applies; users of a competitor have already made the commitment the beta test needs to verify.
Product Hunt — pre-launch 'coming soon' pages
Product Hunt operates a 'coming soon' surface for products that have not yet launched. A listing there is a single-page announcement with an email signup. The platform's own launch guide (producthunt.co/launch) notes that the surface is community-driven; a listing does not guarantee visibility, and upvotes cannot be solicited directly. Treat a Product Hunt 'coming soon' page as a passive channel — useful as a second source of qualified signups, not as a primary one.
Primary source: producthunt.co/launch — Product Hunt's own published launch mechanics.
BetaList — paid submission for a featured slot
BetaList publishes submission criteria (betalist.com/criteria) that constrain what can be listed: the product must be relatively new (private beta, recently launched, or unreleased), the listing must point to a custom landing page (not a free-hosting subdomain, not a LaunchRock-style page), and the listing must give visitors a way to sign up or get access. BetaList is a paid directory with a fast-track option; treat it as a distribution surface for an already-built landing page, not as a recruitment channel for an empty one.
Primary source: betalist.com/criteria — BetaList's own published submission rules. Always verify the current rules page before submitting, since submission criteria change.
Conferences, meetups, and trade shows
The persona-attended conference is a high-signal channel for B2B testers. The cost is the registration fee plus travel; the advantage is that the tester has flown to a conference to think about the problem your product solves. The recruiting move is to attend the conference with a working beta and run three to five 20-minute sessions with people who walk up. This is the channel Graham describes as 'recruit by hand, one at a time' in his 2013 essay.
Primary source: Paul Graham, 'Do Things That Don't Scale' (2013) — the same essay treats in-person recruitment as the standard mechanism for the first ten users.
Paid acquisition at small budget
Paid acquisition — Google Ads, LinkedIn Ads, Reddit Ads, sponsored newsletters — is the most expensive channel and the one that produces the lowest-quality tester. The testers it produces are strangers who clicked because the ad copy was compelling, not because they have lived the problem. The channel has a place once the landing page is converting at a known rate, but it is the wrong first move. Run the free channels first.
Caveat: the conversion quality of paid-acquisition testers depends on the ad targeting and the landing page; treat the channel as a last resort for the first cohort.
The recruitment message
How to write the message that asks a stranger to test
The recruitment message is the highest-leverage piece of writing the founder produces in the beta-recruitment phase. A message that asks about commitments, names the problem in the buyer's words, and makes the ask specific will produce a different cohort than a message that asks about opinions and offers free access.
Template
Subject: 15 minutes next week? (and a beta) Hi [name], You posted about [specific problem they described in their own words]. I am building [product] for [same persona] who run into [same problem]. The beta is rough; I would rather hear what is broken than what looks good. Two asks, you can pick one: 1. A 20-minute call next week where you try the beta live and tell me what is broken. 2. A short reply with the last time you tried to solve [specific problem] and what you used. Either is genuinely useful. If neither fits, no reply is also fine. [your name]
- Name the specific problem the tester described in their own words. Generic copy ('for busy professionals') produces generic testers.
- Offer two specific asks (a call, a reply) instead of one open-ended ask. Two specific asks make it easy for the right tester to pick one; one open-ended ask makes it easy to ignore.
- Say that no reply is fine. The line removes the social pressure that produces polite non-responses and makes the positive responses more honest.
- Do not lead with 'free'. Free access is an incentive that attracts testers who want the product for free, not testers who want to solve the problem. The ask is for feedback, not for charity.
The first ten
Recruit the first ten by hand. Do not scale.
Paul Graham's 'Do Things That Don't Scale' (2013) is the most-cited essay on the topic, and its core argument is unchanged: the first ten users of a startup do not come from a channel, they come from manual outreach. The reason is not that channels do not work — they do, later — but that the first ten are the ones who teach you what the product needs to become, and a channel cannot teach you that, only a conversation can. For the first ten, the right move is to combine two channels: your own network (channel 1) and niche communities (channel 2). Both are channels you can run without a budget; both produce testers you can talk to; both let you apply the Mom Test rule before sending the beta link. The cohort is small on purpose. The cohort is also a working group, not a launch announcement — they should expect to be contacted again, and the founder should expect to do the contacting.
Manually recruit — no channel for the first ten
The recruitment is a list of specific people, not a campaign. Each name on the list is a person you could call. Send each of them the recruitment message above. Expect a positive response from a third of them. The total volume is small; the signal is high.
Run a Mom Test conversation before sending the link
Use the first contact to verify the tester fits the persona and has lived the problem. The conversation is the Mom Test: ask about their last similar commitment, ask about specifics, and make the beta commitment small and specific. The conversation is also where the disqualifying filters below get applied.
Run the beta live, not asynchronously
For the first ten, watch the tester use the product. Twenty-minute live sessions produce ten times more signal than an asynchronous signup with a feedback form. The session is also where the founder learns whether the workflow is broken end-to-end, which an asynchronous tester will silently skip.
Re-contact every tester within seven days
The first ten are a working group, not a launch announcement. Re-contact them in the first week with a question, an update, or a fix. The re-contact is the discipline that turns a signup list into a beta cohort; without it, the cohort decays into a silent install base.
Qualifying filters
Three filters that disqualify the wrong tester
Most beta-recruitment failures come from accepting testers who should have been filtered out. The three filters below are independent; each one is a reason to remove the tester from the cohort.
They will not use it in their own workflow
A tester who will not run the product as part of their own day is a tester who can tell you whether the product opens, not whether the product solves the problem. The disqualifying signal is a tester who asks 'when is the next version?' or 'when does this launch?' — both indicate the tester is evaluating, not using.
Their commitment is verbal only
A tester who says 'sure, I'll try it' is not a tester who has tried it. The Mom Test rule — make a small commitment before the test starts — applies here. The commitment can be a booked call, a signed form, or a deposit; the point is that the test starts with a real action, not a verbal intention.
They have not paid for or worked around a similar problem
A tester who has not paid for a competing product, hired a contractor, or built a manual workaround for the problem does not have a felt need for your solution. Their feedback is about whether the product is good; it is not about whether the product is needed. The test you want is one where the tester has a stake in the answer.
Common mistakes
Five mistakes that turn a beta list into a silent install base
Each mistake below is a way to produce a signup list that looks like a beta cohort but is not. The cost of the mistake is not the time spent recruiting; it is the months of product work built on feedback from the wrong testers.
Friends and family as testers
Friends and family will say the product is great because they want to support you, not because they would use it. The signal is corrupted by an incentive other than the truth. Rob Fitzpatrick's Mom Test (2013) treats this as the canonical example of bad customer research; the same logic applies to beta recruitment.
Asking strangers who have no reason to care
A recruitment message that asks a stranger to test a product they have no stake in is a recruitment message that produces polite non-responses. The ask has to be specific to the persona and the problem; a generic 'we are looking for beta testers' produces a generic non-response.
Free access as the only incentive
Free access attracts testers who want the product for free, not testers who want to solve the problem. The first group produces feedback about features they would like; the second group produces feedback about whether the product solves the problem. The ask is for feedback, not for charity.
No follow-up after signup
A tester who signs up and hears nothing is a tester who has stopped using the product. Re-contact within seven days. The re-contact can be a question, an update, or a fix; the point is that the cohort is a working group, not a launch announcement.
Treating the cohort as a launch announcement
A beta cohort is a working group, not a launch list. The cohort is small on purpose; the cohort expects to be re-contacted; the founder expects to do the re-contacting. A cohort that is treated as an announcement will decay into a silent install base within thirty days.
How Yibud helps
How Yibud connects to finding beta testers
Yibud's tools surface whether the persona, the problem, and the channel are aligned before you spend weeks recruiting the wrong testers. The eight dimensions of the Startup MRI score are the inputs the recruitment message depends on; the persona in particular is the variable the channel has to match.
Check whether the persona is in the inputs
The startup idea evaluator scores the eight dimensions from the inputs you provide — idea, audience, monetization, acquisition channel, and technical background. A weak market score or a weak distribution score usually means the persona is not yet defined sharply enough for any channel to recruit against. Fix the persona first; then start the recruitment.
Open the evaluator →Read the customer discovery playbook
The customer discovery page documents the broader discipline — how to find the right buyers, how to run interviews that do not lie, and how to know when you have enough signal. Beta recruitment is the last step in that discipline; the page is the playbook for the steps before.
Open customer discovery →Decide what the cohort tells you
The decision guide covers what to do with the evidence a beta cohort produces — when the signal supports continuing, when it points to a pivot, and when it points to a kill. The cohort is one input to that decision; the guide is the next read after the beta runs.
Open the decision guide →FAQ
Frequently asked questions
How many beta testers do I need?
Roughly one tester per distinct workflow the product covers. A product with one core workflow can produce useful signal from three to five testers; a product with five workflows needs at least one per workflow, and the cohort grows from there. The number is not the goal — coverage of the workflows is the goal.
What if I cannot find anyone who fits the persona?
That is the answer. A persona that produces no testers in your network and no testers in the niche communities where the buyer already congregates is a persona that has not been validated. The right move is to revisit the persona definition and the problem statement, not to lower the recruiting bar.
Should I pay beta testers?
Generally no — payment attracts testers who want the payment, not testers who have lived the problem. The exceptions are: a tester who would otherwise be unable to take the call (a consultant billing by the hour, a senior practitioner with limited time), and a tester whose specific expertise is required for the workflow. In those cases, the payment is a budget for access, not a substitute for the Mom Test conversation.
How long should the beta run?
Long enough for the tester to use the product in their own workflow across at least two cycles of the workflow, and short enough that the cohort does not decay into a silent install base. Two to four weeks is the standard window for a beta; longer is a soft launch, not a beta.
Should I run a private beta or a public beta?
Private beta for the first ten. Public beta only after the private beta has produced a stable workflow and a clear feedback loop. The private beta is where the founder learns what to build; the public beta is where the founder learns whether the build is being used.
What if my testers are silent after signup?
Re-contact them within seven days. The first contact after signup is the contact that turns a signup into a tester. If the second contact is also silent, the tester has been filtered out by one of the three filters above — they did not have a felt need, they were not using the product in their own workflow, or their commitment was verbal only.
Can I use Product Hunt or BetaList for the first ten?
Not for the first ten. Both platforms are useful once the landing page is converting at a known rate and the beta is stable. For the first ten, the recruitment is manual — your own network and the niche communities where the persona already congregates. The platforms are second-wave channels, not first-wave ones.
How does Yibud help with finding beta testers?
Yibud does not recruit testers for you — there is no algorithm that can. What the tools do is surface whether the persona, the problem, and the channel are aligned. The startup idea evaluator scores the eight dimensions of the idea; a weak market or distribution score usually means the persona is not yet defined sharply enough to recruit against. The customer discovery page documents the broader discipline; the decision guide covers what to do with the evidence the cohort produces.
References
Primary sources
- Rob Fitzpatrick — 'How to Find Beta Testers' — the canonical primary essay on applying the Mom Test rule to beta recruitment (momtestbook.com).
- Rob Fitzpatrick — The Mom Test (2013) — the book on talking to customers and learning if your business ideas are good (momtestbook.com).
- Paul Graham — 'Do Things That Don't Scale' (2013) — the Y Combinator origin essay on manual recruitment of the first ten users (paulgraham.com/ds.html).
- BetaList — Submission Criteria — BetaList's published rules for what can be listed on the directory (betalist.com/criteria).
- Product Hunt — Launch — Product Hunt's official launch guide, including the 'coming soon' surface mechanics (producthunt.co/launch).
Check whether the persona is in the inputs
Run the idea through the Startup MRI evaluator in under a minute. The persona in your inputs is the variable the recruitment channel has to match — and the first place the wrong cohort comes from is a persona the channel cannot reach.
Analyze my idea →