Recruiting is the part that quietly kills founder-led research.
The interview itself looks manageable. You can ask questions. You can read The Mom Test. You can block a few hours on the calendar and promise yourself this is the week you will finally talk to users before building the next feature.
Then Monday arrives and you open the customer list.
Who should you ask? What should the email say? Do you start with power users because they reply, churned users because they have pain, or free users because activation is weak? What if people say yes and do not show up? What if the same five friendly customers answer every time and you slowly build the product around the people who already like you?
This is why recruiting needs a system. Not a heavyweight research operation, just enough structure that you can fill the calendar with the right people without turning research into your full-time job.
This guide is for technical B2B SaaS founders with roughly fifty to a thousand users, no dedicated researcher, and a real product decision waiting. The goal is simple: know who to ask, what to send, how to reduce no-shows, and how to keep a steady cadence.
Start with the question, not the list
Do not start by asking, “Which users should I talk to?”
Start by asking, “What decision am I trying to make?”
Nielsen Norman Group’s overview of interviewing users frames interviews as a way to understand users’ attitudes, beliefs, desires, and experiences. That sounds broad, but the recruiting cannot be broad. The people you invite need to match the experience you are trying to understand.
Three founder examples:
- If you want to understand why new accounts do not invite a teammate, recruit users who signed up recently, completed the first setup step, and still have a team size of one after two weeks.
- If you want to understand why the annual plan does not convert, recruit prospects who viewed pricing, started or seriously considered annual checkout, and then chose monthly or left.
- If you want to understand why API-heavy free users do not upgrade, recruit users who hit the limit and still stayed free.
In each case, the recruiting list becomes obvious once the question is specific.
“Talk to users” produces a vague sample. “Talk to people who hit the rate limit twice last month and did not upgrade” produces a researchable one.
Recruit by behaviour, not goodwill
The easiest people to recruit are rarely the best sample.
Power users reply because they like you. Friendly customers reply because they want to help. People who praised the product in the last survey reply because the relationship is warm. They are useful in some studies, but they are a biased default.
Steve Portigal’s Interviewing Users makes the recruiting job larger than “find people who will talk”. You need people whose behaviour puts them inside the system you are studying: buyers, daily users, admins, influencers, blockers, churned accounts, trial abandoners, and the people who do the workaround outside the product.
The Mom Test principle applies here too. Do not recruit for compliments. Recruit for evidence that the person has recently lived through the problem.
Good recruiting criteria sound like product analytics:
- cancelled in the last 30 days;
- exported more than five reports last week;
- invited teammates but never activated them;
- downgraded after the first renewal;
- opened the new feature twice and never returned;
- contacted support about the same workflow more than once.
Bad recruiting criteria sound like vibes:
- likes the product;
- knows us well;
- gave a high NPS score;
- said they would be happy to help;
- seems representative.
If you recruit by goodwill, you will learn what helpful people say. If you recruit by behaviour, you will learn what happened.
Build three recruiting channels
Most founders rely on one channel: email. Email is useful, but it should not be the whole system.
You want three channels that cover different kinds of users.
1. Behaviour-triggered outreach
This is the highest-signal channel because the ask is tied to something the user just did.
Examples:
- A user abandons setup after connecting one data source.
- A customer exports the same report three times in a week.
- A free account hits a plan limit.
- A paid account’s usage drops sharply.
- A customer cancels or downgrades.
The outreach can be an email, an in-product message, or a short note from the founder. The important part is the timing: the experience is fresh, and the reason for the ask is obvious.
Teresa Torres argues in Tools of the Trade: Recruiting Customer Interview Participants that recruiting has to become a repeatable system if teams are going to talk to customers every week. Behaviour-triggered outreach is how a small team gets there without reinventing the list every Monday.
2. In-product prompts
If the user is already in the product, you do not need to drag them back from a cold inbox.
Use this carefully. A slim banner after a relevant action is better than a modal that interrupts work. The copy should be specific:
We are trying to understand how teams use exports. Would you be open to a 15-minute conversation this week?
Avoid showing the prompt to the same person repeatedly. Cap it by user and by quarter. Otherwise your most active users become your easiest sample forever.
The point is not that in-product prompts are magically better than email in every case. The point is that context matters. If the research question is tied to an action, ask near the action.
3. Relationship-based invites
In B2B, some accounts are much easier to reach through a person they already know. That might be the founder, a customer-success contact, an account manager, or the person who handled onboarding.
ProductPlan’s guidance on B2B customer interviews is useful here because it names the trade-off. Relationship channels can be warmer and more effective, but they can also skew the sample toward happy customers, strategic accounts, or people the team already knows how to reach.
Use relationship channels for access. Do not let them define the whole sample.
Write the invite like a person
Most research invites sound like marketing copy wearing a lab coat.
We are conducting customer research and would love your feedback.
That sentence is easy to ignore because it is about you. A better invite is about the participant and the behaviour that made their perspective useful.
Use this structure:
- Why them.
- What you want to understand.
- How small the ask is.
- One simple way to say yes.
Example:
Subject: Quick question about your exports
Hey {name},
I noticed you have been exporting reports a lot this month. I am trying to understand what happens after teams export, especially where the product stops helping.
Would you have 15 minutes this week or next? Here is my calendar: {link}
No worries if not. Either way, thanks for using {product}.
{your name}
That is enough.
User Interviews’ guide to recruiting by email makes the same practical points: keep the message short, make the subject specific, and avoid asking people to decode a generic research request.
Do not include three links. Do not attach a deck. Do not say “as part of our ongoing UX initiative”. Do not sell the product in the same message. You are asking for a conversation, not warming a sales lead.
Recruit continuously, not heroically
The worst version of research recruiting is a big campaign every quarter.
You send a batch of emails, chase replies, reschedule no-shows, run too many calls in one week, and then avoid the whole process for two months because it was painful.
Continuous recruiting is less dramatic and more useful.
Teresa Torres’ continuous-discovery advice is blunt: if you have to hustle to find a customer every week, you probably will not do it. The goal is to wake up with interviews already appearing on the calendar because the recruiting loop is running in the background.
For a solo founder, the lightweight version looks like this:
- Pick one behavioural trigger that matches a current product question.
- Every Monday, pull 10–20 users who match it.
- Send a short, specific invite.
- Cap yourself at one or two interviews a week.
- Rotate the trigger when the product question changes.
You do not need a perfect automation on day one. A saved SQL query, a spreadsheet, and a recurring calendar block are enough to start. The important part is that recruiting becomes a cadence rather than a campaign.
Protect the sample from convenience
Recruiting is not admin. Recruiting is part of the research.
The answer you get is filtered through who agreed to talk. If only your happiest users answer, the research will sound happier than the product is. If only churned customers answer, the research will sound more broken than the product is. If only senior buyers answer, you may miss the daily user’s workaround that actually determines retention.
Use three simple protections:
- Track invites, not just interviews. If you invited 50 trial abandoners and only 6 replied, the silent 44 are part of the story.
- Balance segments deliberately. If the question spans admins and daily users, recruit both. Do not let whichever group replies first become “the user”.
- Use referrals cautiously. Snowball sampling can be useful for hard-to-reach audiences, but as Scribbr notes, it can trap you inside one network or subgroup.
Jeanette Mellinger makes the founder risk plain in First Round Review: bias from both sides can create false signals. Recruiting is where many of those false signals start.
Incentives in B2B
Should you pay people?
Sometimes.
For consumer research, a modest incentive is often straightforward. For B2B research, it is messier. Some companies have gift policies. Some regulated customers cannot accept vouchers. Some senior participants are not motivated by money, and a gift-card subject line may make the invite feel less serious.
Use incentives to respect time, not to buy a positive answer.
Good B2B options:
- A charity donation in the participant’s name.
- A small product credit for existing customers.
- Early access to a relevant improvement, if it does not bias the answer.
- A simple founder note that makes clear their input will shape the product.
Mention the incentive after the reason for the conversation. If the invite leads with money, you may recruit people who want the incentive more than people who have lived the problem.
Plan for no-shows
Someone saying yes is not the end of recruiting. They still have to appear.
Jeff Sauro’s MeasuringU analysis of moderated-study no-show rates puts a practical planning range around 5–10%: about 5% in MeasuringU’s own moderated studies, and 8.6% in a much larger User Interviews dataset. Founder-led interviews can easily be higher because the process is usually less polished and the participants are often busy B2B users.
Use the basics:
- Send the calendar invite immediately.
- Include the video link in the invite.
- Send a reminder the day before.
- Send a short same-day reminder.
- Make rescheduling easy.
- Over-recruit slightly if you need a fixed number of completed interviews.
User Interviews’ guide to reducing no-shows recommends confirmation and reminder messages for exactly this reason. Most no-shows are not rejections. People forget, meetings move, and the interview was never as important to them as it was to you.
A four-week recruiting plan
If you have never done this before, do not design a research machine. Run the first month.
Week 1: choose the question
Write one product decision at the top of a page. Then define the behaviour that qualifies someone to answer it.
Example:
We need to understand why invited teammates do not activate. Recruit workspace owners who invited at least one teammate in the last 30 days where no teammate completed setup.
Pull 30–50 matching users if you can. If you cannot, your question may be too narrow, or the behaviour may not be tracked well enough yet.
Week 2: send the first invite
Send the invite to 10–15 people. Keep the message short. Use one scheduling link. Track who you invited, who replied, who booked, who showed, and who declined.
Do not rewrite the whole email after two non-responses. If nobody replies, check the segment and the “why you” line first.
Week 3: run calls and send the second batch
Run the calls that booked. Immediately after each one, write down the story, the quote, the surprise, and the follow-up question it creates.
Send a second batch to another 10–15 users. If the first calls suggest a sharper segment, adjust the list.
Week 4: decide whether to continue
By the end of the month, you may have 5–8 completed interviews. That is enough to see early patterns if the segment was focused. If the stories are still all over the place, recruit a narrower group. If the same story keeps repeating, decide what product action it points to and keep one recurring interview slot open.
NN/g’s sample-size guidance is useful here: interview studies need enough participants to reach saturation, and five is often not enough. But the founder version is practical. Start small, listen carefully, and keep going while new interviews are still changing the decision.
Where Maren fits
Maren does not remove the need to recruit the right people.
You still have to decide which users matter for the question. You still need to write the invite, choose the segment, and avoid building the sample from whoever is easiest to reach.
What Maren changes is the bottleneck after someone says yes.
Instead of scheduling a live call, the participant can follow a link and talk to Maren when it suits them. Maren asks warm, specific follow-ups, probes vague answers, and brings back themes from many conversations. That helps with the two parts of recruiting that usually hurt most: time slots and moderation capacity.
The recruiting is still the research. Maren just makes it more realistic to hear from enough of the right people.
The short version
Recruiting good interview participants is not about finding people who are willing to talk. It is about finding people whose recent behaviour can answer the decision in front of you.
Start with the product question. Recruit by behaviour. Use email, in-product prompts, and relationship channels together. Protect the sample from convenience. Send short human invites. Plan for no-shows. Build a weekly cadence instead of a quarterly campaign.
The interview starts when someone joins the conversation. The research starts when you decide who gets invited.