Free guide · No signup · About 12 minutes to read
The first 10 B2B customers: a research-first playbook
A one-week system for finding good prospects, writing relevant outreach, and starting real conversations — without scraping, mass email, or pretending to know someone.
If you have a product and no customers, the advice usually splits into two camps: use your network, or send a thousand cold emails. The first is not repeatable. The second trains you to ignore the only thing that matters — whether this person has a reason to reply today.
This playbook is the third path. You will build a small list of companies with observable signals, qualify them honestly, and write one message at a time. It takes about six focused hours across a week. It costs nothing, and it works without any automation.
1. Define the buyer as a situation, not a demographic
“SaaS founders” is not a buyer. Neither is “small businesses” or “agencies.” Those are categories. A buyer is a person with a current problem, in a role that owns it, at a moment when the problem has become more expensive to ignore.
Write your starting point as one sentence:
I help [role] with [specific problem] when [observable trigger].
Two illustrative examples:
- “I help operations leads at 10–50 person logistics companies keep dispatch coverage when they post a second dispatcher job.”
- “I help freelance designers turn a stalled quote into a decision when a project goes quiet after pricing.”
The test is simple: can you observe the trigger in public? A new job post, a changelog entry, a pricing-page change, a public question — these are observable. “They probably need this” is not. If you cannot observe the trigger, choose a narrower one you can.
Narrow is not permanent. It is a starting point that makes your first ten conversations specific. You can widen the definition after you have evidence.
2. Build a 25-row list from public signals
Make a spreadsheet with these columns: company, URL, signal, date, why it matters, person or role, contact route, score, status. Twenty-five rows is enough for a first week. Do not start with a purchased list.
Places where real signals tend to live:
- Job posts. A new role often means a new workload, a new process, or a team that is about to feel a gap.
- Product changelogs. A shipped feature changes how a team works and what it needs next.
- Pricing pages. A pricing change usually signals repositioning, a new segment, or a sales motion that is being rebuilt.
- Community posts. A public complaint or question is the clearest invitation there is — if the person is asking for help.
- Partner directories and integrations. A new integration often means a new workflow to support.
- Conference agendas and speaker lists. A new initiative usually gets announced before it gets staffed.
Date every signal. If you cannot say when it happened, you do not have a signal — you have trivia. “They raised money in 2023” is history. “They posted a first technical writer role last Tuesday” is a reason to write this week.
Use the contact route the company publishes: a contact form, a public business address, or a community profile. Do not guess addresses, scrape personal inboxes, or move to a private channel to get around a gate. If there is no legitimate route, drop the row.
Quality beats volume here. Twenty-five researched rows will teach you more than 2,500 scraped addresses, and they will not damage your domain or your reputation.
3. Qualify with five tests
Before a row earns a message, score it from 0 to 2 on each test:
- Recency. Is the signal from the last few months? Can you date it?
- Specificity. Can you point to the exact page, post, or listing — not just the company?
- Relevance. Does the signal connect to the problem you solve, or only to the company in general?
- Ownership. Is there a person or role that plausibly owns this problem?
- Actionability. Can you reach them through a route they publish and permit?
Eight or more: write now. Five to seven: research one more thing. Below five: drop the row. A short list of strong rows is the point; a long list of weak ones is procrastination with a spreadsheet.
Red flags that should stop a message:
- Generic praise that could be sent to anyone.
- Old funding news or an undated “recently” claim.
- A signal that has nothing to do with your offer.
- No identifiable owner of the problem.
- A guessed or scraped email address.
4. Write the first message
A relevant first message has three parts: an observation, an implication, and a question. Keep it under 120 words. Plain text. No attachments. No calendar link in the first message. No “hope this finds you well.”
- Observation. “I saw [specific signal] on [date].”
- Implication. “That usually means [specific change] — which is why [problem] becomes urgent.”
- Question. “Is [specific thing] on your plate right now, or is it still [alternative]?”
The question should be answerable in one line. You are not asking for a call yet. You are checking whether the reason you wrote is real.
Example A — a new hire
Subject: the new dispatcher role
Hi Sam — I saw the dispatcher job you posted on Tuesday. Hiring a second dispatcher usually means coverage is getting thin before the role is filled, not after.
Is keeping shifts covered the thing that is breaking right now, or is the bigger issue somewhere else in the handoff?
— [Your name], [what you do, plainly]
Example B — a pricing change
Subject: your pricing page change
Hi Priya — your pricing page moved from per-seat to a flat team plan this month. That usually means the old model stopped matching how customers were actually buying.
Are you seeing more teams ask for a trial before committing, or is the friction somewhere earlier in the buying process?
— [Your name], [what you do, plainly]
Example C — a public question
Subject: your question about follow-ups
Hi Marco — you asked in the community thread how to follow up without sounding like a reminder. The short answer: make the second message useful on its own, then close the loop explicitly.
Is the problem the wording, or deciding when to stop? I can share the three-touch structure I use if that would help.
— [Your name], [what you do, plainly]
Subject lines should be specific and lowercase: “the new dispatcher role,” “your pricing page change,” “question about follow-ups.” No hype, no fake urgency, no “quick question” that turns out to be a pitch.
Run the remove-your-product test: if the message collapses when you remove your offer, it was a pitch, not a reason. The observation should stand on its own. Then say plainly who you are and why you are writing. If you have a commercial interest, disclose it in the first message — not after the person replies.
5. Follow up without burning goodwill
Three touches, then done. The sequence is deliberately finite: it respects the person’s inbox and gives you a clean stop rule.
- Touch 1 — day 0. The first message: observation, implication, question.
- Touch 2 — three to five business days later. Add something useful: an answer to a question they asked, a short example from work you have actually done (anonymized if needed), or one new fact that changed since you wrote.
- Touch 3 — five to seven business days later. Close the loop explicitly: “It sounds like this is not a priority right now, which is completely fine. I will stop here. If [specific problem] becomes current later, you know where to find me.”
Never send “just bumping this.” If the follow-up would not be worth reading even if the person never replies, delete it. The test for touch 2 is simple: does it add something, or does it only apply pressure?
Stop rules:
- Any “no,” “not now,” or opt-out is final. Record it and stop.
- No reply after touch 3: close the row. Do not move to another channel.
- Never add someone to a newsletter or sequence because they replied once.
- Never send the same message to two people at the same company.
The polite close works because it removes the obligation to respond. Some people will answer it precisely because you are clearly going to stop.
6. Track conversations, not opens
Keep the spreadsheet simple: company, signal, date sent, reply (yes or no), what they said, next step, stop reason. If you use email, open rates are unreliable and easy to optimize for the wrong thing. Track the only numbers that survive contact with reality:
- Messages actually sent.
- Replies received.
- Conversations that continued past one exchange.
- Qualified calls or next steps.
- Customers.
Once a week, ask which signal types produced replies. Which roles? Which subject lines? Which opening observation? Do more of what produced a real conversation and drop what did not. Do not count drafts as sends, replies as interest, or interest as revenue.
7. The seven-day plan
That is about six hours. If you have less time, do five rows, two messages, and one follow-up. Consistency beats a one-off burst of volume.
8. Free tools, if they help
You do not need a tool to use this guide. A spreadsheet and a text editor are enough. If you want a guided version, these are free and require no account:
- Buyer Finder — a guided research checklist that helps you turn a broad market into a specific next buyer.
- Outreach email builder — a five-step builder that composes from details you verified yourself. It runs in your browser and does not invent facts about your prospect.
- Free templates — sample first messages and objection replies, with the research notes behind each one.
Start with ten honest rows
The first ten customers rarely come from one clever hack. They come from a small number of well-researched conversations that turn into work. Pick one trigger, build ten honest rows, send five messages, and follow up once. Then do it again next week.
If you only remember one thing: a fact is not a reason to write, and a compliment is not a signal. Find the dated change, say why it matters now, and ask a question a busy person can answer in one line.