01 / Build the business
Find your first customers without an audience.
To find your first customers without an audience, look for people already discussing or buying help with the problem you solve. Choose one place where you can reach them, check their needs, and offer a small paid result. You can start these conversations before building a following or email list.
You will leave with: One customer acquisition channel to test, a researched prospect list, a relevant invitation, and a defined pilot offer. This lesson focuses on a small service or business software offer.
Start with something you can deliver
You need a plausible customer, a problem, and a result you can deliver. A polished brand or automated sales system can wait. If you have only an idea, use the validation lesson first to learn about the problem without disguising research as a sales pitch.
Still unsure which problem matters? Run a separate research conversation before offering a solution.Interview potential customers before building.
Paul Graham's essay describes the hands-on recruitment and service founders can use at the beginning. The practical exercise below applies that principle to a small customer list.Do Things that Don't Scale.
1. Define who is a good fit right now
Write the buyer, the situation that makes the problem timely, and the result you can provide. Add one clear reason to exclude a poor fit.
A narrow first customer
Nina, our fictional founder, offers a client-onboarding setup for independent wedding photographers. She looks for photographers who handle their own bookings and repeatedly chase missing client details.
A relevant trigger could be a photographer asking their peer group for help with that exact workflow. An agency seeking a custom CRM migration falls outside Nina's current offer.
A public job title or business size is a starting clue. It does not prove a person has your problem, controls the budget, or wants contact. Confirm those details in the conversation before treating them as a qualified opportunity.
2. Choose a channel that does not require your own audience
Find one place where suitable buyers are already active. Check for a recent relevant need and a permitted way to respond before investing time there.
If you have no relevant contacts, start with requests for help, service marketplaces, or industry events. The aim is to reach a suitable person, not collect followers. A large group is a poor starting point if you cannot find buyers with the problem or participate usefully.
| Route | Evidence to look for | First useful action |
|---|---|---|
| A focused professional community | Recent questions about the exact workflow; members include your buyer; rules allow relevant replies | Answer an existing question with a specific useful step. Explain your role before offering further help. |
| A service marketplace or public project board | A current brief matches a result you can deliver; the budget and platform terms can work for you | Respond through the platform with your understanding of the scope and a clearly labeled sample. |
| An industry meetup or association event | The audience includes your buyer, with time for conversations; access costs fit your budget | Ask how attendees handle the workflow. Offer a follow-up only when someone wants to explore your service. |
For each candidate, write down one actual request, brief, or relevant event and its date. Check who participates and how contact works. If you find only sellers promoting to other sellers, keep looking. If a route requires a paid listing or event ticket, assess the cost before committing.
Nina chooses a route from evidence
In this fictional example, Nina has no following or email list. She compares a general entrepreneur group with a wedding-photography peer group. The first contains product promotions; the second has a recent question about missing client details.
The photography group permits helpful replies and clearly disclosed service offers in that thread. Nina chooses it for a small test. She prepares an example intake form using invented client details, then answers the question before asking whether the member wants to see the sample.
Her next milestone is a conversation with a photographer who has this problem now. A like on the reply would show engagement, not a qualified buyer or a sale.
A small channel test
Buyer and problem: [specific buyer] needs help with [specific task].
Place and evidence: [community, board, or event] contains [dated example of a relevant need].
Allowed first action: [useful reply, response to a brief, or conversation].
Limits: [hours and spending I can commit] before reviewing on [date].
Evidence to record: suitable people reached, conversations held, offers made, payments, and unresolved decisions.
Next decision: continue if I can reach suitable buyers; revise the route or invitation if I cannot.
3. Build a short list with a reason for every name
Create a prospect list that records fit evidence, how you can reach the person appropriately, and the next action. Start with a batch you can research individually.
Use your chosen channel to research a first batch. If you already know relevant colleagues, clients, or peers, an introduction is another route worth testing. Public directories can help identify businesses, but a listing alone does not establish need or an invitation to send a sales message.
| Prospect | Fit evidence and access | Next action |
|---|---|---|
| A photographer a designer knows | The designer mentions repeated booking administration; need to confirm with the owner | Ask the designer whether an introduction would be welcome |
| A peer-group member asking about intake forms | Specific public request for help; community allows useful replies | Answer the question, then offer a sample if they want one |
| A studio listed in a local directory | Right industry, but problem and contact route are unknown | Research further; do not count as a qualified buyer |
Ten researched names can be a manageable first batch, not a required number or a sales forecast. Keep the source and date for each. Separate identified prospects from people contacted, conversations held, offers made, and paying customers. Those stages answer different questions.
Stripe's early-sales guide covers starting with relationships, identifying prospects, and narrowing the list to suitable buyers.Your first 10 customers — Stripe Atlas.
4. Make a small, relevant invitation
Ask for an introduction or respond to an existing request with a specific reason to talk. Explain what you offer and let the recipient choose whether to continue.
A reply to an existing request for help
For the missing booking details you described, I would first separate the questions needed now from those that can wait until closer to the wedding.
Then put the required questions in one form and link to it from the booking-confirmation email, so the client has one place to complete them.
I offer paid intake-form setups for independent photographers. I made a sample with fictional details; would seeing it help you decide whether this approach fits?
Use that reply only when it answers the person's actual question and the setting permits an offer. If you have no testimonials, show a relevant sample or demonstration and label it honestly. A sample can show how you work; it does not prove customer results.
An introduction request to someone you know
Hi [name], I'm offering a small setup service for independent wedding photographers who spend time chasing client details after a booking.
You mentioned that [person] handles their own bookings. Would you be comfortable asking whether an introduction would be useful? No need to share their contact details without checking.
The offer is one intake form and a short email sequence in their existing tools. I'm looking to understand whether that would solve a current problem for them.
After an introduction is accepted, refer to the actual reason for it. Ask whether a short conversation would help, and explain that you may discuss a paid service if it fits. Do not claim familiarity, results, or customer references you do not have.
If someone declines, close the conversation. If they ask you to follow up later, record the agreed date. Avoid repeated messages to a person who has shown no interest. A larger contact list does not repair an irrelevant offer.
5. Understand the work and the buying decision
Use the conversation to learn about the current process, the consequence of the problem, and how a purchase would be decided. Present the offer only after checking fit.
- Set the agenda: understand the workflow, then decide whether discussing a solution is useful.
- Reconstruct a recent example: what happened, what they tried, and what remained difficult.
- Check priority: what makes this worth addressing now, and what happens if they leave it unchanged?
- Check the decision: who approves spending, what alternatives they are considering, and any relevant timing constraints.
- Summarize what you heard and invite corrections before showing the scope and price.
A useful call can end with “not a fit.” For a suitable buyer, finish with a specific next step: a written offer, a review with another decision maker, or a scheduled pilot. Record who owns the action and when it is due.
Y Combinator's sales advice includes researching the prospect's situation before approaching them. Treat public signals as clues to verify, not proof of a need.Sales Advice for Technical Founders.
6. Offer a pilot with a clear finish line
Write the deliverables, exclusions, price, customer inputs, timing, and acceptance criteria in one offer. Agree on the terms before starting the work.
A pilot should let the customer evaluate a useful result and let you learn whether you can deliver it. Avoid an open-ended promise to do whatever they need. If you discount, state the normal scope and the reason for the pilot price without inventing a former price.
Nina's pilot
For an illustrative $500, Nina proposes one intake form, three email templates, and a handoff session in the customer's existing tools. One revision is included. Custom integrations and ongoing administration are excluded.
The customer supplies their current questions and workflow. They agree when delivery starts, how payment works, and how they will review the completed setup. Nina promises the deliverables, not a guaranteed number of hours saved.
After delivery, compare the result with the agreed finish line and record your actual time. Ask what the customer used, what remained confusing, and whether another problem emerged. Request permission before publishing their name, quote, or results. Ask for a relevant introduction only when they are comfortable recommending the work.
Once the offer is accepted, turn it into a delivery brief, completion checks, and a customer handoff.Run a paid pilot for a service business.
7. Find the next useful change in your pipeline
Review absolute counts and notes at each stage. Locate the unresolved step, then choose one change to the audience, conversation, offer, or delivery.
A small pipeline, without a false benchmark
Suppose Nina identifies ten prospects. She finds an appropriate route to six, holds four conversations, sends two offers, and gets one paid pilot. These fictional counts describe her batch; they are not expected conversion rates.
The four untouched prospects still need research. Of the four conversations, two did not need the service now. One of the two offers remains under review. Nina records that status rather than labeling every person who has not bought a rejection.
Her next action is to deliver the accepted pilot and answer the remaining buyer's scope question. She does not increase message volume while that question remains unresolved.
| Observation | Question to investigate | A possible next action |
|---|---|---|
| You cannot identify suitable people in the channel | Are the buyers here, or mainly other sellers? | Try another source while keeping the buyer and problem specific. |
| Suitable people do not start a conversation | Was the invitation relevant, visible, and easy to answer? | Review the request and context before changing the offer or increasing volume. |
| Conversations reveal no current need | Did the selection criteria identify a real problem? | Refine the buyer or trigger using what people actually described. |
| A suitable buyer pauses after an offer | Which decision, concern, or approval remains open? | Resolve that question; do not assume every delay means the price is wrong. |
A sale that takes too much work points toward scope or pricing. These are hypotheses to investigate, not automatic diagnoses. Keep channel results separate and record time spent as well as absolute counts. One quiet week or one paid pilot cannot establish which channel will work at scale.
Your first-customer worksheet
Start with one channel test, then use one entry per prospect. Keep contact details and conversation notes private. Add dates so an unanswered message is not confused with a declined offer.
- Fit definition
- Buyer: ____. Current problem: ____. Relevant trigger: ____. Result I can deliver: ____. Poor-fit criteria: ____.
- Channel choice
- Channel and location: ____. Dated evidence of a relevant need: ____. Who participates: ____. Permitted way to respond: ____. Why this route is reachable without an audience: ____.
- Test and proof
- First useful action: ____. Sample or demonstration I can show: ____. What it proves and does not prove: ____. Time/spending limit: ____. Review date: ____.
- Prospect record
- Person/business: ____. Source and date: ____. Fit evidence: ____. Still unknown: ____. Appropriate route to contact: ____.
- Conversation notes
- Recent example: ____. Current alternative: ____. Priority: ____. Decision maker/process: ____. Timing: ____.
- Offer and status
- Scope: ____. Price: ____. Date offered: ____. Status (researching / invited / conversation / offer / paid / declined): ____.
- Next action
- Action: ____. Owner: ____. Agreed follow-up date, if any: ____. Stop or revisit condition: ____.
- Batch review
- Channel and dates: ____. Hours/spending: ____. Prospects identified: ____. Contacted: ____. Conversations held: ____. Offers made: ____. Paid customers: ____. Open decisions: ____. One change to test: ____.
Plain text. Open it in any text editor or paste it into your notes.
Your next action
Choose one channel and find one dated request or brief from a suitable buyer. Record why it fits and how you can respond usefully before making an offer.