Home / Blog / First 100 Customers
Day 15Your First 100 Customers: The Founder Sales Motion Nobody Wants to Run
There is no growth hack for the first hundred. There is a list of names and a phone.
Founders spend months building a product and then discover that the part they postponed, selling it, is the part the company actually consists of. So they reach for the thing that feels least like rejection: a launch post, an ad budget, an SEO plan, a growth playbook written by someone with an audience of 200,000 people they do not have.
None of that gets you to 100 customers. What gets you there is a list of specific companies with specific names on it, contacted one at a time, by you, until you can predict what they will say before they say it.
Before you have a repeatable sales process, you are not selling. You are researching, and the invoice is the receipt for the finding.
Paul Graham's essay Do Things That Don't Scale is the canonical description of this phase, and the examples still land because they are so unglamorous. Airbnb's founders flew to New York and went door to door photographing hosts' apartments. Stripe's founders, when someone said they would try it, asked for their laptop and installed it on the spot. Neither is a strategy you can scale. Both are how those companies found out what the product had to be.
The response time numbers come from research published in Harvard Business Review on the short life of online sales leads: contact within an hour, and you are nearly seven times more likely to have a real qualifying conversation than a firm that waits two hours, and more than sixty times more likely than one that waits twenty four. Gartner's buying research supplies the other constraint: a complex B2B purchase typically involves six to ten people, so the person who loves your product is rarely the person who can approve it.
Build the list before you build the pitch
Your first list is 100 named companies, not 10,000 email addresses. Each row needs the company, the specific person whose job your product makes easier, one observable fact that qualified them (they posted a job for the role you replace, they use a tool you integrate with, they are subject to a regulation with a deadline), and a path in: a mutual connection, a community, an event, a comment thread.
Building this list takes two days and feels like procrastination. It is the opposite. A list where every row has a reason converts several times better than a purchased one, and it doubles as your positioning test. If you cannot find 100 companies that fit your profile using public information, you do not have a targetable segment, and no amount of outreach volume fixes that.
Name the 100
Company, person, qualifying fact, way in. No row without a reason.
Reach out warm first
Introductions, then communities, then cold. Same message, different temperature.
Run the 20 minute call
Their process first, your product last. Ten minutes of questions before a single slide.
Ask for the money
A price, a start date, and a next step with a name and a calendar entry attached.
Log what happened
Objection, verbatim quote, outcome. Ten calls in, patterns appear. Change one thing.
The loop runs daily, not weekly. Its output is two things at once: revenue, and a written record of exactly what buyers say, which is the raw material for every piece of marketing you will ever write.
Do the pipeline arithmetic before you panic
Founders abandon outreach in week two because it feels like nothing is working. Usually the numbers were never going to produce a result by week two, and knowing that in advance is the difference between persistence and delusion.
Under six months to 100 customers. That is a working motion. Protect the routine and do not redesign it. Six months to a year. Viable. Raise one rate rather than all four: the reply rate responds fastest to a better list. One to two years. Too slow to learn from. Either the list is wrong or the message is generic. Fix targeting before volume. Over two years at this rate. The arithmetic says stop. A different segment, a different channel, or a different price point, not more hours.
Contacts per week times reply rate times meeting rate times close rate gives customers per week. At 50 contacts, 20 percent replying, 35 percent meeting and 25 percent closing, you win 0.875 customers a week and need 57 contacts per customer. Doubling the reply rate halves the timeline, which is why list quality outranks list size.
Notice which slider moves the answer most. Reply rate is set almost entirely by how well the list is targeted and how specifically the first line names the recipient's situation. That is a research problem, and research is cheaper than volume.
The twenty minute call that teaches you the business
Structure every early sales call the same way so that ten of them produce comparable data.
Minutes 0 to 3. Why you reached out, specifically. Name the qualifying fact. "You posted for a second claims assessor last month" beats "we help companies like yours" by a distance you can measure.
Minutes 3 to 12. Their process, in their words. How do they do this today, who touches it, how long it takes, what it costs, what happens when it goes wrong, when they last tried to fix it. No slides. This is the Mom Test discipline applied inside a sales call, and it is where the actual value of the conversation sits.
Minutes 12 to 17. Show the two parts of the product that address what they just described. Two, not twelve. A demo that covers everything communicates that you do not know what matters.
Minutes 17 to 20. Price and next step. Say a number out loud. Then stop talking. The silence after a price is the most informative three seconds in early-stage business, and founders ruin it by filling it with a discount nobody asked for.
Put one hour in your calendar every weekday morning, before anything else, labelled "outreach." In that hour you do exactly two things: contact ten new names from the list, and follow up with everyone who replied and went quiet. No product work, no email, no Slack. One hour a day is 50 contacts a week, which at the rates above is a customer a week and roughly 200 recorded conversations a year. Nothing else you can do with five hours a week comes close. The founders who fail at this are almost never the ones who tried it and got poor results. They are the ones who did it for nine days.
Speed is the advantage you own
You cannot outspend an incumbent or out-brand them. You can answer in four minutes on a Saturday, and they structurally cannot. That is why the response time research matters so much at this stage: nearly seven times more qualifying conversations for replying inside an hour, sixty times against a competitor who waits a day.
From the Harvard Business Review study of online lead response: replying inside an hour produced close to seven times the odds of a qualifying conversation versus two hours, and over sixty times versus a day. Relative bars, not absolute rates. Set a notification, answer from your phone, win deals a larger competitor never knew were open.
Unscalable work that pays for itself
In the first hundred, do the customer's job for them where you can. Import their data by hand. Configure their account on a screen share. Write the first three reports yourself and send them as if the product produced them. Deliver the outcome manually while the software catches up.
Two things happen. The customer gets value in days instead of months, which is the whole activation problem solved with labour instead of code. And you learn precisely which parts of the manual work are repetitive enough to automate, in the order that matters, which is a better product roadmap than any planning session produces.
The rule is to stop when the work stops teaching you something. Manual onboarding for customers one through twenty is research. Manual onboarding for customer sixty, when the steps have not changed in a month, is a business decision to stay small.
When to stop selling personally
Not when you are busy. When it repeats. You are ready to hand over when you can write one page containing: who to contact, what the opening line says, the four questions on the call, the three objections and their answers, the price, and the close rate you achieve. If that page does not exist, hiring a salesperson means paying someone to rediscover it, usually slower, usually at a higher salary than your own.
Founders who hire too early tend to conclude that sales is hard and salespeople are unreliable. What actually happened is that they outsourced the research phase of the company to a stranger with a quota.
The takeaway
- Build a list of 100 named companies where every row has a qualifying fact and a way in.
- Run the pipeline arithmetic first. Reply rate is the lever, and it is set by targeting, not volume.
- One hour of outreach every weekday beats any campaign you can buy at this stage.
- Reply within the hour. It is worth roughly seven times the conversations, and it costs nothing.
- Do unscalable work until it stops teaching you something. Then automate it in the order the pain appeared.
- Hire a salesperson only after you can write the whole motion on one page.
Frequently asked questions
How do startups get their first 100 customers?
Through manual founder-led work: a named list, direct outreach one at a time, warm introductions, and doing part of the job for the customer by hand. Airbnb photographed listings door to door. Stripe installed the product on prospects' laptops in the meeting. The first hundred are earned in conversations, because you are still learning what to say.
How fast should you respond to an inbound lead?
Within an hour, ideally within five minutes. Research in Harvard Business Review found replying inside an hour produced almost seven times more qualifying conversations than waiting two hours, and over sixty times more than waiting a day. It is one of the few advantages a small company holds over a large one.
How many prospects do I need to contact to get one customer?
Work backwards through your own rates. A well-targeted early-stage list often shows 20 to 30 percent replying, a third of those taking a call, and a quarter to a third closing, which is roughly 40 to 60 contacts per customer. Generic outreach can need ten times that, which is why the list matters more than the volume.
When should a founder hire a salesperson?
Once you can write the motion on one page: who to contact, the opening line, the call questions, the objections and answers, the price, and your close rate. Before that, a salesperson is being paid to run experiments you should be running yourself.
"We'll figure out sales later" is a cause of death.
Kill My Startup covers the distribution failures that look like product failures right up until the money runs out.
Buy on Amazon →Sources
- Paul Graham, "Do Things That Don't Scale," Y Combinator essays, including the Airbnb and Stripe examples.
- Oldroyd, McElheran and Elkington, "The Short Life of Online Sales Leads," Harvard Business Review, on lead response time.
- Gartner B2B buying research on buying group size in complex purchases.
- Rob Fitzpatrick, The Mom Test, on question design in customer conversations.