Finding your first 10 customers doesn't come down to some magic channel. It comes down to your business model, and therefore your go-to-market strategy. A B2B SaaS prospects 100 accounts through LinkedIn outbound. A marketplace seeds one side by hand. A B2C product embeds itself in an existing niche community. A vertical SaaS locks down a specific profession, city by city. Every business model demands its own go-to-market (GTM). This guide gives you the full method, segmented across six business models, with examples of French startups (Doctolib, Qonto, Vinted, Lydia, Ledger) and the universal framework to apply before you start prospecting. If you're looking for "12 tips to find customers", stop here. This article assumes you're the founder of an early-stage startup and you want an operational answer, not a checklist.
Why 10 first customers (and not 1, or 100)?
The number 10 isn't arbitrary. It comes from a Y Combinator rule of thumb: below 10 customers who actually pay, you don't have a startup. You have a hunch.
A single customer can be a fluke, a friend, a cofounder's cousin. Five customers might share a hidden trait that you wrongly project onto the whole market. At 10 customers, the patterns start to emerge. You see which persona converts, at what price, on which channel, with which recurring objection. You've got the first statistically usable signals of an eventual product-market fit.
But 10 customers isn't a scale threshold either. Nobody raises a seed round on 10 logos. The goal of your first 10 isn't revenue. It's information. 10 customers = 10 useful conversations, 10 feedback cycles, 10 chances to discover that your product doesn't do what you thought it did.
The myth of the product that sells itself died with the 2010s. In 2026, in any competitive market, your first 10 customers are won through labor, not code.
The universal framework: the 4 go-to-market questions before you prospect
An early-stage go-to-market comes down to four questions. Before you touch your first prospect, you need to answer them. Not three. Not five. All four, in this order.
1. Who is your ICP (Ideal Customer Profile)?
Not "French SMBs". Not "early-stage startups". Not "CMOs at tech companies". Those answers are admissions of vagueness.
A usable ICP has an industry, a size, a role, a geography, and a buying signal. Example: "Head of Sales at a French B2B SaaS scale-up of 20 to 100 people, that raised a Series A in the last 18 months, based in Paris or Lyon". With that level of precision, you can pull a list of 80 accounts in an afternoon on Sales Navigator. Without it, you'll never be able to prospect.
2. What is your wedge?
The wedge is the sharpest problem your ICP has, not the broadest. Slack didn't go after "enterprise communication". Slack went after "IRC is ugly, and our devs are sick of it". Everything else came later.
3. On which channel do you find your ICP complaining?
The right channel isn't the one you prefer. It's the one where your ICP already hangs out, ideally already searching for a solution to your wedge. Specific subreddits, Discord threads, industry WhatsApp groups, trade shows, community Slacks: if you don't know where your ICP hangs out, you don't have an ICP, you have a fantasy.
4. What is your sales motion?
Four families: sales-led (a human closes the sale), product-led (the product closes the sale after a free trial), community-led (the community brings the customers), marketing-led (content and SEO bring the customers). The choice isn't ideological. It's dictated by your ACV (Annual Contract Value): under €100/month, sales-led is too expensive. Above €10,000/year, product-led alone isn't enough.
Until these four questions have precise answers, don't launch any prospecting. You'll burn social capital and time for nothing.
To frame your target before prospecting, the lean canvas is still the fastest tool for aligning ICP, wedge, and channels on a single page.
Do things that don't scale (the founding principle)
In 2013, Paul Graham (Y Combinator cofounder) wrote the essay Do Things That Don't Scale. It became the reference text on a startup's first customers. His thesis fits in one sentence: the thing that works to get from 0 to 10 customers won't work to get from 1,000 to 10,000. And that's perfectly fine.
Four canonical examples:
Airbnb (2009). Brian Chesky and Joe Gebbia realized their New York listings weren't converting. They got on a plane, went to New York, and photographed hosts' apartments themselves. The platform's revenue doubled that week. They didn't yet know it would become a product ("Pro Photography"), or that it would scale. They did it because it was the thing to do that day.
Stripe (2010). Patrick and John Collison signed their first customers one by one. When a dev said "OK, I'll try it", the founders pulled out their laptop and installed the integration right there, on the spot. No docs, no self-service dashboard. One manual integration per cofounder.
Doctolib (2013). Stanislas Niox-Chateau and his team went door to door among Parisian doctors. No webinar. No cold email. Handing out flyers office by office, with a demo on a laptop, in the waiting room. The first 100 practices were signed by hand.
Vinted (2008). Justas Janauskas (cofounder) manually curated the first clothing items and brought in the first sellers through Lithuanian forums. No algorithm. No growth automation. A founder answering every single message.
The pattern is always the same: founders do what no employee would do. It's the exact opposite of what most founders want to do at the start (build a system, launch an ad, automate). And that's exactly why it works.
Go-to-market by business model: 6 playbooks for your first 10 customers
Here's the heart of the article. The channel that works for a B2B SaaS doesn't work for a marketplace. The pricing that works for a B2C doesn't work for a vertical SaaS. Here are the six go-to-market playbooks by business model, each with its ICP, its #1 channel, its motion, and its pricing rule.
B2B SaaS: founder-led sales + targeted outbound
You sell a software tool to companies. Typical ACV between €100 and €10,000/month.
- ICP: a list of 50 to 100 hyper-targeted accounts. Not 10,000.
- Channel #1: warm intros through your network. Channel #2: personalized LinkedIn outbound (not templating). Channel #3: cold email, only once the first two are saturated.
- Motion: sales-led, founder-led. No SDR, no salesperson. The founder takes every demo. Until you've done 50 demos yourself, don't hire.
- Pricing: paid from the first euro. Founder discount fine (up to 50%), never free. Locked-in price: early customers keep their rate for life.
FR case: Qonto. Before their first sales hire, Alexandre Prot and Steve Anavi did more than 200 founder-led demos to sign their first SMB customers. They answered support in person. They folded feedback in within the week.
B2C / Consumer: community-led + a single organic channel
You sell a product or service to individuals. Low price, high volume, hard conversion.
- ICP: an existing niche community, not "French people aged 25-45". Think Reddit, Discord, Telegram, or offline: associations, clubs, student communities.
- Channel #1: being in the community before the product. Not "launching an Instagram campaign". Living the topic for 6 months before pitching anything.
- Motion: community-led or product-led freemium. No human sales.
- Pricing: freemium or low price from the start. No hidden discount. If it's free, it's free; if it's €9.99, it's €9.99.
FR case: Lydia started with HEC students in 2013 with a hyper-niche product (splitting a coffee among friends) before broadening out. Backmarket positioned itself in geeky refurbished-tech forums before going after the mainstream. The rule is always the same: one segment, in depth, before broadening.
Marketplace: seed one side by hand
You're building a two-sided platform (sellers/buyers, providers/clients, hosts/travelers).
- ICP: identify both sides, but don't go after them at the same time. The side to seed first is the one with the least loyalty to the incumbent (often supply).
- Channel #1: manual seeding. Recruit 10 to 30 suppliers by talking to each one, one by one.
- Motion: three possible patterns. (1) Single-player mode: the seeded side gets value without the other side (Airbnb worked as a listings site before it had travelers). (2) Concierge: you do the service by hand until you have inventory (Lyft paid its first drivers). (3) Geographic constraint: a single neighborhood, a single city, until you reach liquidity.
- Pricing: often zero fees at launch on the seeded side to reduce sign-up friction. On the demand side, market price.
FR case: Vinted seeded the sellers first, manually, in Lithuanian fashion forums and communities. No ads. No app store. Painstaking work for two years before reaching critical liquidity.
Vertical SaaS: lock down a profession, city by city
You sell a software tool to a specific profession (physiotherapists, notaries, mechanics, restaurateurs).
- ICP: one profession × one organization size × one geography. Example: "self-employed physiotherapists, 1-3 practitioners, inner Paris". Not "healthcare professionals".
- Channel #1: professional associations, trade unions, specialized trade shows, and door-to-door. Digital doesn't work until the profession knows you.
- Motion: sales-led, annual contract, in-person demo or a long video call (45 minutes minimum).
- Pricing: annual, billed upfront, 12-month contract, no cancelable monthly plan. You want cash and commitment from day 1.
FR case: Doctolib locked down self-employed Parisian doctors first, door to door. Once Paris was saturated, Lyon. Then Lille. Then Germany. The logic was clear: densify one segment before broadening, to benefit from a referral effect (doctors talk to their peers).
Hardware / physical product: pre-orders + scarcity
You sell a physical object (gadget, electronics, design, consumer goods).
- ICP: tech early adopters or collectors. Not the mainstream.
- Channel #1: crowdfunding platforms (Kickstarter, Indiegogo, Ulule) to validate demand before producing. If you can't pre-sell €10,000, don't start production.
- Motion: marketing-led + community-led. A structured pre-launch campaign, early-bird tiers, a newsletter warming the audience up 3 months ahead.
- Pricing: early bird (-30%), limited founding edition, visible scarcity. No permanent discount, no "20% off for the newsletter".
FR case: Ledger started with crypto early adopters via Bitcoin forums, before Amazon. Withings negotiated distribution in the Apple Store as early as 2009 across 4 products. The pattern: an early distribution channel with strong credibility among the target segment.
Services / agency / done-for-you: one public case study, worked hard
You sell your expertise (consulting, agency, custom services, advanced freelancing).
- ICP: companies that have already tried an alternative solution and failed. Your best customers are the ones who've already burned a budget elsewhere.
- Channel #1: the founder's personal LinkedIn (founder-led content). You post case studies, post-mortems, free tutorials. You don't sell, you demonstrate.
- Motion: sales-led, but inbound. Leads come to you because they've read your posts for 6 months.
- Pricing: high, custom. Never hourly (you're paying for expertise, not time). Per-project fixed fee or monthly retainer.
FR case: most of the AI and growth agencies scaling in 2026 start from the same playbook: a founder who posts 5 times a week on LinkedIn, detailed case studies, and zero ads. Founder-led content is the most durable acquisition mechanism for this business model.
Pricing your first 10 customers
The question comes up in every session with a founder: "Should I charge my first 10 customers, or let them test for free in exchange for feedback?".
Answer: charge them. Always. Even a little, but charge.
Three reasons:
1. A free customer isn't a customer. It's a user. A user doesn't give the same feedback as a customer. A user tolerates bugs, drops off without a word, doesn't invest in the relationship. A customer pays, so they're committed, so they talk.
2. Price is a signal. If you charge €0, your prospect understands your product is worth €0. That's hard to fix afterward. If you charge €99/month with a founder discount to €49/month, you anchor the perceived value.
3. The founder discount is fine. Free isn't. A 30 to 50% reduction for early customers is acceptable and even advisable. One condition: the rate is locked-in, meaning your first 10 keep it for life. It's a gift for their commitment, not a gift born of doubt.
The classic mistake: promising free access in exchange for feedback. Three months later, you've got 10 users who no longer answer your messages, zero revenue, and no credibility left when you try to charge them. To go deeper on early-stage pricing, read our guide on early-stage startup pricing.
The 5 mistakes that kill your first 10 customers
Mistake 1: wanting a scalable channel too early. You launch Google Ads before signing 5 customers by hand. You burn €3,000 in two weeks, you get maybe 1 lead, and you've learned nothing because you never talked to people directly. The rule: no scalable channel before the first PMF signal.
Mistake 2: three ICPs in parallel. You test "industrial SMBs", "SaaS scale-ups", and "digital agencies" at the same time. Three personas × three channels = you dilute everything. None convert. Pick a single ICP for your first 10 customers. You'll broaden later.
Mistake 3: no qualification before the demo. You take 30 demos with anyone, 25 of whom aren't in your ICP at all. You burn your time. Set a minimum filter: company size, role, and a buying signal (for example: "have you already tried a solution like X?"). No qualification, no demo.
Mistake 4: hiring a salesperson before the founder has done 50 demos. The salesperson you hire won't know how to sell your product until you've found the closing arguments yourself. Selling a product that's still finding PMF is a founder's job, not a salesperson's.
Mistake 5: confusing interest with a purchase. "Three super-excited prospects told me they'd 100% sign" is a sentence that comes up every week in coaching. Three months later, none have signed. LOIs (letters of intent) aren't contracts. Until the money is in the account, the customer isn't a customer.
When do you move from 10 to 100 customers?
You've got your 10. Well done. Next question: when do you stop founder-led selling and switch to more scalable channels?
Three signals to watch:
Signal 1: retention. Your first 10 customers renew (or buy again). If you have 30% churn in the first month, you don't have PMF. Stay founder-led, talk to every churner, understand why. For the right metrics to track, see the early-stage startup KPI guide.
Signal 2: organic pull. Prospects you never approached write to you. They heard about you through one of your customers, a post, a conference. You're no longer the only one pushing: the market is pulling.
Signal 3: your playbook is documented. You know what closes and what doesn't. You have a standard demo, an email template that works, a repeatable objection-killer. As long as you improvise at every demo, scaling is impossible.
When those three signals are in place, you can invest in scale (paid, content, hires). Before that, you'll burn capital without accelerating. It's also the moment your early-stage go-to-market stops being manual: it becomes a system. To structure the transition, the startup growth marketing guide covers the channels that take over.
The most expensive mistake is scaling before you understand why it's working. The startup that has 50 customers without understanding its funnel will get stuck at 100, unable to unlock growth, because it automated too early.
FAQ
How long does it take to find 10 customers?
Between 2 and 9 months depending on your business model. A founder-led B2B SaaS can reach 10 logos in 3-6 months if the ICP and wedge are precise. A marketplace usually takes longer, because you have to seed one side before liquidity self-sustains. If you go past 12 months without 10 paying customers, the problem is rarely the channel: it's the product, the wedge, or the ICP.
Do you need a website to find your first customers?
No. A one-page landing page is enough, or even a public Notion or a Typeform. Your first 10 customers don't decide by reading your website, they decide by talking to you. A full website is useful once you want to generate SEO or inbound, so more like between 10 and 100 customers. Before that, it's procrastination in disguise.
Cold email or LinkedIn: which should you choose?
For your first 10 B2B customers, LinkedIn outbound beats cold email 80% of the time, because your prospects see your profile, your activity, your credibility. Cold email needs a volume of 500-1,000 sends to generate 5 qualified leads. LinkedIn lets you send 30 highly personalized messages and get a 30% reply rate. Cold email has its place, but after LinkedIn is saturated. On tooling, Waalaxy automates LinkedIn prospecting where Lemlist runs cold email at volume; our B2B prospecting tools guide breaks down the full stack.
How much should you spend on ads for your first 10 customers?
Zero. Or very little (under €500/month) to test a channel, not to acquire your first customers. Paid ads are an amplifier: they amplify what's already working. If nothing works yet, they amplify nothing. Your first 10 come from founder-led selling, not Google Ads.
Do you need an MVP before prospecting?
Not necessarily. You can pre-sell a solution before it exists (concierge MVP). Stripe sold its integration by installing it manually at the first customer. Plenty of B2B SaaS companies signed paid LOIs on slides. The criterion isn't "do I have a product", it's "can I deliver on my promise to a paying customer".
What is a go-to-market for an early-stage startup?
The go-to-market (GTM) is the strategy that defines how your startup sells, to whom, at what price, and on which channel. For an early-stage startup, it's not a 30-page theoretical plan: it's the combination of a precise ICP, an identified wedge, a #1 channel tested manually, and a sales motion coherent with your ACV. A good early-stage go-to-market fits on one page and is dictated by your business model. That's exactly what this guide breaks down into six concrete GTM playbooks.
Does swanbase help you find your first customers?
It's a big part of our work. The founders we support get weekly reviews of their ICP, their prospecting scripts, their channels, and their pricing. What makes swanbase different: we don't give theoretical lectures, we look at your prospect list and fix it. Our program takes 2% equity and supports you over the long term, not in a 14-week sprint.
