How to Get Your First 5 Customers in 2026

Quick answer: $10k/month is four clients at $2,500. That's it. You get those four from two motions — content (call your shot in public so people come to you) and warm outbound (mine the network you already have, then go find strangers who look like it). Everything else is plumbing you set up once. This guide is the whole system, with the real reply rates from my campaigns, the messages that worked and the ones that didn't, and the pricing floor that makes the math close. No gatekeeping, I promise.

One caveat before you start, and it's a big one.

None of this works if your ICP and your offer aren't already locked in.

This isn't how you find what you're selling. It's how you sell it once you know. If you can't finish the sentence "I help [specific person] get [specific outcome]," close this guide and go do that first.

Everything below is an amplifier. Amplify a vague offer and you just get vague results faster.

The math first

Four clients at $2,500/mo is $10k.

I keep dragging everything back to that arithmetic because it changes what you do on Monday. You're not trying to "grow." You're trying to have ~40 good conversations, because 40 conversations is 4 clients at any sane close rate.

So the job is really just: manufacture conversations.

Two things do that, and they feed each other.

→ Content — you say something specific in public and the people who need it raise their hand
→ Warm outbound — you go to the people who already know you, then to the people who look like them

Content + warm outbound. That's the whole system. Everything below is those two things in detail.

1. Call your shot

Spend 25 minutes writing a bold claim under 10 words. Attach a photo that makes you look like a person. Post it.

Mine was: "I want to help 100 founders grow their startups before summer ends."

The photo was me. Taken in my apartment. Not a headshot, not a conference stage.

That post got me 15 calls in a week. I was booking ~4 calls a day and my DMs were flooded past the point where I could actually reply to everyone.

One post. 25 minutes of writing.

Here's why it works, and it isn't the writing.

A number makes the thing countable. You can't DM a hundred people individually — but you can say one thing that a hundred people need, and the number gives the reader a seat to step into. The bold claim isn't bravado. It's an invitation with a seat count.

The hook is 80% of the post. The first line and the image do almost all the work. Everything after them is just proving you meant it. So spend the 25 minutes on the first line, not the body.

Underneath the hook I run the same 5 beats every time:

Beat 1 — The hook. Bold, first-person, has a number in it. "I want to help 100 founders…" / "I built 9 AI agents to automate my job…"

Beat 2 — Origin, then proof. Two sentences on where this came from, then the receipts as an arrow block. Not adjectives. Numbers.

Beat 3 — The compiled resource. "So I put everything I learned into one guide," plus 3-4 bullets that create curiosity about outcomes, not features.

Beat 4 — The wedge. "No newsletter wall, no catch." Say the catch isn't there, because everyone assumes it is.

Beat 5 — The comment-gate. "Comment 'growth' and I'll send it your way." One word, casual sign-off.

Beat 2 is the one people skip, so here's a real one. My best post ever was about building 9 agents to automate my own job, and the middle of it was just this:

Since late May my 9 agents have
→ processed $35k in invoices
→ onboarded 13 new clients
→ answered 264 questions concerning customer/prospect transcripts, HubSpot notes, Slack conversations, past emails and more

That post did 580 reactions and 1,209 comments. It out-commented everything I'd ever written by ~7x.

And every single one of those comments is an inbound lead.

That's what the comment-gate actually does. It turns a passive impression into a person who has explicitly asked you for something — which means your follow-up DM isn't a cold message anymore. It's a reply.

On cadence: post 3-5x a week. This is not a style preference, it's the entire mechanism.

I took over content for a LinkedIn-automation client as fractional head of content. Zero paid spend. In 31 days:

→ 163,320 impressions — 2.7x the account's entire prior history, in half the time
→ impressions/day went from ~1,000 to ~5,300 (+426%)
→ avg likes/post nearly doubled (37 → 70)
→ cadence tripled (1 post every 8.6 days → 1 every 2.8 days)

The mechanics were not clever. The account just started posting.

You don't need to be clever. You need to be specific, consistent, and willing to give the good stuff away.

1b. Put a URL in your bio

You don't need a landing page. A Calendly or Cal link is fine, and shipping it today beats shipping a site in three weeks.

Do one thing with it: add 2-3 qualifying questions before the call. Team size, funding, whatever your budget signal is.

You're about to get an inbound spike from the post above. The fastest way to waste it is to fill your week with calls that were never going to convert. The questions are the filter and they cost you nothing.

But notice what this is not.

The link goes in your bio — where people find it after they've decided they're interested. It does not go in your opening message.

Hold that thought. It turns out to be the single biggest lever in this entire guide, and I've got the numbers in section 4.

2. Mine the network you already have

Your first clients are already connected to you. You just can't see them, because LinkedIn's UI is built to stop you from doing exactly what you're about to do.

Pull your full connection list with Unipile. It's an API over your own LinkedIn account (~$5/account/month) and it's how I do all programmatic messaging from my own profile.

Pull 1st and 2nd degree. The 2nd-degree list is not a consolation prize — it's the more valuable half, and I'll explain why in a sec.

Then drop the export into Origami or Claude Code and hand it this:

Ask me who my perfect customer is and what problem I'm solving, then find all my best fits in this list who currently have that valid problem.

The "ask me first" clause matters more than it looks. It forces the agent to interrogate your ICP before it filters, so you get a qualified list instead of a keyword match.

(It also usually reveals your ICP is fuzzier than you thought. Cheap lesson to learn on a Tuesday instead of after 400 sends.)

What you want on the other side is a fit-scored table with two message columns:

→ A direct DM for 1st-degree connections. You know them. Just talk to them.
→ An intro-request for 1st-degree connections who know a 2nd-degree target. You're not asking your friend to buy. You're asking your friend to introduce you.

That second column is the entire reason to pull the 2nd-degree list.

In industries where people buy on relationships, a warm intro beats cold outreach and it isn't close. Cold outbound is not hitting a 20% reply rate. A warm intro does it routinely.

Your 1st-degree network is small. Your 1st-degree network as a routing layer into your 2nd-degree network is enormous.

2a. Message them like a person

Short. Like you're texting a friend. Not an email, not a pitch, no signature block.

Lead with something specific you actually noticed — a gap on their site, a page ranking for the wrong term, an obvious missed opportunity. Then ask if they want to jam on it.

Not "can I show you our platform." Ask if they want to think about the problem together.

And the bar for sending is simple: if you genuinely believe you can help them, sending nothing is the worse option.

People treat outreach to their own network like an imposition. It's only an imposition if you're wrong about being able to help. If you're right, staying quiet is the rude thing.

3. Message quasi-strangers

Now you take the handful of warm-network people who fit best and go find a thousand more like them.

For sourcing: go to Origami, Claude Code, Clay, or Apollo, drop in your best-fit customers, and say "find me more of these."

Origami's the one I lean on. Describe your ICP in plain English and it builds the list — find people, find companies, scrape sites, resolve handles to real contacts. Inserting rows kicks off enrichment automatically. You get a fit-scored table with name, company, title, site, phone, and LinkedIn URL.

For the filter: look for something you genuinely relate on. Same college, same city, same industry, same weird former job.

This isn't a personalization gimmick to juice a metric. It's what gives you an honest first line — and the honest first line is the thing that gets replied to.

For sending: either message manually (genuinely fine at low volume, and you'll learn more) or sequence it.

→ LinkedIn → HeyReach. Connection requests and message sequences across multiple seats, one unified inbox, so you're not logging in and out of accounts all day.
→ Email → Instantly. My top pick, and the most reliable I've used for large lists. If you only set up one sending tool, set up this one. (Lemlist and Apollo sequences are the credible alternatives. I've stayed on Instantly.)

Then wire the two ends together so it runs without you.

My core daily job is one scheduled task: every morning, pull 20 fresh leads from Origami, keep only the ones with a LinkedIn URL, push them into the HeyReach campaign.

That's a ten-line bridge between two APIs. That's the whole outbound engine.

On volume (and being honest about it)

HeyReach will pace around 40 requests/account/day, and that's the number I've used to model client campaigns.

But my practical floor is more conservative: 20-25 invites/day, warmed up over 2-3 weeks, with randomized timing.

LinkedIn's genuinely safe ceiling is closer to ~100 requests/week per account. Going from 0 to 100/day overnight looks exactly like what it is.

Caveat I'm not going to dress up: any third-party LinkedIn automation carries irreducible risk. There is no tool that makes it zero. Run more seats instead of pushing one seat hard, and watch two health metrics — keep acceptance above 25% and reply above 15%.

Run enough volume that it's arithmetically hard not to hit your number. Here's what that actually looks like on real campaigns of mine:

→ 330 connection requests out, 62.4% accepted
→ another: 278 connects → 45 accepted → 7 replies → 2-4 meetings
→ a slower one: 203 requests → 30 accepted → 3 positive replies over 8 days

That's the honest range. Positive replies are a small number multiplied against a big one, which is why the big one has to actually be big.

4. The message: what the data says

Read this part twice. It cost a client real money to learn and it's the least intuitive thing I know about outbound.

I pulled every message from a client's LinkedIn campaign — 182 prospects, 345 outbound touches, 89 replies — and tore it apart by variant.

Here's what came out.

Do not put the booking link in the cold opener

Same opener. Same audience. Split only by whether message 1 had a Calendly link in it.

Opener variant	Reply rate
With a booking link in msg 1	5.6% (4 / 71)
Without a link in msg 1	41.0% (25 / 61)

Both samples are big enough to trust.

The biggest lever was not the wording. It was the link.

And the most-used version of that message — link plus attachments, sent 60 times — was the worst performer of the whole campaign at ~3%.

The mechanism is obvious in hindsight. A link in a cold opener flips your message from "a person is talking to me" to "a person is selling to me." And the ask (book a meeting with a stranger) is enormous relative to the trust (zero).

Send the link only after a positive reply. They'll ask for it. That's the point.

Being straight about that 41%: it's the full 3-touch sequence working on the no-link group, not the opener carrying it alone. About half those replies only came back after messages 2 and 3.

Which brings me to the second finding.

Keep all three touches, and never skip the breakup

Touch	Sent	Replied	Reply rate
Msg 1 — cold opener	173	26	15.0%
Msg 2 — availability nudge	81	12	14.8%
Msg 3 — breakup	49	13	26.5%

The breakup — "last time I'll follow up on this" — beat everything.

The follow-ups are responsible for roughly half of all replies. And they're free. Anyone telling you the opener is the whole game is reading a smaller sample than this one.

The winning opener (identity-led, ~60 words, no link, no attachment):

{name},

I'm {your name}, a former {credible role at a company they respect}. We built {product} to {specific outcome for their exact job}.

Currently piloting with teams at {recognizable peers} — would you be open to a 15 minute demo this week or next?

The breakup, which pulled the most:

{name}, last time I will follow up on this — let me know if you are interested in seeing {the specific outcome}.

And one worth knowing about because it's a trap: "just realized I never attached my link" got 0 replies out of 12 sends.

The manufactured-excuse follow-up does not work. People can smell it.

So the rules:

→ Identity-led, not war-story
→ Short — the 60-word versions beat the long pitch with the dashboard screenshot and the PDF, every time
→ No link and no attachment in the opener
→ Keep all three touches

And track the only number that matters. Build a dashboard for sent → replied → positive.

Volume sent is vanity. Positive-reply rate is the number, and it tells you which of three things to go fix: the list, the channel, or the pitch.

5. Price it so four clients is actually $10k

You can run every step above perfectly and still miss, because you priced like a freelancer.

The rules I keep on a card before every sales call:

A managed system is recurring revenue. If you run it, monitor it, or iterate on it, it's a monthly retainer. The floor for managed outbound is ~$2k/mo. Charging $500 once for a 1,000-lead engine you then babysit is a poverty rate.

Price to value, not to their wallet. Never open with "what's your budget?"

Guarantee inputs, never outcomes. "X qualified contacts and Y booked calls a month" — yes. "These will convert" — never.

Sell three months, not one. Whatever they sign on day one is 10% as good as it'll be by month two. One-month pilots don't work because the thing hasn't finished becoming itself yet.

The deal shape: $2,500/mo, 3-month engagement, outbound-only to start, weekly embedded calls, and a performance hedge — if we're not hitting the agreed outreach numbers, we can cut it. No lock-in.

Four of those is your $10k.

Now the honest part

Every deal I quoted converged on $2,500/mo. Regardless of scope. Regardless of the client's ACV. Regardless of how much money they had.

A client doing $3M ARR with $30-50k contracts got the same number as a moonlighting solo founder.

And I closed essentially every price I ever quoted.

A 100% close rate is not a flex. It's the clearest possible evidence you're priced too low.

If nobody has ever pushed back on your number, it's a number you picked to be comfortable — not one you picked to be right. Raise it until someone says no. Then you'll know where the ceiling actually is.

I anchored my pricing to what my old employer charged instead of to what the outcome was worth. That single unexamined habit is the most expensive mistake in this document.

What this actually looked like

I'm not going to pretend the line went straight up.

When I was running this system it was ~$11k MRR across six clients — $3.5k, $3k, $3k, $1k, $500, and one free pilot.

That's the real shape of $10k/month. Not four identical logos. A couple of good retainers, a couple of mediocre ones, and one I should have charged for and didn't.

The two clients under $1k are the tell. They took nearly the same effort as the $3k ones.

If I were rebuilding this from zero, the one change I'd make isn't more volume, a better tool, or a cleverer message.

It's refusing to sign anyone below the $2k floor.

Four clients at $2,500 is $10k. Six clients averaging $1,833 is also $10k — and it's a much worse week.

Start this week

Step 1: Write the bold claim. Under 10 words, with a number in it. 25 minutes on the first line. Photo of you, in your actual apartment. Post it.

Step 2: Put a Cal or Calendly link in your bio with 2-3 qualifying questions behind it.

Step 3: Pull your connections with Unipile. Hand them to Origami or Claude Code with the ICP prompt. Get back a fit-scored list with a DM column and an intro-request column.

Step 4: Message the warm list like a human. Something specific you noticed, then "want to jam on it?" No link.

Step 5: Take your ten best fits and say "find me more of these." Sequence through HeyReach or Instantly. Keep all three touches.

Step 6: Quote $2,500/mo for three months. When they say yes immediately, that's your signal to charge more next time.

Content plus warm outbound. There's no third thing, and there's no version of this where the tooling saves you from not having done step one.

Go call your shot :)