How to Get Your First Leads for Your B2B Startup
Cold email still works for your B2B startup. Your list was the problem. Fresh data, buying signals and a short email, in that order.
On this page
Your B2B startup sends 1,200 cold emails. 49 go to people who already left. Nobody replies, so cold email must be dead.
It isn't. The list was: old data, random timing, and an email that asks too much. This guide fixes all three for your startup, in order, and each section ends with something to do today.
Give this to your agent
Paste the prompt into Cursor, Claude Code, Codex or any agent.
Run THE LEAD PLAYBOOK for my business. My website: [your website]. Read it first. Then read https://www.ekoslabs.com/skills/buying-signal-scan.md and follow it. Find 10 companies that showed a buying signal in the last 30 days (just hired, raised, or posted about the problem I solve). For each, find the one person who owns the problem and check they are still in the role. Draft a two-point email for each: the signal, then one small ask, under 50 words. Show me everything before anything is sent, and ask before using any paid API.
01Fresh data
People change jobs. Every bad address costs twice: a wasted send, and a hit to your sender reputation (inbox trust), which pushes good emails into spam.
First, two questions from our outbound checklist:
- Who exactly? "B2B SaaS, 10 to 500 employees" is a market, not a segment. Add a situation, like "just hired their first sales person". 50 of them in a room should nod at the same problem.
- How clean is the list? Check 20 rows by hand: right person, still there, working email. That count is your real list quality.
Then three rules:
- Still in the role. Same title, same company on LinkedIn. "Ex-" or a new company means off the list.
- Under a year old. No confirmation date? Treat it as stale. Old bought exports cause most bounces (returned emails).
- Verify before send. Run every address through a verifier right before the campaign, not when you built the list. Drop invalid or unknown. Send from a domain set up for cold email, or even a clean list lands in spam.
Old data is also a legal risk in some countries, like Germany. Check the rules where your buyers are.
Your 1,200 contacts might become a few hundred. That's a win.
Apollo, Clay, ZoomInfo and LinkedIn Sales Navigator all work as sources. None is fresh by default: freshness is a step you run.
Do this today: the freshness check
Do the first 20 rows by hand.
| Check | Pass if | Action if it fails |
|---|---|---|
| Still in the role | Same title, same company on LinkedIn | Remove, or find their replacement |
| Record age | Confirmed in the last 12 months | Check again before sending |
| Email verified | Verifier says "valid" | Remove. Never send to "unknown" |
| Right person | Owns the problem you solve | Swap for the person who does |
| Sending domain | Set up properly for cold email | Fix it before the first send |
- Write your segment as a market plus a situation
- Export your current outbound list
- Check the first 20 contacts against the table
- Count how many fail. That number is your bounce risk
- Verify the whole list before your next send
02Buying signals
Fresh data says the person is real. A signal says why now.
It's not an "intent score". It's a change at a company that you can explain in one sentence, and you write about it. Start with these three:
| Signal | What changed | Why it matters to them | When to write |
|---|---|---|---|
| Just hired | They posted a role which exists to solve the problem you solve | They decided the problem is worth a salary, and may be open to a faster way | Within two weeks of the post |
| Just raised | A funding round or a launch | New money and new goals usually mean hiring and new spend | After the noise of week one, within the first month |
| Just posted | They wrote about the problem in public, or asked about it | The strongest signal there is, they said it themselves | The same day if you can |
Timings are rules of thumb, not data.
Six more from our signal map: a new executive (weeks 3 to 10 in the role), a pricing change, a tech-stack switch, a job post spelling out their workflow, a comparison page, repeat engagement with your content. Pick two or three, not all nine.
The signal loop
A tool does a task. A system changes your next decision. Step five is what makes this a system.
- Signal: what changed, date, link. No date means a guess
- Diagnosis: what it means for them and what usually goes wrong next. Not your problem to solve? Drop the account
- Message: what you noticed, what happens next, one small ask (section 03)
- Result: no reply, not now, interested, meeting or wrong person, plus their exact words
- Next iteration: keep, change or drop the signal
Every Friday, 30 minutes: count replies per signal, put their best phrase into next week's first line, change one thing. Change everything at once and you learn nothing.
For step 1, our buying-signal-scan skill checks targets for funding, hiring, tech changes and news, and ranks who to contact this week, dated.
Our buying-signal-scan skill. Made in Germany.
Do this today: run one signal through the loop
- Write the problem you solve in one sentence
- Pick the signal that best shows a company has that problem now
- Find 10 companies with that signal in the last 30 days
- For each, find the problem owner and run the freshness check from 01
- Fill the first three columns before writing any email
| Account | Signal + date + link | Diagnosis | Message sent | Result | Their words | Next iteration |
|---|---|---|---|---|---|---|
Ten real, timed, verified contacts: your first batch.
03The email
Two points.
Point one: the signal. Something real and recent about them.
Point two: the ask. One small, easy question.
A middle line, if any, is the diagnosis: what usually happens next for them.
The test from our checklist: could the buyer have written your first line? If it opens with who you are and what you built, it fails.
Subject: quick one
Hi Maya,
Saw you just raised. Congrats.
Worth a 10-min chat this week?Honest, and five seconds to read. Three more by signal type. Copy the shape, not the words.
Hi [Name],
Saw you're hiring your first [role]. Usually that means [problem] is about to get real.
Worth a 10-min chat on how others set that up?Hi [Name],
Congrats on the new role at [Company]. New [title]s usually spend the first month figuring out which [numbers / tools] to trust.
What's on your list?Hi [Name],
Read your post on [topic]. The part about [specific line] is exactly what we see too.
Open to a 10-min chat on how we'd approach it?Not only email: four outreach styles that get seen
When your buyer gets fifty cold emails a week, a different format gets noticed. The signal and two-point rule still apply.
- Meme outreach. A meme of the problem your signal points at. A founder who just hired their first SDR (sales rep) gets "SDR on day one vs day thirty". Then send the two-point email.
- Loom outreach. A 60 to 90 second screen video fixing one thing on their site, ad or post. Put their name and the thumbnail in the first line. Save it for your best 10 to 20 leads.
- Public outreach. Reply usefully to their post, quote their thread, or tag them in a breakdown of something they did well. Give value first, DM after they reply.
- Meme plus attack outreach. A meme poking at the problem or the old way, then a sharp take on why it fails. Attack the problem, never the person. It can read as mean, so test on a few warm leads (people who know you) first.
| Style | Best for | Effort per lead | Use it when |
|---|---|---|---|
| Two-point email | Volume with a real signal | Low | You have a clear, recent signal |
| Meme | Founders and creative buyers | Low | Their problem is easy to show in one picture |
| Loom | Your top 10 to 20 leads | High | You can show one fix on their own site or ads |
| Public | Buyers who post a lot | Medium | They are active on LinkedIn or X this week |
| Meme plus attack | Buyers stuck on an old way | Medium | You have a strong take on why the old way fails |
Forty good ones beat 1,200 bad ones
Forty fresh, timed, short emails teach you more than 1,200 random ones: you can read every reply and fix your message by Friday. Earn volume after the message works.
After you send: who answers replies and how fast, what each follow-up adds ("just bumping this up" adds nothing), and replies per signal, not opens.
Do this today: write your first five
| # | Name | Signal (one line) | Diagnosis (one line) | Ask | Sent? |
|---|---|---|---|---|---|
| 1 | |||||
| 2 | |||||
| 3 | |||||
| 4 | |||||
| 5 |
Before you send:
- First line names a real, recent signal
- Under 50 words
- One small ask
- No links or attachments in the first email
- You'd reply to it yourself
After you send:
- You know who answers a reply, and how fast
- Your first follow-up says something new
- Replies are logged per signal in the loop table from 02
In one line
Fresh data, the right timing, a short email. Forty of those beat 1,200 of anything else.
Any startupHow to Create Your ICP for Your Startupwho should be on your list in the first place.
ekoslabs.comekoslabs.com

Want this run for you?
We build it on your tools, held for your approval.
