What is different about lead generation at a startup?
A company with a sales team needs more pipeline. A startup with five people and twelve customers needs to find out who its customers really are. Every outbound campaign at this stage doubles as research: which segment replies, which pain they name, which price they accept without flinching.
The old advice still holds. In Do Things That Don't Scale, Paul Graham argues that founders have to recruit their early users by hand. Outbound is a modern version of that: the founder still writes to people one by one, with help to find the right people and the right moment. Our guide to your first B2B customers covers the rest of that early playbook.
Who buys from a two-year-old company?
Not everyone. Large companies with procurement departments rarely bet on a vendor that might not exist next year. Your early buyers tend to share a few traits:
- They feel the problem every week and have already tried to patch it with a spreadsheet, a script or an intern.
- They can buy without a committee: a founder, a team lead with a company card, the head of a function at a 10 to 150 person company.
- They are often startups themselves, used to young tools and happy to give feedback.
- They value direct access to the founder and influence over the roadmap more than a long feature list.
How do you turn an ICP guess into a test?
Write each ICP as a hypothesis you can prove wrong in a few weeks. The ideal customer profile template keeps your hypotheses comparable; run them one after another:
Pick one segment
Industry, size, buyer role and one disqualifier. "Seed-stage B2B SaaS, 5 to 30 people, the founder buys, no in-house data team" is testable. "SMBs" is not.
Pick one trigger
The event that makes the problem urgent for that segment: a launch, a first hire in a function, a complaint about the tool they use today.
Write one promise
One sentence on the result you deliver for that segment. If you need three sentences, the segment is too broad.
Run it for two to three weeks
Contact enough companies to see a pattern, a few hundred rather than a few thousand. Read every reply, including the rude ones.
Keep, change or kill
Real conversations and repeated pain: keep it and widen the segment. Polite interest only: change the trigger or the promise. Silence: kill the hypothesis and test the next one.
What does the first month of founder-led outbound look like?
A realistic plan for a founder with no sales help and a few hours a day to spare. It runs two hypotheses back to back and ends with a decision, not a dashboard.
Days 1 to 3: set up sending you can afford to lose
Register one or two separate domains for outreach, configure SPF, DKIM and DMARC, and keep your main domain out of it. Gmail and Yahoo require bulk senders to authenticate their mail and offer an easy unsubscribe (Google sender guidelines). New inboxes need a gradual ramp before they send at full volume, which is why this step comes first.
Days 4 to 7: write two hypotheses and collect 30 signals for each
Use the test format above. For each hypothesis, find 30 real companies with a recent signal and note the signal next to each one. If you cannot find 30 in a week, the segment is too small or the trigger too rare.
Week 2: send hypothesis A and write every email yourself
Founder to founder, 15 to 25 a day. Log each reply in one sheet with the exact words the prospect used.
Week 3: send hypothesis B and take the calls from A
Run discovery calls, not demos. Ask what they use today, what the problem costs them and what would make them change something this quarter.
Week 4: decide, then automate only what worked
Keep the hypothesis that produced real conversations. Turn its best email into a template with a slot for the signal, and only then hand the research and sending to a tool.
Which signals fit an early-stage seller?
Startups that sell to other startups have an advantage: their buyers announce themselves in public. Signals that work well at this stage:
- New companies and launches. A team that just launched on Product Hunt or appeared in a startup directory is choosing its first tools now (new companies, product launches).
- Pain posts by founders. Founders write openly on X and Reddit about what is not working. A post about a manual process you automate is close to a perfect lead (pain-point posts).
- Fresh funding. A seed or Series A round often comes with hiring and new tools; pair it with a second signal before you write (funding signals).
- Requests for recommendations. "What do early-stage teams use for X?" threads are small but warm (tool requests).
Example: a launch post becomes a first email
A fictional example. You sell usage-based billing for API products, and the hypothesis you are testing is "API startups that just launched and have no paid plan yet".
The ICP being tested
- B2B API or developer products with 2 to 20 people.
- Launched publicly in the last 30 days.
- No pricing page, "free during beta" or a single flat plan.
- The founder or CTO decides; there is no finance team yet.
The signal
A fictional startup launches a document-parsing API on Product Hunt. In the comments, the founder says pricing will be "per page, once we figure out metering".
The first line
"Congrats on the launch. You mentioned per-page pricing once metering is sorted out: that is the part we handle, from counting pages to sending the invoice. Want the setup guide we wrote for teams pricing their first API plan?"
Reach companies with a reason to buy this week
Startories finds the buying signal, verifies the decision-maker and runs the outreach until they book a call.
What does a founder-to-founder sequence sound like?
Founders answer founders who write like people. Here is a three-email sequence for the billing example above, with what each email is doing. Keep each one under about 90 words, in plain text, signed with your first name.
| When | Subject | What the email says | Why it works |
|---|---|---|---|
| Day 0 | "metering for your parsing API" | The launch congratulations and the per-page pricing line shown above. | It quotes the founder's own words from the launch comments, so it cannot be mistaken for a mass email. |
| Day 4 | "how API teams price a first plan" | "Three common ways to price an API: flat tiers, per unit with a free allowance, and prepaid credits. Here is a one-page comparison of the trade-offs, no signup needed." | It gives something useful with no ask attached and shows you understand their pricing decision. |
| Day 10 | "design partner spot" | "We are taking on four design partners this quarter: half price for a year and a shared channel with me. In return, a 20-minute call every two weeks. Interested?" | The offer is specific, the limit has a real reason (your time) and the trade is stated plainly. |
What offer gets an early customer to say yes?
A young product needs an offer that makes yes cheap to say. Pick one offer per hypothesis and state it by the second email at the latest:
| Offer | What the buyer gets | What you get | Watch out for |
|---|---|---|---|
| Design partner | A lower price, a direct line to the founder, input on the roadmap | Regular feedback and an honest view of the gaps | Partners whose requests pull you away from your ICP |
| Paid pilot | A fixed scope, a short timeline and a defined success test | Commitment, and a clear yes or no at the end | Pilots that drift on without an end date |
| Done-with-you setup | You configure the product with them on a call | Faster activation and first-hand onboarding lessons | Setup work you cannot repeat at 50 customers |
| Monthly terms with a clean export | The freedom to leave at any time | Lower perceived risk | Churn that hides a real product gap |
How much founder time does outbound take?
Founders underestimate the research, not the writing. A worked example with assumptions to replace with your own: you want 10 discovery calls in three weeks, and you assume one call per 25 well-chosen companies contacted. That is 250 companies.
By hand, finding the signal, confirming the company fits, identifying the decision-maker and verifying the email takes around 8 minutes per company once you have a routine (our estimate; time yourself on the first 20). That is about 33 hours for 250 companies, before you write a word. A personal first line adds 2 to 3 minutes each, so plan for 40 to 45 hours in total: roughly one full working week out of the three.
That is a good use of time for the first hypothesis, because you learn things no tool will tell you. It stops being one at the second or third, when the research looks the same every time and your hours belong in sales calls. Hand off the research and sending at that point, and keep the conversations. If you are weighing a first sales hire instead, AI SDR vs human SDR compares the options and their costs.
How do you answer buyers who doubt a young vendor?
Every startup hears the same doubts. Answer them head-on; buyers notice when you dodge.
| What they say | What they mean | A straight answer |
|---|---|---|
| "You are very early." | Will you still exist next year? | Offer monthly terms and a clean data export, so leaving is cheap if things go wrong. |
| "Who else uses this?" | Is anyone like us taking the risk? | Be honest about the number. Offer design-partner terms: a lower price in exchange for regular feedback calls. |
| "We need SSO and a security review." | Our IT team has a checklist. | Say what exists today, what is on the roadmap and when, and whether that is enough for this team right now. |
| "Not a priority this quarter." | The pain is real but not urgent. | Ask what would make it urgent, then watch for that event and write again when it happens. |
Which mistakes cost early-stage founders the most?
- Buying 10,000 contacts before you know which 300 matter. The list is the cheap part. The domain reputation it burns is not.
- Sending from your main domain. If outreach hurts its reputation, your product emails, invoices and investor updates suffer with it.
- Testing four segments at once. With a few hundred emails, you cannot tell which change moved the result.
- Writing in pitch-deck language. "The operating system for X" means something to investors and nothing to a buyer. Use the words prospects used in their replies.
- Automating before you have written 50 emails by hand. You will automate the wrong message, faster.
- Hiring an SDR to find product-market fit. Finding fit is the founder's job; an SDR scales a message that already works (what an SDR really costs).
What should a founder measure each week?
Keep the scoreboard small enough to review every Friday:
- Conversations started per ICP hypothesis, not emails sent.
- The exact words prospects use for the problem; they become your next subject lines and landing page copy.
- Founder hours per conversation, to decide when to hand the research and sending to a tool or a service.
- Bounce and unsubscribe rates per inbox. A burned domain is an expensive mistake at this stage (deliverability basics).
- The signal behind each new customer, so your second funnel copies the first one that worked.
Getting started without a sales team
Startories works like an AI SDR for founders who do not have one. It finds launches, pain posts and recommendation requests that match your hypothesis, checks each company against your ICP, identifies and verifies the founder or team lead, and drafts a first email about the event. You approve what goes out until you trust the funnel. Emails are sent from separate, warmed-up domains with capped daily volumes, so your main domain is never exposed and the setup week above shrinks to choosing your hypothesis.
Starter is built for one hypothesis at a time: one funnel, one ICP and one intent source for $99 a month, with your first project starting on a 3-day full-access trial for $1. See plans. If founder time is the real bottleneck, the done-for-you option has our team set it up and run it. Past your first customers with a SaaS product? Continue with lead generation for SaaS, or AI companies if AI is at the core of what you sell.
Frequently asked questions
When should a startup start doing outbound?
As soon as you can describe one buyer and one problem precisely. Outbound before product-market fit is fine when the goal is learning: keep volumes small, test one segment at a time and read every reply.
Should founders sell themselves or hire an SDR?
Founders should run the first sales conversations, because they learn the most from them. An SDR hired before the message is proven usually struggles. Tools that find leads and draft emails let a founder stay in the conversations without doing all the research.
How many companies should I contact to test an ICP?
Enough to see a pattern: a few hundred well-chosen companies over two or three weeks is usually enough to tell a promising segment from a dead one. If nobody replies at all, change the segment or the trigger before adding volume.
Is cold email still worth it for an early-stage startup?
Yes, when each email has a real reason behind it and goes out from a separate, warmed-up domain. Generic cold email to bought lists burns domains and teaches you little. Signal-based emails start conversations you can learn from.
What rules apply to startup cold email in the US?
The CAN-SPAM Act applies to B2B email too: honest sender and subject line, a physical postal address, and a working opt-out honored promptly (FTC compliance guide). Buyers abroad may fall under stricter laws, so check before you target them.
What does Startories cost for a startup?
Starter is $99 a month for one funnel and about 250 qualified, enriched leads a month, and your first project starts with a 3-day full-access trial for $1. There is no free plan.