Cold email for SaaS founders

Start with one narrow segment you can describe without adjectives, send a short plain email from a real person, and ask a question you actually want answered. Run it at a human pace on a verified list, read every reply yourself, and change the message only after a few hundred sends.

Updated October 26, 2026

Pick a segment you could name in a sentence

The first campaign fails when the audience is “companies that need better marketing”. A segment you can name in a sentence is one where you know the problem, the tool they use now, and why they would reply. Fifteen such companies are more useful than two thousand random contacts.

Write the segment as a sentence before writing the email: “agencies with three to ten people running paid ads for local clients”. If you cannot finish the sentence with a specific problem, the segment is not ready.

  • One problem, one audience, one sentence.
  • A list you could verify by hand if you had to.
  • A reason to contact them this month.

Specific openings are easier than they look; personalisation shows where the facts come from.

Write like a founder, not a funnel

A founder's email is short because founders are busy. Three sentences and one question is a complete message. No header image, no feature list, no calendar link in the first contact, because all three signal a campaign.

Say what you built in plain words, then ask whether the problem exists for them. If the answer is no, that is a real answer and it is more valuable than a click.

  • First sentence: why them, specifically.
  • Second: what you built, in one line.
  • Third: one question with a yes or no answer.
  • No attachments, no tracking pixel, no signature block longer than two lines.

The sending setup a founder actually needs

You do not need a marketing stack. You need an authenticated sending domain, one warmed mailbox, a verified list, and a daily cap low enough to answer the replies. At one message per minute a single mailbox can send a few hundred a day, which is far more than a founder can follow up on.

If you use a managed platform, check that it sends from your own domain, reports replies rather than opens, and stops a sequence when someone answers. Those three things matter more than the feature list.

  • Your own authenticated domain, not a shared sender.
  • One mailbox, paced, with a low daily cap.
  • Replies routed to an inbox you read.

That combination is what AI cold email with SMTP included provides, with published prices from $18 a month.

Reading the first hundred replies

Classify every reply into four buckets: problem exists, problem does not exist, wrong person, not now. The first bucket is your pipeline; the third tells you the list is wrong; the fourth tells you the timing. None of them require a dashboard.

Worked example: a founder sent 220 messages over three weeks to agencies. Eleven replied. Four described the same workflow problem in their own words, two asked for a call, three were the wrong contact, two said not now. The four descriptions became the homepage headline, which is worth more than the campaign's reply rate.

Keep the measurement simple; reply rate shows the formula and the window.

When to stop the first campaign

Stop when the evidence says the segment is wrong, not when the campaign feels slow. Three signals are decisive: a majority of replies are wrong-person, no reply arrives after a few hundred verified sends, or a complaint appears. Each points at a different fix.

Wrong-person replies mean the list is off, so fix the source before writing anything new. No replies at all after a few hundred sends usually means the first line does not name a real problem. A complaint means stop immediately and review the basis for contacting that person.

Pausing a campaign is cheap. Continuing one that has already produced a complaint is how a founder loses the domain they planned to use for the next year.

  • Wrong-person majority: fix the list.
  • No replies after a few hundred: rewrite the opening.
  • Any complaint: stop and review.
  • Write the next campaign before the first one ends. The replies will tell you which sentence worked, and rewriting from that evidence is faster than starting from a blank page after the results fade from memory.
  • Keep the first hundred replies in one document, unedited. Patterns appear that no dashboard shows: the same objection phrased three ways, a feature nobody mentions, a competitor named twice. That document is the input for the next month of product and marketing work, and it costs nothing to keep.

What to do with each reply type

ReplyMeaningNext step
Describes the problemFitAsk one question, then book a call
Problem does not existWrong segmentRemove that segment, not the offer
Wrong personList qualityFind the right role before resending
Not nowTimingAsk permission to follow up later

Track these by hand for the first hundred replies. A spreadsheet is enough.

What founder outreach will not do

Honest boundaries for a first campaign.

  • It will not validate a product on its own; a reply is interest, not revenue.
  • It will not scale without a sending setup you can maintain.
  • It will not fix a vague segment or a generic message.
  • It will not guarantee meetings, trials or sales.

Frequently asked questions

How many emails should a founder send per day?

As many as you can personally follow up on. A few hundred a week at a human interval is a realistic starting point.

Should I use my main domain?

Use a separate outreach domain or subdomain so complaints do not affect password resets and billing mail.

Do I need a warm-up service?

Not necessarily, but a new mailbox needs a history before volume. Real correspondence to people who reply is the strongest warm-up.

What if nobody replies?

Check three things in order: did the messages deliver, was the list verified, and did the first line name a real problem. The answer is usually the third.

Ready to run a campaign? Compare published plans or create a free account. Trials cannot send until a paid plan is active.

Lifetime deals — one payment each

All thirteen are sold at softwareprogrammer.app. Published prices are one-time; code SALE20 takes 15% off off. One-time payment · instant order · seat activated by hand.

Prices shown are the published one-time prices at softwareprogrammer.app. Applying SALE20 reduces the total by 15% off; nothing on this page is a ranking or income promise.

Other tools in this group

Each product is operated by the same team. Links below are editorial, not paid placements.

Looking for the product this page describes? Compare plans and prices.