Cold Email Sequence Examples for B2B SaaS Companies

Most "cold email template" posts give you one email. That's not the thing that works. The thing that works is a sequence: four or five messages that each say something different, spaced far enough apart that you're not annoying, with a clear reason to stop. Writing the first email is easy. Writing follow-ups two through five without repeating yourself is where almost every B2B SaaS team falls apart — which is why follow-up #3 so often reads "just bumping this to the top of your inbox."

This post is worked examples, not theory. Four complete sequences, with the actual copy, the day each step fires, and a note on why each message exists. If you want the reasoning behind timing, angle rotation, and stop rules, read how to write a cold email sequence first. Then come back and steal these.

{

h2

}

The skeleton all four share

Every sequence below follows the same shape. Different market, different words, same bones.

Steps 1 and 2 share a thread. Step 3 starts a fresh thread with a new subject line, because by day eight your original subject has already been judged and ignored. Step 4 ends it. Four steps is enough for most SaaS motions; five if your deal size justifies it.

Example 1: Dev tool selling to an engineering leader

Product: CI build-caching tool. Buyer: head of platform at a 60-engineer startup. These people can smell a template from the preview pane, so every claim has to be falsifiable and the ask has to be small.

Step 1 — Day 0. Subject: cold starts on your first CI job

"Hi Priya — saw from your careers page that Nimbus runs Buildkite across roughly 40 repos. Teams at that shape usually hit 6–9 minute cold starts on the first job of the morning, because the layer cache re-downloads per agent instead of living on the host. We pinned that cache at the host level for two teams your size and got first-run times under 90 seconds. Is build time something your team is actually tracking this quarter, or is it noise right now?"

Step 2 — Day 3, same thread.

"One more data point: Parabola's p50 CI run went from 11 minutes to 4 in the first two weeks, and their monthly spend on build agents dropped about $3,100. The change was a config file, not a migration. Want me to send the config so your team can test it without talking to us?"

That offer is the whole email. Engineers reply to "here's the thing, try it yourself" far more often than to "15 minutes on Thursday?"

Step 3 — Day 8, new thread. Subject: flaky tests and the on-call rotation

"Hi Priya — different angle than my last note. The other complaint we hear from platform leads around 60 engineers is that flaky tests quietly eat the on-call rotation: reruns, Slack archaeology, someone merging on red because the suite is 'usually' fine. We fingerprint reruns so you can see which 12 tests cause 80% of the noise. Most teams find it's a dozen files. If that's a bigger problem than build time, say so and I'll send what we've got on it."

Step 4 — Day 16, same thread as step 3.

"Priya — I'll stop here. If platform work isn't getting budget before Q3, I'm happy to check back in autumn. And if developer experience sits with someone else, reply with a name and I'll leave you alone."

Example 2: Horizontal SaaS selling to a VP of Sales

Product: call-coaching software. Buyer: VP Sales at a 120-person org. The trigger here is hiring — open AE roles are public, dated, and directly connected to the problem the product solves.

Step 1 — Day 0. Subject: ramp time for the six AE roles

"Hi Marcus — you've got six AE roles open across Chicago and Austin right now. At most mid-market orgs those reps take 4–5 months to reach full quota, mostly because call coaching is whatever time a manager has left on Friday. We score every recorded call against your own top performers' calls, so a new rep learns in week two what they'd otherwise pick up in month five. Does ramp sit with you, or with enablement?"

Step 2 — Day 4, same thread.

"Concrete version: Fieldbase took average ramp from 143 days to 97 over two quarters with 40 reps. At your average deal size that's roughly one extra closed deal per new hire before they're even 'ramped.' Worth 15 minutes, or should I send the breakdown instead?"

Step 3 — Day 9, new thread. Subject: ramp benchmarks, no pitch

"Marcus — no ask in this one. We pulled ramp data from 70 mid-market sales orgs (headcount, comp structure, time to first closed-won, time to quota) into a two-page summary. The medians surprised a few people, especially on hybrid teams. Want me to send it? Reply 'send' and it's in your inbox in a minute."

Step 4 — Day 17.

"Last one from me, Marcus. If this is the wrong quarter, reply 'later' and I'll come back in October. If it's the wrong company entirely, 'no' works and you won't hear from me again."

Giving people a one-word exit is not weakness. "Later" and "no" are both better outcomes than silence, because both let you clean your pipeline honestly.

Example 3: Vertical SaaS selling to a non-technical operator

Product: field-service dispatch software. Buyer: operations manager at a 38-truck HVAC contractor. This buyer isn't reading sales emails all day, so plainer language wins and the proof has to be local.

Step 1 — Day 0. Subject: the 4:30 call

"Hi Dana — Hallmark runs about 38 trucks out of two yards, from what I can tell. At that size the last call of the day is usually the one that gets dropped, because after 4pm dispatch is working off a whiteboard and the radio. Whoever answers gets the job, not whoever's closest. We put live drive times on a map so that call goes to the right truck. Three contractors in the Carolinas added about 1.5 extra calls per tech per week. Worth 15 minutes next week?"

Step 2 — Day 4, same thread. Pre-empt the real objection.

"The question I always get next: will the techs actually use it? It's one screen on the phone — accept, navigate, done. No timesheets, no checklists. Adoption at Carolina Comfort was 31 of 34 techs in the first week."

Step 3 — Day 10, new thread. Subject: Bill at Carolina Comfort

"Dana — Bill Hartley at Carolina Comfort runs a similar two-yard setup and switched off whiteboards last spring. Happy to introduce you so you can ask him what went wrong, not just what went right. He's candid about the first month. Want the intro?"

Step 4 — Day 18.

"I'll leave it there, Dana. If dispatch is fine as is, that's a good problem to have. If you want the Bill intro later in the season, just reply to this."

Example 4: The trigger sequence (three steps, tight window)

When you're sending off a live event — funding, a new VP hire, a competitor's price increase, a job post for the role your product replaces — compress everything. Relevance decays fast.

Step 1 — Day 0. Subject: congrats on the Series A — one thought on support volume

"Hi Jen — saw the $14M round. Congratulations. The pattern after a raise like that: support volume roughly doubles in nine months while headcount lags by a quarter. The teams that come out fine are the ones that deflect tier-one tickets before hiring, not after. If you're already planning that, ignore me. If not, I can show you what three post-Series-A teams did in 15 minutes."

Step 2 — Day 3. "Quick follow-up: here's the deflection breakdown from one of those teams — 38% of tier-one tickets were four article topics. Took them a week to fix. Want it?"

Step 3 — Day 9. "Jen, closing the loop. If support planning isn't yours, point me at whoever owns it and I'll stop filling your inbox during a busy month."

What to change before you send any of these

• The observation in line one must be real. "Saw you're hiring six AEs" works because it's checkable. "I was impressed by your innovative approach" is noise. If you can't write a true, specific first line for a lead, that lead doesn't belong in the campaign.
• Keep variables to two or three per email. Company name, one fact, maybe a geography. Long chains of {{tokens}} break in ugly ways and read like a mail merge. More on this in personalizing at scale without sounding automated.
• Swap your numbers in. Do not ship "143 days to 97" unless it's yours. Made-up metrics get caught on the first call and cost you the deal you were trying to open.
• Drop a line of signature, not five. Name, company, one link. No logos, no banners, no tracking-pixel-stuffed footer.
• Add the unsubscribe line. One sentence at the bottom of each email: reply 'unsubscribe' or a real one-click header. It costs you nothing and keeps you on the right side of Gmail and Yahoo's requirements.
• Cut 20%. Every one of these drafts was longer before I trimmed it. Yours is too.

How to judge whether a sequence is working

Don't score these on opens. Open rates on cold email are inflated by security scanners and image proxies, and a sequence that "gets 62% opens" with two replies is a failing sequence. Score on replies, positive reply rate, and meetings booked — broken out by step, so you can see which message is doing the work. If step 1 produces nothing and step 3 produces everything, your opening angle is wrong and the second angle should be your opener.

Also watch where replies die. A sequence that generates "interesting, send more info" and nothing after has a step-4 problem, not a copy problem.

Running these in Sendvanta

Sendvanta sends multi-step sequences like these through your own Gmail, Microsoft 365, or SMTP mailboxes, with per-step thread control so steps 1–2 stay in one thread and step 3 starts fresh. Drafts get SpamAssassin-scored before send, engagement is classified bot-aware so scanner opens don't pollute your reply data, and deliverability health scoring will slow a campaign down automatically if mailbox, domain, or DKIM signals start sliding. One-click unsubscribe and suppression are on by default. The free plan is $0 forever with 1,000 active leads, 3,000 emails a month, and up to 10 of your own SMTP mailboxes — enough to run all four of these sequences against a real list before you pay anything. Paid plans start at $29/mo. See the full feature list.