Most replies to cold email do not come from the first message. Across the outbound programs we have looked at, somewhere between 55 and 70 percent of all positive replies arrive on a follow-up, and the second and third emails alone usually outperform the first. Yet most sequences stop after one or two nudges that say “just checking in.” This article gives you a five-email sequence with timing, the reply-rate pattern to expect from each step, and the rules that keep follow-ups from tipping into spam.
What the data says about follow-ups
Numbers vary by list quality and industry, but the shape is consistent enough to plan around. For a well-targeted B2B list with a relevant offer in 2026:
| Day sent | Typical reply rate on this step | Share of total replies | Positive reply share | |
|---|---|---|---|---|
| 1. Opener | Day 1 | 2% to 5% | 30% to 40% | Moderate |
| 2. Value add | Day 3 or 4 | 1.5% to 4% | 25% to 30% | Highest of the sequence |
| 3. Different angle | Day 8 | 1% to 3% | 15% to 20% | High |
| 4. Social proof | Day 14 | 0.5% to 2% | 8% to 12% | Moderate |
| 5. Close the loop | Day 21 | 1% to 3% | 10% to 15% | Mixed, includes polite declines |
Two patterns matter. The second email often produces the best-quality replies of the whole sequence, because the opener did the introduction and the second message gets to be useful. And the last email spikes because a clear “I will stop now” prompts people to answer, sometimes with a no, sometimes with “actually, next quarter.” Both are worth having.
Sequences longer than five or six emails show sharply diminishing returns and rising spam complaints. If nobody has replied after five well-spaced, genuinely different messages, the answer is no for now, and the contact should go into a quarterly re-engagement list instead.
The rules before the templates
- Every follow-up must add something. A new piece of information, a different framing, a relevant example. “Bumping this to the top of your inbox” adds nothing and trains recipients to ignore you.
- Reply in the same thread for emails 2 and 3, then start a new thread for 4 and 5. Threading keeps context. A fresh subject line after two weeks gets a second look from people who dismissed the first thread.
- Keep every email under 120 words. Follow-ups should be shorter than the opener, not longer.
- One question per email. Two questions get zero answers.
- Stop on any reply. Automation that sends email 3 after someone replied “not interested” is the fastest way to a domain reputation problem.
- Send Tuesday to Thursday, between 7 and 10 in the recipient’s local morning. Monday and Friday sends reply 15 to 25 percent worse in most datasets.
The five-email sequence
Placeholders are in curly braces. The example is a hypothetical company selling a lead qualification tool to marketing directors, but the structure transfers to any B2B offer.
Email 1: Opener (day 1)
Subject: {company} and unqualified demo requests
Hi {first name},
Saw {company} is hiring two SDRs, which usually means inbound volume is up and someone is sorting the good demo requests from the rest by hand.
We help marketing teams score and route inbound so sales only sees leads that fit. {Similar company} cut their MQL-to-SQL review time from four days to same-day with it.
Is qualification something you are working on this quarter?
{signature}
Why it works: a specific, verifiable trigger, one concrete result with a named peer, and a low-commitment question that can be answered with a sentence.
Email 2: Value add (day 3 or 4, same thread)
Subject: Re: {company} and unqualified demo requests
{first name}, one thing that might be useful regardless of whether we talk.
We benchmarked cost per qualified lead across about 300 B2B companies this year. For {industry}, the median is {figure}, and the teams under that number almost all had one thing in common: a written definition of “qualified” that sales and marketing both signed. Here is the short version: {link}.
Would it help to see where {company} lands?
Why it works: genuinely useful even if they never reply, and it reframes the conversation from your product to their metric.
Email 3: Different angle (day 8, same thread)
Subject: Re: {company} and unqualified demo requests
Different angle, {first name}.
The usual objection we hear is “our CRM already scores leads.” It does, but the score is rarely tied to what sales actually accepts, so it drifts. The fix is closing that loop weekly, and it is more of a process than a tool.
If you already have that loop running, I will get out of your inbox. If not, happy to share the 20-minute weekly agenda we use.
Which is it?
Why it works: it names the objection before they raise it, offers a way out, and makes the question binary.
Email 4: Social proof (day 14, new thread)
Subject: How {similar company} handled Q3 inbound
Hi {first name},
{Similar company}, about your size and also selling to {their market}, had 400 inbound leads a quarter and sales working 40 of them. After they defined qualification and automated routing, acceptance went from 18 percent to 31 percent in one quarter with no change to the marketing spend.
The head of marketing there is happy to compare notes if that would be useful. Want an intro?
Why it works: a fresh subject line, a real number, and an offer of a peer conversation rather than a sales call.
Email 5: Close the loop (day 21, same thread as email 4)
Subject: Re: How {similar company} handled Q3 inbound
{first name}, I will stop here rather than keep filling your inbox.
If lead qualification comes up later this year, the benchmark and the weekly agenda are yours whenever you want them. And if there is someone else at {company} who owns this, I would appreciate a pointer.
Either way, thanks for reading.
Why it works: respectful, explicit, and it asks for a referral, which is the single most under-used follow-up ask.
Measuring the sequence
Track four numbers per step, not just for the whole sequence:
- Reply rate per step. If email 2 is not your best or second-best performer, the value add is not valuable enough.
- Positive reply rate per step. A reply is not a lead. Separate “interested” from “unsubscribe” and “wrong person.”
- Meeting booked rate per sequence. The number that matters. For a good list, 1 to 3 percent of contacts booking is a healthy outbound program.
- Spam complaint rate. Keep it under 0.1 percent. Above 0.3 percent your domain is at risk. Slow down, shorten the sequence, tighten the list.
Feed booked meetings back to the same qualification process you use for inbound so outbound and inbound are compared on the same cost per qualified lead basis. The method is in our CPQL benchmarks, and the definition of qualified should be the one in your MQL to SQL handoff.
Three ways sequences fail
- Same email five times. Changing “checking in” to “following up” is not a new angle. Each step needs a reason to exist.
- Too fast. Five emails in seven days reads as desperation. Spread the sequence over three weeks.
- Personalization theatre. “I loved your recent post” with no reference to what it said is worse than no personalization. One real, specific observation in email 1 is enough; the follow-ups can be templated.
For the opener itself, including subject line tests and templates by industry, see our cold email templates. The sequence above assumes the first email is already doing its job. The follow-ups are where the majority of your replies live, and they deserve at least as much attention.