TL;DR: Platform benchmarks recommend four to seven follow-ups because averages across all customers show 42% to 44% of replies coming from follow-up steps. Practitioner Eric Nowoslawski's analysis of 1,517 campaigns found the opposite: two steps is optimal, three is acceptable, and anything beyond three causes a sharp drop in replies. The disagreement comes down to measurement: benchmarks count all replies including auto-responders and rejections, while a practitioner tracking his own money counts positive responses only. The practical answer is to default to two emails, test before adding more, and judge results on positive replies rather than raw reply rate.

Key Takeaways

• Platform benchmark reports recommend four to seven follow-up emails because aggregate data shows a large share of replies come from later steps in a sequence.
• Eric Nowoslawski's analysis of 1,517 campaigns found the optimal number of follow-up steps is two, with three acceptable and a sharp drop in replies after that.
• The disagreement between benchmarks and practitioners largely comes down to measurement: reply rate counts every response including auto-replies and rejections, while positive reply rate counts only genuine interest.
• The optimal email length in Nowoslawski's dataset was 70 to 90 words, and custom-variable personalization built into the template beat generic AI-written openers.
• A default cold email sequence should start at two emails, with additional steps added only if a clean test shows they increase positive replies rather than just raw replies.
• Deliverability and a strong first email matter more to results than sequence length, and adding follow-up volume on a weak sender reputation increases the risk of landing in spam.

What Do Practitioners Say About Cold Email Follow-Ups in 2026?

Platform benchmark reports say a cold email sequence should run four to seven touches. One of the most visible practitioners in cold email, Eric Nowoslawski of Growth Engine X, published campaign data from 1,517 campaigns in 2025 showing the opposite: the optimal number of steps is two, three is acceptable, and anything past three causes a sharp drop in replies. Both claims come from real data. They disagree because they measure different things, and the gap between them is the most useful thing a founder can understand about follow-ups right now.

The short version: benchmark reports describe averages across every customer on a platform. A practitioner's number describes a decision he makes with his own money on the line.

Why do platform benchmarks recommend longer sequences than practitioners do?

Because the aggregate data genuinely supports adding steps, and because averages hide how badly the median campaign is targeted.

The numbers from the platforms are consistent with each other. According to Instantly's 2026 Cold Email Benchmark Report, 58% of all replies come from the first email in a sequence, which leaves 42% coming from follow-ups. That is a large share of your pipeline sitting behind steps two and beyond. Instantly's own recommendation, based on billions of analyzed emails, is a minimum of four total touches.

Other platforms report the same direction of travel:

• In an analysis of 53 million cold emails, Saleshandy found that 44% of all positive replies came from follow-up emails rather than the initial outreach.
• According to Woodpecker's 2026 cold email statistics, campaigns with three to five follow-up steps hit 8.3% reply rates versus 4.1% for campaigns with no follow-ups, more than double.
• Salesforge, citing lemlist's analysis of millions of campaigns, reports that a single email typically gets around a 4.5% reply rate while a 10-email sequence can push that as high as 22.37%.

None of that is fabricated. But read what it actually describes. These are platform-wide averages across every account on the platform, including campaigns sent to scraped lists with no real targeting, campaigns with an offer that only makes sense to the person who wrote it, and campaigns where the first email was so weak that of course the follow-ups had to carry the load. When the first email is bad, more steps look like they help.

There is also a structural point worth naming plainly. The companies publishing "send four to seven emails" also sell sending volume. That is not an accusation of dishonesty. Their data is real. It is simply a reason to weight an independent practitioner's number higher when the two disagree.

Why does one practitioner's data show reply drop-off after three follow-ups?

Because he is counting positive responses, not replies, and because he stops sending when sending stops working.

In a LinkedIn post, Eric Nowoslawski wrote: "In 2025, we launched 1,517 campaigns and generated over 60k positive responses." On sequence length, his conclusion was that the optimal number of steps is two. In his words: "3 is ok. Anything over 3, HUGE reply drop off."

He is not saying three steps is a mistake. He explicitly says three is fine. He is saying the cliff sits right after it.

Two other findings from the same post are worth carrying with you:

• 70 to 90 words was the optimal email length in that dataset. Short, not clever-short.
• Custom-variable personalization written to fit the template beat generic AI lines. The AI-generated opener that reads "Hey [Name], I saw you help people send cold email" performed poorly. Personalization that was built as a variable slotting into a sentence that was designed around it performed better.

The measurement difference matters more than most founders realize. Reply rate and positive reply rate are not the same thing. Every auto-responder, every out-of-office, every "please remove me", and every "not interested" counts as a reply in nearly every sending tool. Add three more steps to a sequence and your raw reply rate will probably go up. Whether your booked meetings go up is a completely separate question.

That is why I read one operator's positive-response data differently than a platform average. He sees which replies were actually good, he pays for the sending infrastructure, and he carries the reputation damage if the extra steps annoy people. His number is a decision. A benchmark report is an average.

What goes wrong when founders follow up too little or too much?

Both failure modes are common, and they fail for different reasons. Too few touches leaves the 42% of replies that come from follow-ups on the table. Too many touches does not just plateau, it actively costs you.

Following up too little. A large share of senders never attempt a follow-up at all, or quit after one. Given the Instantly and Saleshandy figures above, that is walking away from the single largest source of additional replies in a sequence. One email is not a campaign. It is a test.

Following up too much. Pushing past the point where extra steps earn positive replies increases spam complaints and unsubscribes. In 2026 that is not a soft cost. Gmail and Outlook enforcement is tighter than it was two years ago, and complaint rate feeds directly back into whether your next campaign reaches the inbox at all. Reply rates have been compressing industry-wide as inbox saturation and a flood of low-effort AI-generated outreach make any single message harder to land.

Following up badly. Semir, CEO and co-founder of the B2B account intelligence platform Salesmotion, points to generic "just checking in" follow-ups as a primary reason outreach fails. A follow-up that restates the first email in different words gives the recipient nothing new to react to. They already decided about the first message. Repeating it asks them to make the same decision again.

If your reply rate fell off a cliff recently, sequence length is rarely the first thing to check. I wrote a diagnosis order for that in why a cold email reply rate can suddenly drop, and deliverability sits above messaging in that order for a reason.

How should a founder actually structure a follow-up sequence in 2026?

Start short, make the second email genuinely different, and only add steps when a test proves the step earns positive replies.

How many steps should you run for your list size?

Nowoslawski's own answer to that is not a fixed number, it is a ratio to how many companies you can actually reach. In his video Cold Email Follow-Up Strategy for 2026 (Data-Backed) he puts it plainly: "How do you decide if you should be sending two emails in a sequence, one email in a sequence, three emails in a sequence? I think about it as it just depends on how large your TAM is."

His tiers:

• A million contacts in your total addressable market: "I would never send a follow-up email. I would just send one email, hit everybody in your TAM, and then come back with new insights, and then hit everybody again."
• 50,000 to 100,000 contacts: "send two emails in a sequence."
• Under 50,000 contacts: "send three emails in a sequence."

The logic is that a follow-up and a re-send are the same thing viewed from different distances. With a huge market, the second at-bat comes from mailing everyone again in three months with a new angle, which reaches far more people than a follow-up to the slice you already touched. With a small market, you have nobody left to add, so the extra step inside the sequence is the only at-bat you get.

That is why my own default sits where it does. The lists I build are narrow on purpose, so most campaigns run two emails, and a third only shows up on a genuinely small market where there is no wider audience to go back to.

How do I run follow-ups by default?

New cold campaigns I build default to two emails: the first email and one follow-up.

• The first email carries most of the work. That is where the targeting, the offer framing and the personalization effort go. If the first email needs five follow-ups to get noticed, the first email is the problem.
• The follow-up is a different angle on the same problem. Never "just bumping this", never "circling back". A different way into the same pain, or a different consequence of not solving it.
• A third step needs a reason: a small market, or a test. On a small addressable market, where there is no wider audience to re-mail, a third email is the only extra at-bat available. Otherwise it has to earn its place: if a clean test on the same list shows the step adds positive replies, it stays. If it only adds raw replies, it goes.
• Campaigns get judged on positive replies per contacts reached. Not raw reply rate. Auto-replies and "not interested" are replies, and counting them flatters a campaign that is not working.
• I do not track opens. The tracking pixel is a spam signal, and image prefetching inflates the number until it means nothing. If you want to know whether a campaign is working, count positive replies and booked calls.
• No follow-ups go out on Saturdays or Sundays.

In my experience, the biggest gains in cold email almost never come from adding another email. They come from a better list and a first email that sounds like it was written by a person who understands the recipient's business. Founders reach for sequence length because it is the only lever that can be pulled without thinking hard, and that is exactly why it is usually the wrong one.

If you do choose to run longer, use a fixed cadence rather than gut feel. The Digital Bloom's 2025 cold outbound analysis found that a Day 0, Day 3, Day 10, Day 17 cadence captures 93% of total replies by day 10. Read that number carefully: it is an argument for front-loading your effort, not for extending the tail.

Before you judge any of this against your own campaign, make sure you know what a realistic target looks like. I broke that down in what counts as a good cold email reply rate.

Why do benchmark reports disagree so much on what a good reply rate is?

Because different vendors divide by different denominators, and almost nobody checks which one a report used.

Some platforms calculate reply rate against emails opened. Others calculate it against total emails sent. The same campaign can look strong in one report and weak in another with no change to the campaign at all.

The spread is dramatic. Belkins' 2025 Cold Email Benchmark Report measured reply rate strictly as replies divided by total emails sent and found an average of 0.45% across all 2025 campaigns, which they describe as reflecting strict cold outreach to net-new contacts. Instantly's 2026 report puts the overall average at 3.43%, with top performers exceeding 10%.

That is not a contradiction about reality. It is a difference in arithmetic and in what counted as a campaign. If you compare a low reply rate against Belkins you feel great. Compare the same number against Instantly and you feel like you are failing. Neither feeling tells you anything useful. Pick one definition, apply it consistently to your own campaigns, and track the trend instead of the absolute number.

When is it not worth adding more follow-up steps?

When the test does not support it, when your deliverability is not solid, or when the sequence has no real reason to exist beyond persistence.

When a clean test says no. Run the added step on the same list, same offer, same window. If positive replies do not move, stop there. "More is always better" is an assumption, not a finding, and Eric Nowoslawski's data is a reasonable argument that it is often wrong.

When the infrastructure is not ready. Deliverability is a prerequisite, not something you fix later. That means SPF, DKIM and DMARC authenticated on your sending domains, inboxes warmed before they carry volume, sending spread across multiple domains, and bounce rate held to a low level. Adding follow-up volume on top of shaky sender reputation does not get you more meetings, it gets you into the spam folder faster. I run inbox placement tests with a seed list rather than watching opens, and I hold campaigns to a high placement threshold before letting them scale.

When the follow-up has no trigger. A follow-up anchored to something real, a funding round, a leadership change, a hiring surge, a visible project, gives the recipient a reason to respond now. A generic reminder gives them a reason to archive. A short sequence built on a genuine buying signal will usually beat a long generic one, and the research pointing at signal-based outreach lines up with what I see when I build these pipelines. That research and enrichment layer is where most of the leverage sits, which I covered in how AI and automation generate B2B leads.

What is the next step if your follow-ups are not converting?

If you are running outbound and cannot tell whether your extra steps are earning positive replies or just noise, the fix is usually upstream of the sequence: the list, the first email, or the sending infrastructure. If you want a direct read on which one it is for your offer, book a Growth Mapping Call and we will go through your current setup together.

Frequently Asked Questions

How many follow-up emails should a cold email sequence have?

It depends on what you're optimizing for. Platform benchmarks like Instantly recommend a minimum of four touches based on aggregate reply data, but practitioner Eric Nowoslawski's analysis of 1,517 campaigns found two steps optimal, with a sharp drop in replies after three.

Why do cold email benchmark reports recommend more follow-ups than practitioners do?

Benchmark reports average results across every campaign on a platform, including poorly targeted ones where weak first emails make follow-ups look necessary. Practitioners measuring their own campaigns count positive responses specifically, which reveals that extra steps often add noise rather than real interest.

What percentage of cold email replies come from follow-ups?

According to Instantly's 2026 Cold Email Benchmark Report, 58% of replies come from the first email, leaving 42% from follow-ups. Saleshandy's analysis of 53 million cold emails found 44% of positive replies came from follow-up emails rather than the initial message.

Is it bad to send too many cold email follow-ups?

Yes. Eric Nowoslawski's data shows a sharp drop-off in replies after three follow-up steps, and pushing past that point increases spam complaints and unsubscribes, which damages sender reputation and hurts future deliverability rather than just plateauing results.

What is the ideal length of a cold email based on practitioner data?

According to Eric Nowoslawski's analysis of 1,517 campaigns, the optimal email length was 70 to 90 words. Custom-variable personalization written to fit the template outperformed generic AI-generated openers in the same dataset.

Why do cold email reply rate benchmarks vary so widely between reports?

Different platforms use different denominators. Belkins' 2025 report calculates reply rate against total emails sent and found a 0.45% average, while Instantly's 2026 report found 3.43%. The gap reflects differing methodology, not a real difference in campaign performance.

What cadence should follow-up emails follow?

The Digital Bloom's 2025 cold outbound analysis found a Day 0, Day 3, Day 10, Day 17 cadence captures 93% of total replies by day 10, supporting front-loading effort early rather than extending the sequence with more steps.