How to write cold emails that get replies (not just opens)
The average reply rate is 3.4%. Elite campaigns hit 8-12%. The difference is not the subject line. It's these three structural changes to your copy.
The average cold email reply rate in 2026 is 3.4%.
Elite campaigns hit 8–12%. The best campaigns we have seen reach 14% on tight ICPs.
The gap between average and elite is not luck. It is structure. Here are the three changes that move reply rates from 3% to 10%+.
1. Write for one person, not a segment
Every cold email tool teaches you to think in segments. "Series B SaaS CTOs in fintech." You write one email. You send it to 500 people who match that filter.
That approach produces a 3% reply rate because the email reads like it was written for 500 people. It is generic by design.
The fix is to flip the process. Pick one person from your list. Write the email for them specifically. Then send that same email to the other 49 people who are structurally similar.
Generic (3% reply rate):
"Hi {{firstName}}, I help SaaS CTOs improve their data pipelines. Curious if this is a priority for you right now?"
Specific (10% reply rate):
"Hi Sarah, saw you just posted a backend eng role focused on Postgres optimization. We helped Stripe's data team cut query times by 60% last quarter. Worth a quick call?"
The second email works because it references something real about Sarah. It name-drops a credible customer. It ties the value prop to a specific problem Sarah is actively hiring to solve.
You can send that same structure to 50 other CTOs who are also hiring backend engineers for database work. It will feel personal to all of them because the signal you referenced is shared.
2. Prove you understand their world
Most cold emails open with a value proposition. "We help X do Y." That structure fails because the recipient has no reason to believe you.
The fix is to lead with proof that you understand their world before you pitch anything.
Generic opening:
"We help sales teams book more meetings."
Proof-first opening:
"Your outbound team is probably hitting the same wall everyone else is right now: reply rates dropping, deliverability getting stricter, and your reps spending half their day on manual list-building instead of selling."
The second version does not pitch. It describes the recipient's current reality in detail. If you get the reality right, the recipient reads the next paragraph. If you get it wrong, they delete.
This is why tight targeting matters. If you are emailing 500 people across ten different industries, you cannot write a proof-first opening that resonates with all of them. If you are emailing 50 people who all have the same problem, you can.
3. Make the ask small and specific
Most cold emails ask for a 30-minute call. That is too much friction for someone who does not know you.
The fix is to ask for something smaller that still moves the conversation forward.
High-friction ask:
"Do you have 30 minutes next week to discuss how we can help?"
Low-friction ask:
"Worth sending over the case study, or is this not a priority right now?"
The second ask gives the recipient two clear options: yes or no. Both are low-effort. If they say yes, you send the case study and book the call in the follow-up. If they say no, you know this is not the right time and you move on.
The structure works because it respects the recipient's time. You are not asking them to commit to a call with a stranger. You are asking them if a piece of content is relevant. That is easy to answer.
The full structure that works
Here is the template that consistently hits 8–12% reply rates:
Line 1: Reference a specific recent signal about the recipient (their last post, their last hire, their last launch).
Line 2: Explain why that signal matters and connect it to a problem they probably have.
Line 3: Name-drop one credible customer you helped solve that same problem, with a real number.
Line 4: Ask a small, specific question that moves the conversation forward.
Example:
"Hi Sarah,
Saw you just posted a backend eng role focused on Postgres optimization. That usually means query times are becoming a bottleneck as you scale.
We helped Stripe's data team cut query times by 60% last quarter without a full re-architecture.
Worth sending over the case study, or is this not a priority right now?
— Adam"
Total length: 63 words. Four sentences. One specific signal. One credible customer. One small ask.
That structure works because every sentence does a job. Nothing is filler.
What ruins reply rates
Lying about mutual connections.
"John suggested I reach out." If John did not actually suggest it, the recipient will check. You will get caught. Your campaign is dead.
Fabricating urgency.
"Following up on my last email." If this is your first email, that is a lie. It ruins trust before the conversation starts.
Generic compliments.
"I love what you are building at [Company]." If you loved it, you would reference something specific about it. Generic compliments read as templates.
Long paragraphs.
Cold emails are read on mobile. If your email requires scrolling to see the ask, most people will not scroll. Keep it under 75 words.
Asking for a favor.
"Would love to pick your brain." You are asking a stranger to give you free advice. That is not a value exchange. Lead with value, then ask.
The constraints that force better copy
The best cold email writers operate under three constraints:
- Under 75 words. This forces you to cut filler and get to the point.
- One specific signal per email. This forces you to do research instead of blasting a template.
- One credible customer name-drop. This forces you to have customer proof before you send.
If you cannot write a good cold email under those constraints, the problem is not your writing. The problem is your targeting, your research, or your lack of customer proof — all of which sit upstream in modern lead generation, not in the copy itself.
Fix those first. The copy follows.
How GenSend enforces this
Every cold email GenSend drafts follows this structure automatically. It pulls your customer list for credible name-drops. It searches for the recipient's recent public signals. It grounds the copy in your actual voice, not a generic template.
The output is not a template with {{firstName}} swapped in. It is a real email written for a real person, using real context.
This is the structure that separates the 8–12% reply rates from the 3.4% average — and GenSend enforces it at the system level, so it happens by default instead of being left to the operator to remember.
When the copy is grounded, replies follow.
