How to improve a low cold email reply rate: a diagnostic order

Do not rewrite the copy first. Work down five layers in order: did the email reach the inbox, did it go to the right people, is the offer worth answering, does the copy make the offer clear, and did you follow up. Each layer has a cheap test. Fixing a higher layer is worth more than polishing a lower one, because a perfect email in the spam folder gets zero replies.

Most advice on a low reply rate starts with subject lines. That is the layer people can see, so it gets the attention, but it is rarely the cause. This post gives a fixed order for the diagnosis, a test for each step, and a made up example showing how the order saves time. Figures about mailbox providers come from Google's own sender guidelines, read on 25 October 2026.

First, what is your reply rate, exactly?

You cannot fix a number you have not defined. Our post on how we define reply rate sets out the three choices that change the result: the denominator (people contacted or emails sent), whether automatic replies count, and whether a refusal counts. Pick one definition and use it for every campaign. Otherwise you will compare your 2% with a stranger's 8% that was counted another way.

Also separate the two ways a campaign can fail. Low replies with a healthy inbox rate is a message problem. Low replies because the mail never arrived is an infrastructure problem. The order below finds out which.

The five layers

Layer 1: did it reach the inbox?

Start here because it is the only layer where the answer can be zero. Three quick checks.

1. Authentication. Google's guidelines say all senders need SPF or DKIM, and bulk senders (over 5,000 messages a day to Gmail) need SPF, DKIM and DMARC. Even below that volume, set up all three. Our cold email infrastructure post lists the records.
2. Spam rate. Google says to keep the spam rate shown in Postmaster Tools below 0.3% and recommends staying below 0.10%. If you send too little for Postmaster Tools to show data, treat any complaint as a warning.
3. Bounces. A rising bounce rate means the list has bad addresses and providers will notice. Verify before you send; see email verification before send.

Do not use open rate as proof of placement. Privacy features inflate and distort opens, which is the subject of our post on open rates. Warm up helps a new mailbox, but warm up is not enough if the content or list is the problem.

A direct test: send the exact email from your sending mailbox to three or four inboxes you control at different providers. If it lands in spam or Promotions on all of them, stop and fix this layer. If it lands in the inbox, move on.

Layer 2: is the list right?

A weak list is the most common cause of a reply rate that never improves, because no wording can make the wrong person want the offer. Open the file and read 30 random rows as if you were the recipient.

• Would this person plausibly have the problem you solve?
• Do they have the authority to answer, or to pass it on?
• Is the company the right size and type?
• Is the email address one a human reads (not info@ or a role inbox that nobody owns)?

If more than a few of the 30 fail, the list is the problem. Rebuild it from a clear ideal customer profile, and tighten the filters until the 30 row test passes. A smaller list that fits beats a larger one that does not.

Layer 3: is there an offer worth answering?

An offer is what the person gets for replying. "Can I tell you about our platform?" is not an offer. "A list of the 15 hotels in your region that changed their booking page this month" is. Write the offer in one sentence with no product words. If you cannot, the email has nothing to reply to.

Three common offer fixes:

1. Make the ask small. A yes or no question beats a request for 30 minutes.
2. Make it about them. Lead with something specific you saw, not a feature.
3. Name the result in their terms. Their metric, not yours.

Layer 4: is the copy clear in ten seconds?

Only now is it worth editing sentences. Read the email aloud and check:

• Is the first line about the reader?
• Does it fit on a phone screen without scrolling? See how long should a cold email be.
• Is there one clear ask?
• Is anything in it a claim you cannot back?

Change one element at a time and keep the batches the same size, as described in A/B testing cold emails. With small batches a difference of one or two replies is noise.

Layer 5: did you follow up?

A one email campaign gives up every reply that a later message would have earned, because people miss or postpone the first one. Sellinger's email defaults are 5 follow ups before a lead is treated as exhausted, with gaps starting at 1 to 3 days, and both can be changed per campaign. Treat these as a starting point, not a rule. Each follow up should add something new rather than say "just checking in"; the examples in follow up copy that worked and failed show the difference.

A made up example

This is an invented example to show the order, not a customer.

A consultancy sends 400 emails and gets 3 replies, a 0.75% reply rate. The owner wants to rewrite the subject line.

1. Deliverability. Seed test: two of four inboxes put the email in spam. The domain has SPF but no DKIM. They add DKIM and DMARC, pause for a few days and retest: four of four inboxes.
2. List. Of 30 random rows, 9 are assistants or the wrong department. They remove the role types and re-verify.
3. Offer. The ask was "a quick intro call". They change it to one question about how the team handles a specific task.
4. Copy. Cut from 140 words to 70.
5. Follow ups. They add two follow ups, each with a new angle.

Notice that the subject line was never touched. The first two fixes were free and probably explain most of the change. In a real campaign you would still test, because you will not know which fix did the work unless you change one thing per batch.

When to stop and re-plan

If layers 1 to 3 pass, the copy is clear and you have followed up, and replies are still low, the market may not want the offer. Then the test is no longer email craft but whether this audience has this problem now. Try a different segment or a different trigger, or an intent signal that shows who is active, before writing more email. Our note on counting meetings honestly helps you judge success by what comes after the reply.

This post is general guidance, not legal advice. Cold email rules differ by country; see is cold email legal.