All articles

Outreach strategy

Your Cold Email Open Rate Looks Good. Why Aren’t You Getting Replies?

What email opens actually measure, how to classify replies, and how to review the conversations behind your campaign numbers.

A lavender magnifying glass examines an envelope and analytics bars beside a mint reply bubble.

A campaign report shows a sixty percent open rate. The inbox has two replies: an out-of-office message and someone asking to be removed.

Was the subject line working? Did the offer miss? Should you send another follow-up?

The open rate cannot answer those questions on its own. Yet it can easily become the number that determines what a team changes next.

For teams running outreach through Gmail, the useful question is what happened between sending a message and getting a meaningful response. That requires a closer look at what an “open” actually measures, how replies are counted, and whether the campaign reached people with a reason to care.

An open is a tracking event

Email open tracking commonly relies on a small, invisible image embedded in a message. When that image downloads, the system records an open.

That creates two problems. Someone can read an email with images blocked and leave no recorded open. Automated activity can also load content without the recipient reading the message. Mailchimp documents both limitations in its explanation of how open tracking works.

Apple adds another complication. Its Mail Privacy Protection prevents senders from reliably seeing whether a recipient opened a message. Mailchimp also warns that privacy features and bot activity can inflate reported engagement. Apple’s explanation makes the purpose clear: the recipient’s reading activity is private.

A recorded open therefore gives you an imperfect signal. It does not establish that someone read the offer, understood it, or considered replying.

Google makes a related point in its sender documentation: it does not track open rates, cannot verify third-party open-rate reports, and says low opens do not necessarily indicate a delivery problem. Google’s email sender guidelines are more cautious than many campaign dashboards.

That matters when interpreting results. “High opens, low replies” is a reason to investigate. It is insufficient evidence to declare the subject line successful and the body copy broken.

Count the people who answered—and what they said

Reply rate gets closer to the outcome you want, but it still needs a definition.

An automatic holiday response, a removal request, a referral to a colleague, and a request for pricing all produce activity in an inbox. They have very different meanings for the business.

A useful review separates:

  • Automatic responses: absence notices and other system-generated messages.
  • Negative responses: declines and requests to stop contact.
  • Routing responses: introductions or suggestions to contact someone else.
  • Positive responses: relevant questions, requests for details, or agreement to a next step.

Then count unique prospects, so a long conversation with one person does not look like several successful contacts.

Consider this hypothetical comparison. Both campaigns reached 500 prospects, with no recorded bounces, and had the same time to collect replies.

ResultCampaign ACampaign B
Reported open rate64%39%
Unique human respondents2015
Positive respondents38
Meetings booked14

Campaign A leads on opens and total human replies. Campaign B produces more positive responses and meetings.

Using contacted prospects as the denominator, the positive reply rates are 0.6% and 1.6%. Those figures are calculations for this example, not industry benchmarks.

The point is practical: choosing a winner depends on what the campaign was supposed to accomplish.

Read the replies before rewriting the email

Once responses are classified, read their actual wording.

“We already use another provider” suggests a different problem from “I don’t handle this.” The first may raise questions about differentiation or switching costs. The second raises questions about contact selection.

“Send more information” needs interpretation too. It could signal interest, or it could be a polite way to end the exchange. What happens after you send the information helps distinguish the two.

Try reviewing twenty recent conversations alongside the original lead records. For each prospect, ask:

  • What evidence made this person a suitable contact?
  • What specific problem did the message address?
  • Did the response support that assumption?
  • What happened after the first reply?

This is a suggested review process, not a validated scoring model. Its value comes from connecting the result to the decision that produced it.

If the same objection appears repeatedly, you have something concrete to investigate. If responses vary widely, your audience may need narrower segments before another copy test will tell you much.

Personalization needs a factual reason to exist

A first name and company name make an email look addressed. They leave the sender’s central task unfinished: explaining why the offer belongs in this person’s inbox.

Take an outreach product as an example. A message to an agency owner might focus on coordinating follow-ups across client campaigns. A message to a founder might focus on managing an initial prospect list alongside other work.

Those are different situations. Giving both people the same paragraph with different company names loses useful context.

Specificity also creates a responsibility to check the facts. Seeing that a company has hired salespeople does not prove its follow-up process is failing. A credible message distinguishes an observation from an assumption.

For example:

“I saw your two sales hiring posts. If outbound is part of the expansion, how are you planning to manage follow-ups across the team?”

That is an illustrative opening, not a tested template. It offers a relevant question without pretending to know the company’s internal problems.

Keep delivery checks separate from message experiments

Poor results can have several causes at once. Rewriting the offer will not resolve an authentication error. Fixing authentication will not make an irrelevant offer useful.

Google recommends authenticating sending domains, maintaining consistent sending patterns, and monitoring delivery errors and recipient feedback. It also advises reducing volume when messages start bouncing or being deferred. These are operational checks to make alongside campaign analysis. Google’s sending guidance explains the details.

Before testing another message, inspect any available rejection notices, bounce information, and authentication results. Delivery acceptance still does not prove inbox placement, but these checks provide evidence that an open-rate percentage cannot.

How this applies to a SendRush workflow

SendRush describes a workflow that connects Gmail, imports leads from CSV or Google Sheets, uses lead fields and business context for personalization, and coordinates scheduled messages and follow-ups. Its public site also describes stopping sequences on reply and keeping conversations in their original Gmail threads. SendRush’s product overview outlines those capabilities.

That workflow gives a team a practical place to apply this approach:

Start with one clearly defined audience. Review the evidence behind the personalization. Decide what counts as a positive response before sending. When replies arrive, classify them and record the next step, using a separate sheet if needed.

Stopping on reply is an automation rule. Understanding the reply remains a judgment call.

For the next campaign review, bring three things: the audience definition, the response breakdown, and a handful of actual conversations. Use open rates as supporting context.

A prospect asking a relevant question gives you something you can act on. Build the review around finding more of those conversations.

Research and sources are linked throughout. Examples and suggested workflows are illustrative.
Back to the journal

Keep reading