Open rates are the most visible number in cold email, and for a lot of agencies they become the target. The problem is that opens are recorded after the recipient has already decided whether your email deserved attention. They are a record of a judgment, not an input to it. By the time the number exists, the thing it measured is over.

That framing comes from a post by Nithyanandan Ramakrishna titled "Don't Chase Open Rates, Fix Your Email's Soil," which uses a farming analogy worth sitting with. A farmer plants rice. The shoots come up slowly, so he tugs each one a little higher to help. By evening the field looks taller. By morning every plant is dead.

You can read the full post here: the original LinkedIn post

Why open rate is a lagging indicator

Most metrics in a funnel are inputs you can act on. Open rate is not one of them. It is downstream of everything that actually determines whether someone engages: who the email is from, whether the recipient recognizes the sender, whether the subject line is honest, and whether the first line gives a real reason for the email to exist.

When you treat open rate as the target, you start writing for the open rather than for the reader. The post lists the familiar tactics: "Re:" on a first contact, "Quick question," a subject line implying a conversation that never happened. These work in the narrow sense that opens go up. Then the recipient opens the email, realizes they were tricked, and you have done worse than fail to persuade them. You have told them something about yourself before saying anything at all.

That is the part agencies tend to miss. A deceptive subject line is not a neutral optimization. It is a first impression, and it is a negative one.

The mechanism, not a side effect

Watch what happens after you optimize for opens with tricks. Opens rise. Replies fall. Complaints rise. The metric you optimized went up and every metric that mattered went down.

That is not a side effect. It is the mechanism. You are selecting for people who will open anything, including people who will then mark you as spam because the content did not match the promise. Complaint rates are one of the signals mailbox providers watch when deciding how to treat your future mail, so a spike here does not stay contained to one campaign.

There is also a measurement problem layered on top. Many opens are not real opens. Some mailbox providers prefetch or proxy images, and security scanners can trigger a pixel load without a human ever seeing the message. So even the number you are chasing is partly noise. If you are making decisions on open rate alone, you may be optimizing against a signal that includes bots and scanners alongside real readers.

Fix the soil, not the shoots

The post's advice is direct: stop pulling on the only part of the plant you can reach. Fix the soil instead. That means three things.

1. Who it is from

The sender identity is the first thing a recipient evaluates, often before the subject line. Is the from name a real person or a vague brand? Does the domain match where the email claims to come from? Is the reply-to address monitored by a human? These are foundational. If the sender looks unfamiliar or slightly off, no subject line rescues the email.

This is also where authentication lives. SPF, DKIM, and DMARC are the technical layer that tells mailbox providers your mail is legitimately from your domain. They do not guarantee inbox placement on their own, but missing or misconfigured authentication makes everything else harder. For agencies sending on behalf of clients, getting the sending domain and authentication right is table stakes, not an optimization.

2. Where the list came from

A purchased or scraped list is bad soil no matter how good the seed is. Recipients who never opted in are more likely to ignore, complain, or mark as spam, and those signals accumulate against your sending reputation. A smaller list of people who have a plausible reason to hear from you will outperform a large list of strangers, even if the large list produces more opens on paper.

3. Whether the first line gives a real reason this email exists

This is the part most cold emails skip. The first line should tell the recipient why they are hearing from you and why now. Not a compliment about their website, not a fake follow-up, but an actual reason. If you cannot articulate one, the email probably should not be sent.

None of these three fixes moves a number today. All of them decide whether the number ever meant anything.

What to measure instead

If open rate is a poor target, what should agencies watch? The post points toward intent signals: what the recipient actually did with the sent email. That is a more honest read than a pixel firing.

In practice, that means leaning on:

  • Replies, including negative replies. A "not interested" is still a human response and a signal that a human read the message.
  • Clicks that are verified rather than assumed. A click is a stronger signal than an open because it requires deliberate action.
  • Bounces, especially hard bounces, which tell you about list quality and can hurt reputation if they pile up.
  • Complaints and unsubscribes, which tell you whether your targeting and messaging match the audience.
  • Reply sentiment, which tells you whether the message landed with the right people even when the answer is no.

The post notes that Mailheight is strict about what counts, filtering bots out of opens, verifying clicks rather than assuming them, and scoring intent from what the recipient did with the sent email. The broader point applies regardless of tooling: if your measurement includes noise, your optimization will chase noise.

A practical reset for agencies

If your team has been reporting on open rate as a primary KPI, here is a reasonable way to reset without disrupting client relationships.

First, stop reporting open rate as a headline number. Move it to a secondary line, or drop it entirely if you cannot separate human opens from machine ones. Second, shift the primary conversation to replies and positive reply rate. That is the metric closest to revenue and hardest to fake. Third, audit the soil: sender identity, domain authentication, list source, and the honesty of your subject lines and first lines. Fourth, watch complaint and bounce rates as guardrails, not afterthoughts.

This will feel slower. It is slower. The rice farmer's field looked taller by evening, and that was the whole problem.

The takeaway

Open rate is a record of a decision the recipient already made. You cannot optimize it directly without distorting the thing it was supposed to measure. What you can do is fix the conditions that make a real person want to open, read, and reply: a recognizable sender, a list that belongs to you, and a first line that earns the next sentence. None of that shows up in today's dashboard. All of it decides whether the dashboard means anything next quarter.