Open Rate

For a long time, sending an email was the last thing you knew about it. The record showed that ten thousand messages went out, and everything after that was blank. Whether those ten thousand people read it, deleted it, or ever received it at all was simply unavailable.

So the conversation stalled in the same place every time. Someone asked whether last week's email worked, and the answer was that revenue had ticked up but nobody could say if the email had anything to do with it. Since you could not tell whether a new subject line was better, choosing subject lines stayed a matter of taste.

The mechanism email marketing tools built to fill that gap is the tracking pixel. A single invisible one-pixel image is embedded in the message body, and when the recipient's mail program fetches that image from the server, the message is counted as opened.

Once opens could be counted, subject lines became something you could test. Send the same content under two subject lines, see which gets opened more — that comparison is what this number was built for, and it is still what it does best.

Divide the number of opens recorded by the number of messages that reached an inbox.

The first thing to notice is that the denominator is deliveries, not sends. Send 10,000 messages and have 600 land in spam or bounce back from dead addresses, and deliveries come to 9,400. Against 2,350 opens, that is 25 percent; against sends it would read 23.5 percent. Tools differ in which they use by default, so before comparing against last month's figure, check that both numbers share a denominator.

You also have to decide how many times one person counts. If someone opens the same email on the commute and again at lunch, that is two opens and one person. The first is called total opens, the second unique opens, and open rate almost always uses unique.

The calculation has three steps.

1. Subtract bounces and spam-filtered messages from sends to get deliveries.
2. Collapse repeat opens by the same person to get the number of people who opened.
3. Divide step 2 by step 1. Of 9,400 delivered, 2,350 people opening gives 25 percent.

What the calculation actually measures is not the act of opening but the record of an image being fetched. Someone who blocks images can read every word without registering as an open, and a mail program that prefetches images can register an open for a message nobody looked at.

Pulling an industry benchmark in the low twenties and grading your own number against it is the most common mistake here, because the same company's emails vary threefold depending on which kind they are.

Compare by type instead. Messages tied to something the recipient just did — a failed payment notice, an invitation alert — routinely clear 50 percent, because the reader was already waiting for them. A regular newsletter sits in the twenties. A discount or event announcement sits around ten percent.

Three reference points do the work. Your own figure for the same type of email last time. The figure before and after a subject line change. And the figure this number would have to reach for a target click count to be possible, worked backwards.

It also cannot be read alone. Ahead of it sits deliverability, and when a lot of mail lands in spam, the open rate computed over the surviving few looks high. Behind it sit click-to-open rate, the share of openers who clicked a link, and unsubscribe rate, the share of recipients who ended their subscription.

Read together, the four narrow down what to fix. A 40 percent open rate with a 2 percent click-to-open rate means the body is the problem, not the subject. A 5 percent open rate with a 3 percent unsubscribe rate means you are mailing people who never wanted it. And if the open rate has slid from 25 percent to 12 percent over six months, the list has aged — mail providers factor unengaged recipients into sender reputation, so leaving it alone drags deliverability down with it.

"We split the subject line. A got 31 percent, B got 19. Sending A to everyone."

Using the number for the comparison it was designed for.

"Open rate is 38 percent and almost nobody clicked. I think the subject promised more than the body delivered."

The expectation that earned the open and the content behind it did not match.

"I'm not setting an open rate target. The measurement is too shaky, so we'll work off clicks."

Moving the decision criterion because of a known limitation in how the number is collected.

At Pagedeck, Ravi splits the subject line on the weekly product email. Two versions go to 10 percent of the list, and a day later the winner goes to the rest. He also tries personalization in the subject — the recipient's name, or the title of the last document they opened — though how much it helps depends on the type of email.

Mia keeps open rate and click-to-open rate on one screen. In a week where clicks do not follow, the team leaves the subject alone and looks at what the body promised. In a week where the unsubscribe rate clears 0.5 percent, sending frequency gets cut first.

The most common error is reading a high open rate as a successful campaign. It is the first number on the report and the largest one, so it invites that reading. But an open only tells you the subject line and sender name worked. If there is no reason to go further, the open rate stays high with nothing behind it. Optimize this number alone and subject lines get louder over time. People who felt misled accumulate, and six months later the open rate is flat while the unsubscribe rate has climbed.

The second error is trusting the figure as precise. It arrives with a decimal point, which makes it look exact. But it is an indirect measurement based on whether an image loaded. Environments that block images by default drop real readers out of the count, and environments that prefetch images add people who never opened anything. The second kind grew sharply after 2021, which pushed many teams' open rates up overnight and made figures from before and after that shift impossible to compare on one line. That is why a lot of teams now treat it as a subject-line comparison tool rather than a target, or move to click-based measures instead.