For years the first number every sender checked was the open rate. It is now one of the least reliable. Some opens are real people, some are machines, and some real people are invisible. If you are still deciding what to change in a campaign because of it, you may be reading noise.
How an open is measured
Most tools add a tiny invisible image to the email. When the email program loads the image, the tool records an open. That works only when a real person opens the message and their program loads images. That used to be a fair guess. It is not any more.
Why the number is unreliable
- Apple Mail Privacy Protection. Apple says it "helps users prevent senders from knowing when they open an email." In practice it loads images ahead of time, so messages can show as opened when nobody has read them.
- Image proxies. Some mail programs fetch images through their own servers, so the open reflects the server, not the person, and the location it reports is not theirs.
- Security scanners. Company filters often open links and images to check for threats, and count as opens.
- Blocked images. Some real readers never load images, so a real open is not counted.
The result is a number that can be too high and too low in the same campaign. If most of your list uses Apple's Mail app, it can look wonderful while nobody is reading.
What to track instead
- Reply rate. The only number that cannot be faked by a machine. See what is a good cold email reply rate.
- Positive replies. Count the ones that ask a question or agree to talk, and keep them separate.
- Meetings booked per 100 emails. It ties your emails to the result.
- Bounce rate. If it climbs, your list is stale. See cold email bounce rate.
- Spam complaints. Visible in Postmaster Tools if your volume is high enough.
How to test a subject line without opens
Split a list in two, send each half a different subject line to the same kind of person, and compare replies after a week. You need enough sends for a difference to mean something. Fifty per side is a rough floor, and a gap of one or two replies is not a result. For ideas, see subject lines that get opened.
Should you turn off open tracking?
Some senders turn it off, because the image can look like a marketing email and the number misleads. Others keep it for a rough signal. Either way, do not let it decide your next move. Decide on replies.
A fair way to use opens, if you use them
If you still look at opens, treat them as a rough smoke alarm rather than a score. A sudden drop from a normal level to almost nothing can point to a deliverability problem, for example mail landing in spam, and is worth checking. A smooth number that never changes tells you little. Pair any change with the counts that cannot be inflated: replies and bounces. When the two disagree, believe the replies. And never compare your open rate with a figure from a blog post; the mix of mail programs on your list is different from theirs.
If you run outreach from your own mailbox at low volume, you can skip tracking pixels entirely and lose nothing you could trust. A plain text email with no images also looks more like a message from a person, which is what you are. Read each reply yourself, mark it as positive, neutral or negative and keep the tally in a sheet. Within a month you will know your real response rate better than any dashboard could tell you.
A weekly check
Every Friday, write down emails sent, replies, positive replies and meetings. Four weeks of that is more useful than a year of open rates.
Rocketship reports replies and meetings, since those are the numbers that matter, and answers replies for you. Try it for 7 days, card required to begin.