How to improve email open rates
Open rate stopped being a reliable number in 2021. What it still tells you, what actually moves it, and how to optimise without fooling yourself.
Performance · 5 min read · updated
Read this before you optimise anything
Open tracking works by embedding a 1×1 pixel and counting the request for it. Since Apple Mail Privacy Protection shipped in 2021, Apple Mail fetches that pixel on its own — before anyone has read anything, and for every message that arrives. Gmail proxies and caches images too. A large share of the opens in your dashboard were never a human being.
In practice reported open rates now land somewhere in the 35–45% range for lists with a normal Apple share, while the human number underneath is closer to 20–30%. Nobody can tell you which of your opens are which, because the machines do not identify themselves.
That does not make the metric useless. It makes it a trend line rather than a score. On a stable audience, a sudden fall in opens is a real signal — usually deliverability, occasionally a broken send. What you must not do is A/B test on opens, report opens to a board, or celebrate a number that moved because your Apple share moved.
The four things that actually move opens
In descending order of leverage. Most people spend their time on the third one.
- Inbox placement. Nothing else matters if you are in the spam folder. Placement is the ceiling; the subject line only decides how much of that ceiling you use. If opens fell off a cliff, start here, not with copy.
- Who it is from. The sender name is the first thing read and the biggest single driver of the decision. Recognition beats cleverness every time.
- Subject line and preheader. Worth real effort — but it is the third lever, not the first, and the gains are measured in single-digit percentages.
- Who you sent it to. Mailing people who stopped caring drags the average down and damages placement at the same time. Sending less, to fewer, more relevant people, raises this number twice over.
The sender name is worth more than the subject line
People decide whether to open based on whether they recognise you. Keep the From name consistent — changing it every campaign resets that recognition and looks, to a filter, like a sender trying not to be pinned down.
A human name plus the brand (“Sam at InkPigeon”) usually outperforms the brand alone for newsletters and lifecycle mail. For receipts and alerts, the brand alone is right — people are looking for the company, not a person.
Send from an address that can receive replies. noreply@ tells the reader you are not interested in hearing back, and it throws away the reply signal that mailbox providers read as a strong positive.
Subject line and preheader are one unit
The preheader is the text the client pulls from the top of the body and shows next to the subject. It is a second subject line you get for free, and most senders waste it: if you do not set one deliberately, the client shows whatever comes first — often “View this email in your browser”.
Write them together, as a sentence and its completion. The subject makes a specific promise; the preheader adds the detail that makes it worth acting on.
Front-load. An iPhone shows roughly 35–45 characters of subject before it truncates, and most people see your mail on a phone. The payload goes at the start, not after a brand prefix.
Send to people who want it
- Segment before you write. A message written for one segment beats a message written for everyone, and the writing gets easier because you know who is reading.
- Run a sunset policy. Decide the window — 90, 180, 365 days of no opens and no clicks — try one re-engagement message, then stop mailing them. This feels like shrinking your list. It is removing the part of the list that was suppressing the rest.
- Let people choose frequency. A preference centre that offers “monthly instead” catches people who would otherwise unsubscribe or, worse, hit the spam button.
- Be boring about timing. Best-time-to-send studies contradict each other because the answer is specific to your audience. Test it on your own list, expect a small effect, and do not rebuild your calendar around somebody else's blog post.
How to test without lying to yourself
Change one variable per test. Two changes and a result tells you nothing about which one did it.
Give each arm enough recipients that the difference is not noise. As a rough floor, a few thousand per arm to detect the couple of percentage points a subject-line change actually produces — with 200 recipients you are reading randomness.
Judge the test on clicks or conversions, never on opens alone. A subject line that wins on opens and loses on clicks did not win; it overpromised, and the reader paid for it.
Common questions
What is a good email open rate?
Reported open rates commonly sit between 35% and 45% today, but that number includes machine opens from Apple Mail Privacy Protection and image proxies. The honest human figure is nearer 20–30%. Compare against your own history on the same audience, not against an industry table.
Why did my open rate suddenly jump?
Usually a change in audience mix rather than performance — more Apple Mail users, or a new segment. Check whether clicks moved by the same proportion. If opens rose and clicks did not, nothing real happened.
Should I still track opens at all?
Yes, as a trend on a fixed audience and as an early warning for deliverability problems. Do not use it as the success metric for a campaign, and do not put it in a bonus target.
Check your addresses now
The free InkPigeon checker runs syntax, DNS, disposable, typo and role-account analysis on any address — instantly, no signup.