Average email open rate: what it measures now

Ask two current benchmark reports for the average email open rate and you get answers eight to fifteen points apart. MailerLite's 2026 benchmark, built on 3.6 million campaigns, says 43.46%. Omnisend's 2026 guide (April 2026) calls 28% to 35% healthy. Both reports carry the same footnote: Apple has been opening a large share of everyone's mail automatically since September 2021, so the number being averaged is partly a machine artifact.

What Apple broke, mechanically

Open tracking was always a trick: a one-pixel image, unique per recipient, whose download counted as an open. Mail Privacy Protection ended the trick's usefulness. When an Apple Mail user turns it on (Apple presents the choice at setup, and ESP estimates have put acceptance above 90% since launch; Paubox's write-up is one of several), Apple's proxy servers fetch the message content, remote images included, ahead of and regardless of any human reading it. The pixel fires. Your dashboard records an open. The subscriber may never have seen the subject line.

That shipped on September 20, 2021 with iOS 15, iPadOS 15, and macOS Monterey, and the distortion showed up in industry data within the year: Campaign Monitor's year-over-year figures, reported by MarTech, had opens up 3.5% while click-to-open fell 3.6%. Opens rose because a robot was doing the opening; clicks did not follow because robots do not read.

Nothing here changes what a newsletter is for, and none of it matters until you have one; if you are starting from zero, begin with running a newsletter on your own domain and worry about measurement after issue one.

What the benchmark numbers measure now

A benchmark open rate is now a blend of three populations: real opens, Apple's proxy fetches, and image-blocked reads that were never counted in the first place. The blend differs per audience, which is why honest reports disagree.

MailerLite's report is the most transparent of the current crop, which is why we cite it: 43.46% average open rate across 3.6 million campaigns from 181,000+ accounts, December 2024 through November 2025, published December 2025 and updated April 2026. Their own caveat, quoted: "Apple Mail Privacy Protection automatically marks emails sent to Apple Mail clients as opened. This means real open rates are lower than the metrics you see in your dashboard." They call click rate "currently the most accurate indicator of email newsletter engagement." Omnisend's April 2026 guide lands its healthy range at 28% to 35%, with the same MPP warning attached.

A 43.46% average and a 28-to-35% healthy range are both defensible and both true of their own customer bases. That is the finding. These numbers are measurements of pixel fires across dissimilar populations, useful for watching direction inside one dataset and useless as a pass/fail grade for yours. "Is my open rate good" stopped being a well-formed question in September 2021.

One more number from the MailerLite set is quietly instructive: their average unsubscribe rate more than doubled in a year, to 0.22%, which they attribute to Gmail making unsubscribing easier. A metric moved that hard by a client's interface change is a metric that describes the client as much as the sender. Opens have the same disease.

Our policy: clicks and replies, never opens

Our deliverability playbook scores engagement on clicks and replies only, and has since we wrote it. This is not contrarianism; it is copying the graders. Mailbox providers weight engagement in a hierarchy with replies at the top, clicks below them, and opens a distant third. Gmail rewards recency of engagement, Outlook rewards consistent history, and neither cares what your dashboard's open column says. Opens were the weakest signal available before Apple got involved; now they are partly fiction, so we dropped them from scoring entirely.

The practical consequence shows up in small places. Our welcome email asks the new subscriber one easy question, because a reply is the strongest signal a subscriber can send on your behalf. At our own scale, small real sends from a live workspace since June 2026, replies are countable by hand. We count them anyway, because the habit is the point.

The metrics that still mean something

Click rate on delivered. Divide clicks by delivered mail, not by opens. Delivered is a verifiable denominator; opens no longer are, which also kills click-to-open rate as a serious metric. For scale: MailerLite's cross-industry average click rate is 2.09%.

Reply rate. The one engagement signal no proxy inflates. Apple fetches images; it does not answer your question.

Spam complaint rate. The metric with teeth. Google's bulk sender rules (February 2024) want you under 0.1%, and at or above 0.3% Gmail withholds delivery mitigation until you have stayed under for 7 consecutive days. Watch it in Google Postmaster Tools; since Google retired its reputation dashboards around October 2025, the spam rate is the proxy that remains. We take this seriously enough that our panel plots every program against the 0.1% and 0.3% lines and the send path pauses itself at the ceiling.

Unsubscribe-to-complaint ratio. An unsubscribe is a clean exit; a complaint is a reputation wound. A healthy program loses far more people to the one-click unsubscribe than to the spam button, and if the two numbers ever get close, your unsubscribe is too hard to find. Every campaign we send carries the RFC 8058 one-click headers and honors them immediately; the rule allows two days, and there is no reason to use them.

Bounce rate. Keep it under 2%, suppress hard bounces permanently, and treat a rising trend as a list problem, not a mail problem.

Re-baseline your own program

1. Pull your last ten sends and compute the median open rate and median click-per-delivered. Those are your baselines. From here on you compare against yourself, not against a table.
2. If your platform labels Apple-proxy opens, segment them out once to learn your inflation share. Do not build reporting on the segmented number; the point is to stop trusting the blended one.
3. Add one easy reply prompt to your welcome email, and keep your subject lines honest; humans still decide whether to read, even when a robot has already "opened."
4. Check your spam rate in Postmaster Tools weekly at any volume. Gmail's "bulk sender" label formally starts at 5,000 a day, but the complaint thresholds bite at every size. We follow them at dozens of sends a day, and say so.
5. Grade each send by clicks per delivered and replies, watched as a direction over your trailing median.

Stop asking whether 43% is good. It is not your number, and half of it may be Apple's. Compare this month's clicks per delivered against last month's, keep complaints under 0.1%, and read your replies.