No, your open rate is not accurate, and it has not been since 2021. Treat it as a weak smoke alarm for your sending setup and nothing more. The number that tells you whether a campaign is working is replies, and underneath that, the number of real conversations those replies started.
That is the answer. The rest of this is why the industry keeps quoting opens anyway, and what to put on the dashboard instead.
The number breaks at the source
An open is not something a person does. A sending tool hides a one-pixel image in the email, and the open gets recorded when something downloads that pixel.
For years, the something was a person. Now it is often a machine.
As of September 2026, Apple's Mail Privacy Protection, which is on for anyone who tapped through the prompt, downloads remote content in the background whether or not the recipient looks at the message, and routes it through relays so the sender never sees a real IP address. Corporate security gateways do their own version: they fetch every image and every link in an inbound message to inspect it before a human sees the subject line.
Your tool logs an open. Nobody opened anything.
I am not going to hand you a percentage for how much this inflates the number, because nobody can measure a measurement that is broken in an unknown direction. What you can say honestly is that some unknown share of your opens are robots, that the share moves with the audience you are emailing, and that it changes again every time a provider changes a default. A metric whose error moves on its own is not a metric you can steer by.
Collecting it is not free either
Worth knowing: to count opens, your tool rewrites images through a tracking domain. A fresh tracking domain with no reputation is one more thing Gmail has to decide whether to trust, on top of everything else it already judges about how your sending is set up. Plenty of senders switch open tracking off completely and lose nothing they were actually using.
What the number was standing in for
Nobody really cares about opens. They care about whether the message landed and whether it got read, and the open rate is the only proxy the tool shows by default.
Published averages make it worse by looking authoritative. One widely cited 2026 benchmark report puts the B2B average reply rate at 3.43%, drawn from that vendor's own platform data. A companion report from the same vendor puts the average open rate at 44%, up from 39% a year earlier, and says so itself: open rates "have nearly doubled" because of Apple's changes. The reply figure is useful. The open figure is an average of a broken measurement taken across millions of campaigns, which is a very precise way of saying nothing at all.
Count conversations, not calls
The column that belongs at the center of the dashboard is conversations. A conversation is a human writing back in their own words. Not a bounce, not an out of office, not a pixel firing in a data center in Ohio. I have made the case for setting the monthly goal in conversations instead of booked calls before, and it applies here unchanged: calls are what conversations turn into when there are enough of them.
So the dashboard reads: delivered, replies, positive replies, conversations, calls booked, calls held. Six columns, every one of them countable.
The one job opens can still do
An honest exception. If your open rate falls off a cliff from one week to the next, on the same list with the same sequence, something in your sending probably broke. A domain got flagged, a mailbox got throttled, somebody edited an authentication record.
That is all the number is good for now: a crude alarm that something changed. Even then, bounce rate and reply rate usually tell you the same story sooner, and the sending setup is where you go to fix what they found.
What you must not do is run a test off opens. If you are testing subject lines, the honest readout is the reply rate of each version, which means you need enough volume for the gap between them to mean something.
The numbers that change a decision
Ask what decision a metric changes. Opens change none. You cannot email somebody a better open.
Replies change what you write. Positive replies change who you write to. The gap between calls booked and calls held changes how you handle the booking itself. The gap between a first reply and a second one changes how you follow up.
And those numbers come with benchmarks worth arguing about. Clear 5% on replies and you are doing better than most B2B senders, again on a vendor's own numbers, and 10% is genuinely good.
In the campaign dashboards I look at every week, the open column is the first one I collapse. No decision has ever come out of it.