Why open rates are approximate

Open rates look precise and aren't, in either direction. An open is really "the tracking pixel loaded," and between privacy proxies that load it for people who never read the email and image-blocking clients that never load it for people who did, the number is an estimate wearing a percentage sign. SEMAOS shows open rates because they're still useful directionally, and flags them with a caveat everywhere they appear. Here's what's behind that caveat.

How do opens get overcounted?

Apple Mail Privacy Protection is the big one. Since 2021, Apple Mail fetches every image in every email through Apple's proxy at delivery time, whether or not the recipient ever opens the message. To a tracking pixel, that fetch is indistinguishable from a real open, so every Apple Mail recipient registers as an opener. With Apple Mail's share of the inbox market, this alone can inflate an open rate by 20 points or more on a consumer-heavy list.

Security scanners do the same thing on corporate mail: link-scanning and image-prefetching gateways open messages before employees see them. SEMAOS filters the opens and clicks it can identify as non-human (known scanner and bot signatures), which removes the obvious machine noise, but proxy opens like Apple's are engineered to be indistinguishable and no vendor filters them honestly.

How do opens get undercounted?

The mirror image: an open only registers if images load, and plenty of real readers never load them. Outlook's default configurations, corporate security policies, and privacy-minded individuals all block images, so their genuine reads count as silence. Plain-text-preference readers never fire a pixel at all.

One more SEMAOS-specific case: mailbox-channel sends don't track opens by default. Tracking on 1:1 mail is off unless you've opted in under Settings → Sending, so a sequence running through your mailbox shows no opens by design, not by failure. If a send shows zero opens, check which channel it used before concluding anything.

So which numbers can I trust?

Clicks and replies. A click is a deliberate human action on a specific link (and SEMAOS filters known bot clicks), and a reply is the strongest signal email produces. Engagement scoring weighs them accordingly, and the sane reading order for any campaign is replies first, clicks second, opens last and loosely.

Opens still earn their place for relative comparisons: the same list, the same channel, two subject lines. The biases roughly cancel when everything else is held constant, which is why subject-line A/B tests can use open rate as a signal with floors and margins built in. What opens can't support is absolute claims ("42% of prospects read this") or comparisons across different lists and clients.

What should I do differently?

Judge campaigns by replies and clicks. Use opens to compare like against like, never as absolute truth. And when a number looks impossible in either direction, check the audience's inbox mix before blaming the mail: a list full of Apple Mail users will "open" everything, and a list behind corporate image-blocking will "open" nothing, while both read exactly as much as they ever did.