Quick answer
No. An open is logged when a tiny tracking image loads, and that no longer tells you a person read your email. Apple Mail with Mail Privacy Protection downloads remote content in the background "regardless of whether you engage with the email". Corporate security scanners open messages to check them. Classic Outlook blocks pictures by default, and Gmail holds back images from senders it finds suspicious, so some real reads never get counted. Your open rate is inflated in some inboxes and missing in others, and you can't tell which. Instantly, Smartlead, lemlist and Salesforge all link open tracking to deliverability risk in their help pages, and in Hunter's 2025 data sequences with tracking on got a 4.4% reply rate against 7.4% without. For cold email, switch open tracking off and judge campaigns on replies, positive replies and booked meetings.
How open tracking works
Who's writing: a GTM consultant working on cold email and LinkedIn outbound with B2B teams, with 2000+ meetings booked for clients. Before deciding whether to trust an open rate, it helps to know what an open is.
Your sending tool puts an invisible image, usually 1x1 pixel, into the HTML of each email, with a unique address per recipient. When something requests that image from the tool's server, the tool logs an open. lemlist's help center describes it the same way, and Microsoft's Outlook help calls tracking pixels "invisible images that can tell a sender you've read the email".
So an open proves one thing: something loaded the image. It doesn't say who or what loaded it, or whether anyone read a word. A plain-text email carries no image at all, which is why a text-only send can't track opens.
Four reasons the number is wrong
| Where the email lands | What happens to the pixel | Effect on your open rate | Source |
|---|---|---|---|
| Apple Mail with Mail Privacy Protection | Remote content downloads in the background when the email arrives | Opens counted for emails nobody read | Apple |
| Gmail | Images served through Google's proxy, and not shown when Gmail finds the sender or message suspicious | Opens counted when read, missed when images are held back | |
| Classic Outlook for Windows | Automatic picture downloads blocked by default | Real reads missed until the reader allows pictures | Microsoft |
| Corporate security scanners | Mail servers open messages and follow links to check for threats | Opens and clicks from machines | lemlist, Smartlead |
Apple Mail. Apple's page on Mail Privacy Protection is plain about it: when you receive an email in the Mail app, "Protect Mail Activity downloads remote content in the background by default, regardless of whether you engage with the email." It's about the Mail app, not the email provider, so a Google Workspace or Microsoft 365 address read in Apple Mail is affected too.
Your tool won't show you how much of your list that covers. Litmus, which measures opens for email marketers, put Apple at 62.26% of over 1 billion opens in July 2026, per its market share report, and says privacy-protected opens "are not considered reliable opens". That's marketing email, though. A B2B list will have its own mix.
Gmail. Google's admin help says Gmail "uses Google's secure proxy servers to serve images" when people open messages, so an open still registers, without the reader's location or device. But Gmail's own help adds that if it "thinks a sender or message is suspicious, images aren't shown", and says only that "sometimes" senders may know an email was opened. Google doesn't say how often that happens, so you can't correct for it.
Outlook. Microsoft's support page says classic Outlook "is configured by default to block automatic picture downloads from the Internet", and lists avoiding tracking pixels as one of the reasons.
Security scanners. lemlist's help center says some email servers "open emails and click links to scan for threats". Smartlead's guide to open rate accuracy says a corporate security gateway "can fire your tracking pixel in milliseconds", before the person even sees the notification.
Does open tracking hurt deliverability?
Every major sending tool can track opens. Here is what four of them say about it in their own help pages, checked on 15 October 2026:
| Tool | What its help center says |
|---|---|
| Instantly | Its text-only Delivery Optimization setting "disables open tracking to improve deliverability". Set up a custom tracking domain before you turn open tracking on. |
| Smartlead | Opens and clicks are tracked by default, and the setting is locked once a campaign starts sending. Disabling tracking "may help in certain deliverability strategies", and Apple's Mail Privacy Protection "can inflate or suppress opens". |
| lemlist | Opens and clicks are tracked by default. Its lemcoach guide says to stop tracking clicks and opens in cold email and to focus on reply rate. |
| Salesforge | "We do not recommend tracking open rates because it might impact the deliverability", and it "shows to the recipient a warning message which can hurt the credibility of the e-mail." |
Salesforge is what I run for clients. I've written guest articles for its blog, and no tool in this article pays me.
Two published datasets point the same way. Hunter's State of Cold Email report, built on 31 million emails its users sent in 2025, found sequences with tracking on had a 4.4% reply rate, against 7.4% without, and advises: "don't track your sequence open rate and link clicks unless you absolutely need to." Belkins switched open tracking off for every campaign in its 2025 reply rate study and says the tracking pixel was hurting deliverability across the industry in 2024.
Neither proves the pixel alone costs replies: teams that switch tracking off may also send plain text to clean lists. But none of the four tools argues that tracking helps deliverability, and turning it off only costs you a number you couldn't trust.
Where open data leads you wrong
- Subject line tests. If Apple Mail opens everything and Outlook opens nothing, the "winning" subject line is the one that went to more Apple Mail readers. lemlist's guide calls open rates "too unreliable for accurate A/B testing". Judge subject lines by replies, as in these subject line formulas.
- "Opened but didn't reply" steps. Tools let you act on opens: Salesforge has a "Pause on open" setting, and Instantly shows the leads who opened but haven't replied. A scanner or Apple Mail can put someone in that bucket who never saw the email.
- Chasing "hot" openers. A list of people who opened three times is often a list of Apple Mail users and security gateways.
- Reports to a founder or a client. A steady open rate can sit next to a falling reply rate. Only one of them books meetings.
And never mention the open in a follow-up. It reads like surveillance, and often it's wrong:
"Hi Anna, I noticed you opened my last email a few times, so I thought I'd follow up."
Write it as if you know nothing, with a new angle:
"Hi Anna, one more thought. Teams opening a second office usually decide whether outbound for it runs from HQ or a local hire. Is that a call you're making this quarter?"
What to measure instead
Count the things only a person can do, plus the two numbers that tell you whether the email can reach a person at all:
| Metric | What it tells you | How to count it, or the limit |
|---|---|---|
| Reply rate | Someone read the email and cared enough to type | Unique people who replied, auto-replies removed, divided by contacts emailed |
| Positive reply rate | Interest: yes, send more, ask me next quarter | Tag every reply, then divide positive replies by contacts emailed |
| Booked meetings | The number your pipeline runs on | Meetings booked per 1,000 contacts |
| Hard bounce rate | List quality, and risk to the sending domain | Keep at or below 2%, per Instantly |
| Spam rate | Complaints from Gmail users | Keep below 0.10% in Postmaster Tools and never reach 0.30%, per Google |
For what normal looks like on each, see realistic cold email benchmarks. If replies are low, work through why your reply rate is low in order: deliverability first, then list, offer and copy.
How to catch a deliverability problem without opens
The last argument for opens is as an alarm: if they fall to near zero, something broke. It's a weak alarm, because the number already mixes machine opens with missed reads, so a drop can come from a change in the list as easily as from the spam folder. Great copy means nothing if it lands in spam, and you can watch for that without a pixel:
- Run a placement test before you scale and after any change. Send the real email from the real mailbox to test inboxes at Gmail and Outlook, business versions included. The steps are in the cold email deliverability checklist.
- Watch replies per mailbox and per domain. A domain whose replies stop while the others carry on is the alarm opens used to be.
- Track hard bounces and spam rate every week. They move before replies do.
Open rate at 90% and no replies? Don't celebrate the opens. Smartlead says any open rate above 70% "should be treated with suspicion", and that opens above 60% with replies under 3% almost certainly mean inflated opens. Check placement and the list before you touch the copy.
If you keep open tracking on anyway
If a report or a manager still wants opens, limit the damage:
- Use a custom tracking domain. Instantly recommends one before you turn open tracking on, and lemlist says it isolates your reputation from other senders on the shared domain.
- Ignore opens within seconds of sending. Smartlead treats them as likely scanner activity, and lemlist says repeated opens within seconds are often false positives.
- Never let an open trigger a step, a pause or a sales call.
- Decide before launch. In Smartlead the tracking setting can't be changed once a campaign starts sending.
For cold email, my advice stays the same: switch it off and judge the campaign on the people who wrote back.
Key takeaways
- An open only means a tracking image loaded. Apple Mail and scanners load it with nobody reading. Outlook and Gmail can block it when someone does.
- Instantly, Smartlead, lemlist and Salesforge all tie open tracking to deliverability risk in their own help pages.
- Hunter's 2025 data: 4.4% reply rate with tracking on, 7.4% without. Belkins turned tracking off for its 2025 campaigns.
- Measure reply rate, positive reply rate and booked meetings, plus hard bounces and Google spam rate.
- Catch deliverability problems with placement tests and replies per domain, not with opens.
FAQ
Are cold email open rates accurate?
No. Apple Mail can load the tracking pixel without anyone reading, and security scanners open emails to check them. Outlook and Gmail can block the pixel when someone does read. The number is inflated and undercounted at once.
Does open tracking hurt cold email deliverability?
Instantly, Smartlead, lemlist and Salesforge all link it to deliverability risk, and Salesforge says outright that it doesn't recommend it. Hunter's 2025 data showed 4.4% replies with tracking on against 7.4% without, though that doesn't prove the pixel alone caused the gap.
Why is my cold email open rate high but nobody replies?
Usually because machines are opening the emails: Apple Mail Privacy Protection loads content on arrival and corporate security gateways scan messages. Smartlead treats opens above 60% with replies under 3% as almost certainly inflated. Check inbox placement and your list next.
What should I track instead of open rate?
Reply rate, positive reply rate and booked meetings, plus hard bounce rate and the spam rate in Google Postmaster Tools. Use an inbox placement test to see where your emails land.
How does Apple Mail Privacy Protection affect open rates?
When it's on, Apple Mail downloads remote content, including tracking pixels, in the background when an email arrives, whether or not the person engages with it. Each of those emails can show as opened even if nobody read it.