Shadow Inbox/blog
Subscribe
← back to indexblog / cold email / cold-email-open-tracking
● Cold Email

Cold Email Open Tracking: Turn It Off or Keep It?

Apple pre-loads your tracking pixel, security bots click your links, and Google's rules mention neither. What to switch off in cold email, and what to watch.

A
ArthurFounder, Shadow Inbox
publishedNov 12, 2026
read6 min
Cold Email Open Tracking: Turn It Off or Keep It?

You send a 30-email batch on Tuesday morning and check the dashboard on Tuesday afternoon. 28 opens. That is a 93% open rate, and your inbox contains zero replies. You feel briefly like a genius and then like someone who was just lied to. T

You send a 30-email batch on Tuesday morning and check the dashboard on Tuesday afternoon. 28 opens. That is a 93% open rate, and your inbox contains zero replies. You feel briefly like a genius and then like someone who was just lied to. The question of whether to leave cold email open tracking on has a short answer and a longer one, and the longer one is mostly about who is actually doing the opening.

Apple's servers open your email, and your prospect never has to

Open tracking works by embedding an invisible image. Hunter's cold email guide describes the mechanism: the pixel has to be fetched from a server, and the server hosting it logs the request as an open. Click tracking works by redirects, with each link sent through a tracking address before it lands on your page.

Hunter also lists the problems on its own page. Some email providers pre-load your message before the recipient reads it, which can inflate the count. A forwarded email fires the pixel again for every later reader. Some clients cache the pixel, so repeat reads vanish.

Apple's own description of Mail Privacy Protection is more direct. On its Mail Privacy Protection page, Apple says senders may otherwise learn "when and how many times you opened their email." With Protect Mail Activity on, Mail "downloads remote content in the background by default — regardless of whether you engage with the email." Apple routes those downloads through two relays so the sender sees no real IP address either. Your pixel fires, your tool records an open, and the prospect has not looked at the email.

1invisible image is all an open count rests on
0mentions of tracking pixels or click tracking in Google's email sender guidelines, as of 12 Nov 2026
1link per email is lemlist's maximum, and ideally none in the first email

We ran into this in our cold email A/B testing post: once the open number stops meaning anything, every subject-line test built on it measures noise.

Click tracking adds a bot to the funnel

Clicks look sturdier than opens, but they have the same flaw. lemlist's help center says that security tools at Gmail, Outlook and corporate gateways scan incoming mail and follow the links in it, which produces clicks that no person made. It gives an example of a campaign showing a 40% click rate with no human engagement behind it. That is a vendor's example, not a measured study, so read it as a mechanism and not as a statistic.

The mechanism is enough. If a scanner follows your tracked link first, your tool cannot tell it from a buyer, and you follow up on the scanner.

The temptation is to say Gmail punishes tracking pixels. We read Google's email sender guidelines looking for exactly that, and the words "tracking" and "redirect" do not appear. The nearest rule is about links: "Web links in the message body should be visible and easy to understand. Recipients should know what to expect when they click a link."

A link that bounces through track.vendor.com before it reaches your site is not obviously a violation of that sentence, and Google does not call it one. We found no primary source saying a pixel lowers inbox placement. What exists are vendor pages making the argument, and a vendor has a reason to give you a tidy rule.

The vendors that sell tracking now say to turn it off

The sending tools themselves lean toward off. lemlist's help article says open and click tracking both need a verified custom tracking domain, advises one link per email at most (ideally none in the first), and recommends measuring reply rate. Instantly's write-up on tracking pixels argues that pixels and shared tracking domains can hurt inbox placement, and says to disable open tracking on active campaigns while the DNS for a custom tracking domain propagates.

Notice what these pages do not contain: the test. Neither lemlist's article nor Instantly's shows a split of tracked and untracked sends with deliverability numbers. Their reasoning is plausible (redirects resemble phishing, a shared domain shares its reputation with every sender on it), and we still file it as plausible, not proven. A custom tracking domain is the fix they all point to, and it is also the same logic we used in our separate domain for cold email post: keep the thing that can get burned away from the thing you cannot afford to lose.

An open count tells you whether Apple's servers are awake. Nobody has found a way to read a prospect's mind from a pixel.

What we would do with a small cold email campaign

This section is our opinion, drawn from the sources above and not from a split test of our own.

Turn open tracking off. The number is contaminated by pre-loading, you cannot clean it, and the one decision people make from it (rewrite the subject line) is decided on bad data. If you cannot bring yourself to, keep it on only behind a custom tracking domain you own, and read the count as "the email was delivered somewhere."

Turn click tracking off for the first email, and put no link in it. Ask a question instead. If the prospect replies, send the link in a plain form in your reply, where a human will click it and the scanner problem is gone. That is lemlist's permission-based approach, and it matches how a real conversation goes.

Then judge the campaign on replies. Our reply rate math post explains why 20 messages that reach a human beat 500 that reach a pixel, and the cold email playbook for 2026 has the volume rules to go with it. If deliverability is your worry, the numbers worth watching are the ones the providers do publish, such as the bounce thresholds and the spam rate in Postmaster Tools.

A public reply needs no pixel

Some outreach has an honest engagement signal by construction. When someone posts "anyone recommend a tool for X" on Reddit or Hacker News and you answer in the thread, nothing is pre-loaded and nothing is scanned.

Shadow Inbox is our product. It does not send email and it has nothing to say about your tracking settings. What it does is watch Reddit and Hacker News for people who asked for what you sell, so your first contact is a reply to a question they put in writing. The signal that follows (a reply, an upvote, a DM that they started) is human by definition, which is the one thing an open rate could never promise.

● FAQ

Should I turn off open tracking in cold email?
For a small founder-run campaign, yes. The number it produces is inflated by mail clients that load images before a person reads anything, and you cannot tell the real opens from the machine ones. The only metric that proves a human read your email is a reply.
Does open tracking hurt cold email deliverability?
No mailbox provider says so in the pages we checked. Sending-tool vendors argue that the tracking pixel, and especially a shared tracking domain, can attract spam filters. That is a claim without published test data, but turning the pixel off costs you a number you could not trust anyway.
What is a custom tracking domain?
A subdomain you own, pointed at your sending tool with a CNAME record, so tracked links and pixels use your domain and not the tool's shared one. Hunter, Instantly and lemlist all describe it as the mitigation when you keep tracking on. lemlist will not send with open or click tracking on until you have a verified one.
Can I tell if a prospect actually read my cold email?
Not from an open count. Apple's Mail privacy feature loads remote content in the background whether or not the person engages, and other clients and security tools pre-load too. A reply, or a visit to a page you control after they asked for it, is the evidence that holds up.
Is click tracking in cold email worth keeping?
We would turn it off for the first email and send plain links only after someone replies. Link scanners can register clicks no human made, and a redirect through a tracking domain adds one more thing a filter can dislike. One plain link or none at all is the safer default.
— share
— keep reading

Three more from the log.