Should you turn off open tracking in cold email?
Yes, at least on the first email. What the reply data shows, why opens mislead, the rule in France, and what to measure once the open rate is gone.
Yes. Turn off open tracking in cold email, at least on the first message, and switch off link tracking with it. In Hunter's 2026 report, sequences without tracking got a 7.4% reply rate against 4.4% with it. The number you'd give up is barely readable anyway: Litmus says Apple's Mail Privacy Protection now affects roughly 55–60% of all email opens. Keep tracking only for a short, deliberate subject line test. Below is how strong the evidence really is, and what to watch once the open rate is gone.
What the data says, and what it doesn't
- reply rate in sequences without tracking
- 7.4%
- reply rate in sequences with tracking
- 4.4%
- emails in the full report; the size of the tracking split isn't given
- 31M
This is the best public number I could find, and it has a hole in it. Hunter compares two groups of its users and doesn't describe any controls. Teams that switch tracking off may simply be more careful senders with better lists, and some of that 3-point gap would then belong to the list. So don't expect your replies to jump 68% the day you flip the setting.
I'd still make the change, because the two sides of the bet are lopsided. If the effect is real, you get more replies. If it's zero, you've lost a metric you couldn't trust. You'll see bigger figures quoted around the web as well, and I left them out because I couldn't trace them to a first-hand dataset.
Why the open rate stopped meaning much
An open was never a person reading your email. It was a server logging that an invisible image had been requested. That used to be a fair proxy, and it isn't anymore.
Apple's own description of Mail Privacy Protection says it downloads remote content in the background by default, "regardless of whether you engage with the email." Litmus's September 2026 count, from over 1.5 billion opens, attributes 76.56% to Apple clients and Apple's privacy feature combined. That count covers email in general, so your B2B list will look different, likely with more Outlook and Google Workspace. Those inboxes bring their own noise.
Clicks aren't clean either. Microsoft's Safe Links scans URLs before a message is delivered and verifies them again when someone clicks, and Smartlead notes that security scanners pre-click links. A click from a corporate domain seconds after delivery may well be a machine.
The real damage shows up when a team acts on these signals:
Act on this
- A reply, even a “no”
- A meeting on the calendar
- A bounce or a spam complaint
Ignore this
- Three opens in one minute
- A click seconds after delivery
- A “hot lead” alert based on opens
Never write "saw you opened my email." Quite possibly nobody did, and if someone did, you've just told a stranger you were watching.
What tracking adds to your email
Open tracking puts a tiny remote image in the message, with an address unique to the recipient. Link tracking rewrites each link so it goes through a redirect first. By default, both live on a tracking domain that your tool shares across its customers.
Hunter's guide gives the two reasons this matters: spam filters check the domain hosting the pixel, and if other users of the tool ruin that domain's reputation, you pay for it. A one-to-one note from a colleague doesn't contain a remote image on a stranger's domain, so tracked mail also looks less like the thing you're pretending to send.
I haven't found a mailbox provider stating how much a pixel weighs in filtering. This part is mechanism and vendor observation, not proof. A custom tracking domain reduces the shared-reputation risk, but Instantly, which sells them, still advises keeping the pixel off on the first step of every sequence.
If you write to people in France
This one isn't about reply rates. France's data protection regulator, the CNIL, published its recommendation on email pixels on April 14, 2026. According to BCLP's summary, the opt-out regime for B2B marketing survives, but a pixel inside that email needs consent. The exceptions cover requested emails and don't extend to promotional messages. Addresses collected before publication got a transition: tracking could continue if recipients were informed by July 14, 2026.
You can't ask a stranger for consent before your first email, so for cold outreach to French recipients the practical answer is no pixel. I'm not a lawyer, and I read the law firms' summaries of this rather than relying on my own reading of the French text. If France matters to your pipeline, ask counsel, including whether it reaches you as a sender based abroad.
What to measure instead
| Metric | What it tells you | Where to get it |
|---|---|---|
| Reply rate | Whether people are reading and reacting at all | Your sending tool |
| Positive replies | Whether the list and the offer are right | Tag replies by hand or with your tool |
| Meetings booked | The only number your boss cares about | Calendar or CRM |
| Bounce rate | List quality | Your sending tool |
| Spam rate | Whether recipients are complaining | Google Postmaster Tools |
For scale, Instantly reports an average reply rate of 3.43% and treats 2% or more as a sign that your mail is getting through. Our reply rate breakdown goes further on what counts as good.
The usual objection is that without opens you can't tell a delivery problem from a copy problem. You can, and in a better order. If replies fall below 2%, check bounces first (over 2% is a list problem). Next, check your Gmail spam rate. Then send to a few seed mailboxes you own across Gmail and Outlook and look at which folder the message lands in. If all three are clean, the mail is arriving and people don't want what it says. Change the list or the offer.
When tracking is worth keeping
One case holds up. Hunter itself concedes that opens help test subject lines, so if you're choosing between two subject lines, track a small test batch of the same audience, sent from the same mailboxes on the same days. Compare the two arms with each other and ignore the absolute numbers, since the Apple noise is roughly equal on both sides. Then turn tracking off again.
Links are easier. If you need to know whether anyone visited your page, send people to a plain URL on your own domain with a campaign parameter, and read it in your site analytics. You get the visit without a redirect through someone else's domain. Better still, leave links out of the first email altogether and ask a question they can answer in a line.
How to make the switch
Turning off tracking without losing visibility
- Disable open and click trackingDo it at campaign level in your sending tool and check that the setting applied to every active sequence.
- Clean up the first emailSend plain text with no links, images or signature banners, and end on a question that takes one line to answer.
- Remove open-based automationDelete every rule that branches, alerts or re-sends on an open. Base follow-ups on time passed and no reply.
- Run a split testSplit one list evenly, using the same copy and mailboxes, with tracking on one side. Compare replies, not opens.
- Move your weekly reportReplace open rate with reply rate, positive replies, meetings, bounces and spam rate.
On the test size: to tell a 3% reply rate from 5% with the usual statistical confidence (95% confidence, 80% power), you need about 1,500 people in each arm. That's my own calculation from the standard two-proportion formula. With a smaller list you won't get a clean answer, so skip the test and go with the default.
Where this fits with follow-ups and automation
Dropping opens means your follow-up logic runs on one simple fact: did this person reply? That's also the sturdiest trigger you can use. Stop on a reply, send a few follow-ups spaced out, then let it go.
When Nitrino runs an email campaign through a sending provider, it switches off the open pixel and link rewriting and takes a person out of the sequence once they reply. The agent answers that reply in the voice of the account's owner and brings you in when a decision is yours to make. It's built for teams that sell through conversations, where a reply is the signal that matters.
If you do one thing after reading this, open your active sequences and check the tracking setting on step one. A setting left on in an old template is easy to miss.
