Should You Track Email Opens on Cold Email? (We Turned It Off)

Should you track email opens on cold email? Usually no. The tracking pixel can drag down your deliverability, and modern open data is unreliable because Apple and Gmail inflate it. The smarter move is to turn open tracking off on cold sends, then turn it back on only for replies, where the number is honest and actually tells you something.

We run a cold email agency, and this is the exact call we made for every client we send for. Open tracking is off by default on the cold send, and on for replies. Here is the reasoning, and the one place tracking still earns its keep.

What is open tracking, and how does it work?

Open tracking works by dropping a tiny invisible image into your email, usually a 1x1 transparent pixel served from a tracking domain. When the recipient opens the email, their mail client loads that image, the tracking server records the request, and you get an "open." It is the same mechanic almost every cold email tool uses, and most of them turn it on by default.

That sounds harmless. It is not. The pixel introduces two problems that quietly work against the only thing that matters in cold email: landing in the inbox and getting a reply.

Does open tracking hurt cold email deliverability?

Yes, it can. An open pixel is a remote image loaded from a tracking domain on every single email you send. Spam filters have seen that pattern for years, and a fresh sending domain that loads an external tracking image on every message looks more like bulk marketing than a one-to-one note from a founder.

So you are making a trade most people never think about. You spend a little deliverability to buy an "open rate" number. In cold email, deliverability is the whole game. If your message lands in spam, nothing else you do matters. Trading inbox placement for a vanity metric is a bad deal, and on cold outreach it is almost always the wrong one. Our entire approach to deliverability is built around protecting that inbox placement, which is why the pixel was one of the first things we cut.

Why are open rates unreliable in 2026?

Here is the part that makes the trade even worse: the number you buy with your deliverability is mostly fiction.

• Apple Mail Privacy Protection pre-loads images before the human ever opens the message. That logs an "open" the prospect never made. A large share of inboxes now run through Apple Mail, so a big chunk of your open rate is a machine in an Apple data center, not a buyer.

• Gmail proxies images through its own servers, which both masks the real open and adds another layer of noise.

• Security scanners and bots open links and images automatically to check for threats, firing the pixel before a person sees a thing.

Put it together and your cold email open rate is a blend of real humans, privacy prefetch, proxies, and bots, with no clean way to separate them. You are degrading your deliverability to collect a metric you cannot trust. That is why we stopped tracking opens on cold sends entirely. We would rather land in the inbox than know a robot at Apple loaded a pixel at 3am.

So why track opens at all? The one exception: replies

Killing open tracking on cold sends is easy to defend. But there is one moment where tracking flips from liability to genuinely useful, and that is the reply.

Once a prospect writes back, everything changes. It is a real conversation now, not a cold blast into a stranger's inbox. The deliverability stakes are different, and more importantly, you suddenly have a real question worth answering: did they actually read my follow-up, or am I being ghosted?

That is why, when our team sends a reply inside the master inbox, we drop a tracking pixel on that reply. Not on the cold send. On the warm, in-conversation reply, where knowing whether your message got read is the difference between following up smartly and following up blind. If you run outbound yourself, this is the kind of decision our founder-led outbound playbook gets right by default.

Open tracking on a cold send vs. on a reply

Question

Cold send

Reply

Is tracking on by default?

No, we turn it off

Yes, we turn it on

Deliverability risk

High, pixel on every send looks like bulk mail

Low, it is a one-to-one conversation

Is the data useful?

No, inflated by prefetch and bots

Yes, and we filter the fake opens

What it tells you

Almost nothing trustworthy

Whether a warm lead read you or ghosted you

How do you track reply opens without hurting your brand or lying with the data?

If you are going to track reply opens, two things have to be true: it cannot put your brand at risk, and the number has to be honest. Here is how we built both.

The pixel is first-party to the sender, and your brand never touches it

The tracking image is served from the sender's own subdomain, not from some shared vendor domain that thousands of other senders also use. That is the strongest possible reputation position for a pixel. Behind the scenes, that subdomain quietly points at a single neutral infrastructure domain we own and use for nothing else, locked down so it can never be spoofed and reads as pure infrastructure. Our brand domain never appears in a cold email tracking graph. If that tracking origin ever lands on a blocklist, we swap it without any client's reputation moving with it.

The counter throws out the fake opens

This is the part that makes the reply-open number worth looking at. We built the open counter to be honest about the same prefetch and proxy noise that ruins cold-send open rates:

• It filters out Apple Mail Privacy Protection prefetches, but only when both the browser signature and the network origin line up as Apple. The user agent alone is not enough to call it fake, so we do not throw out real opens by accident.

• It keeps Gmail's proxied opens, because Gmail proxies real human opens too. Filtering those would throw away genuine signal.

• It filters known crawlers and automated scanners out of the human count, while still storing the raw hit so nothing is silently lost.

The result is a reply-open count you can actually trust. When it says a prospect opened your reply three times and did not write back, that is a real human who read you and went quiet, not a privacy bot.

What does "they opened it three times and didn't reply" actually tell you?

This is the fun part, and the practical one. An honest reply-open count turns a silent inbox into signal.

If a warm prospect opened your reply three times late at night and still has not responded, they are not gone. They are interested and stalling, weighing it, or waiting for a reason to move. That is a completely different follow-up than someone who never opened it at all. One needs a nudge or a new angle. The other needs a different channel entirely.

You finally get to see the thing everyone secretly wants to see in a sales conversation: who is ghosting you, with receipts. And because the data is clean, you can actually act on it instead of guessing.

The takeaway: tracking is a tool, not a default

Open tracking is not good or bad on its own. It is a tool, and the only question that matters is where you point it. On a cold send it costs you the inbox and hands you a number you cannot trust. On a warm reply it tells you who is actually interested and who is just being polite.

So we turned it off where it hurts and on where it helps. That single distinction, applied across every campaign, is a small example of how we think about the whole stack. If you want to see the rest of it, you can look through how Ken works, check the pricing, or just book a founder call and we will walk you through the backend.

Frequently asked questions

Should I turn off open tracking on my cold emails?

For cold outreach, yes, in most cases. The pixel can hurt deliverability and the open data is unreliable thanks to Apple Mail Privacy Protection, Gmail's image proxy, and security bots. You are usually trading real inbox placement for a number you cannot trust.

Will disabling open tracking improve my deliverability?

It removes one risk signal: a remote tracking image loaded on every message. It is not a magic fix on its own, but on cold sends it is a sensible default, especially on newer domains. Deliverability is the sum of many decisions, from infrastructure to copy, and this is one of the easy ones to get right.

Are email open rates accurate?

Not really, not anymore. Apple Mail pre-loads images and logs opens no human made, Gmail proxies images, and bots fire the pixel automatically. Your open rate is a blend of real opens and noise with no clean way to separate them, which is why reply rate and positive reply rate are far better signals.

Why track opens on replies but not on cold sends?

Because the context is different. A cold send goes to a stranger, where the pixel costs deliverability and the data is junk. A reply goes inside a live conversation, where the stakes are lower and knowing whether your message was read is genuinely useful for how you follow up.

How can I tell if a prospect is ghosting me?

If you track opens on your replies with an honest counter, you can see whether a warm prospect actually read your last message. Someone who opened it several times and went quiet is interested and stalling. Someone who never opened it at all is a different problem. Both call for a different next move.