Rémi KokabiNovember 4, 20268 min read

Open tracking in cold email

Open tracking embeds an invisible pixel in every HTML email you send. When the pixel loads, your platform records an open. The problem is that the pixel fires for bots, security scanners, and privacy proxies, and fails to fire when real humans read your email with images blocked. The number on your dashboard is unreliable in both directions.

This article covers how open tracking works, what it costs your deliverability and reply rates, and why turning off open tracking is the right call for cold email. We will compare campaigns with open tracking on versus off, show the reply rate and bounce rate differences, and explain which metrics to report on instead.

How open tracking works

Every HTML email with open tracking contains a 1x1 invisible pixel image with a unique URL. When the recipient's email client downloads that image, the download fires a signal back to the sender's platform. That signal is the entire basis of open rate reporting. If the image never loads, no open is recorded. If something other than a human loads it, a false open is recorded.

The pixel URL and any tracked link URLs route through a tracking domain. When multiple senders share the same tracking domain, spam filters see that domain across thousands of unrelated campaigns and treat it as a risk signal. That shared reputation drags your deliverability down. If you keep open or click tracking on, set up a custom tracking domain tied to your own sending infrastructure.

Why open rates get inflated

Three sources generate opens that no human made. Each one is common enough to move your numbers.

Apple Mail Privacy Protection

Apple Mail Privacy Protection routes email content through Apple's proxy servers. Those servers preload all images, including the tracking pixel, before the user has done anything. The email registers as opened whether or not it was read. The sender loses visibility into timing, location, and device.

Gmail image proxy and caching

Gmail loads remote images through Google's image proxy and caches them. The tracking server sees a request from Google infrastructure, not the recipient's device. Google may prefetch the image before the user opens the email, triggering a false open. And because Gmail reuses the cached copy on repeat views instead of making another request, the same mechanism that inflates some opens can hide others.

Security scanners and bots

Enterprise security scanners like Microsoft Defender, Proofpoint, and Mimecast load every image and click every link in your emails before delivery. You will see opens fire within seconds of send and clicks on every link in the email, all in the same few seconds. No human reads that fast. In cold outreach to mid-market and enterprise accounts, bot activity is the norm.

Why open rates get undercounted

Opens are also missed when a real person reads your email. Some desktop clients block remote images by default. Outlook on Windows has done this for years. If images never load, the tracking pixel never fires, and the open is never recorded.

Plain-text email views skip image loading entirely. So do preview panes that render a snippet of the message without downloading remote content. A prospect can read your entire email, think about it, and reply the next day, yet your dashboard shows them as "never opened."

What median open rates look like

Even with all the noise from bots, proxies, and image blocking, here is what the median campaign open rate looks like across markets.

Market

Median campaign open rate

Europe

49.1%

US

42.9%

Source: lemlist internal data from its users.

These numbers mix real reads with bot opens and privacy proxy fires. They are directional, not precise. Use them to spot trends against your own history, not as a target.

Does open tracking hurt deliverability

The tracking pixel adds an invisible image load to every email you send. Some spam filters treat tracking pixels as a signal that the email is bulk or automated. The pixel also increases the HTML weight and adds a redirect through your tracking domain, which gives filters more to inspect.

Campaigns with open tracking turned off have lower bounce rates than campaigns with it on, in both Europe and the US.

Market

Bounce rate (open tracking off)

Bounce rate (open tracking on)

Europe

6.46%

7.57%

US

5.59%

7.40%

Source: lemlist internal data from its users.

Lower bounces mean more of your emails reach the inbox. Removing the pixel removes one element filters can flag, and the deliverability numbers reflect that. If you want to keep your emails out of the Gmail promotions tab, turning off open tracking helps.

Turn off open tracking, get more replies

Campaigns with open tracking turned off reply at higher rates than campaigns with it on. Here is the breakdown by market.

Market

Email reply rate (open tracking off)

Email reply rate (open tracking on)

Europe

4.02%

2.67%

US

2.47%

1.70%

Source: lemlist internal data from its users.

Campaigns that turn off open tracking in Europe reply at 4.02%. Campaigns that keep it on reply at 2.67%. In the US, the gap is 2.47% vs 1.70%. The pattern holds across both markets. Better deliverability means more emails in the primary inbox, which means more replies.

What to track instead of open rates

If open rates are unreliable, what do you report on? The table below covers the metrics that tie to real human behavior, along with where each one can still break.

Metric

What it measures

Where it breaks

Reply rate

Percentage of contacted leads who respond

Includes negative replies (objections, unsubscribes sent as replies)

Positive reply rate

Replies marked as interested

Requires consistent tagging by reps or AI

Meetings booked

Conversations that convert to a scheduled call

Attribution can be unclear in multichannel sequences

Bounce rate

Invalid or undeliverable addresses

Does not distinguish hard bounces from temporary failures without classification

Unsubscribe rate

Leads who opt out

Low volume can make the rate noisy on small campaigns

LinkedIn acceptance rate

Connection requests accepted

Acceptance without a reply is passive engagement

LinkedIn reply rate

Messages replied to on LinkedIn

Replies can be short rejections

For reference, across all markets email reply rate averages 2.72%, reply rate across all channels averages 4.07%, LinkedIn reply rate averages 6.6%, and call connect rate averages 13.8%.

Source: lemlist internal data from its users.

Reply rate and positive reply rate are the two numbers that tell you whether your outreach is working. Meetings booked tells you whether it is working well enough. Everything else is a health check.

When a falling open rate still matters

Opens have one useful job left. A sudden drop across all campaigns at the same time, not one sequence, can signal that your emails are landing in spam or that your domain has a reputation problem. That kind of drop is a deliverability alarm worth investigating right away.

The key word is "sudden" and "across all campaigns." A single sequence with lower opens could mean a bad list, a subject line that did not land, or a higher share of image-blocking clients. But when every campaign drops at once, check your inbox placement and domain health before you touch your copy.

Compare against your own history. Public benchmarks are averages across industries, list sizes, and sending patterns that look nothing like yours. They tell you what other people's numbers are, not whether yours are healthy.

How to split-test without open rates

Most teams test subject lines by comparing open rates across variants. If opens are inflated by bots and privacy proxies, the "winning" subject line is the one that happened to land in more Apple Mail inboxes or trigger more security scanners. You are optimizing for machines.

Judge every test on reply rate and positive replies. That includes subject line tests. A subject line that gets more replies is doing its real job: getting the right people to read and respond. One that gets more opens but fewer replies is noise.

Run enough volume per variant to trust the result. Splitting 50 leads across two subject lines tells you nothing. Each variant needs enough sends that a few extra replies are not random variation.

Click tracking works the same way as open tracking: it rewrites your links through a redirect. That redirect gives spam filters another signal to inspect. Keep click tracking only when a specific link matters to the campaign, like a case study or a booking page where you need to know whether leads engaged with it. If your email has no links, turn click tracking off too.

For a deeper look at how links affect deliverability, see our guide on links in cold email.

How to set this up in lemlist

You can disable open and click tracking per campaign in campaign settings. This is a toggle, not a global setting, so you can turn it off for cold sequences and keep it on for warmer campaigns where you want the signal.

If you keep tracking on, lemlist requires a custom tracking domain. No shared tracking domains.

Split tests run on a single step or across the whole sequence. Reply rate decides the winner. You are never forced to judge a test on opens.

When replies come in, mark them as Interested or Not interested. You can do this manually or turn on AI auto-marking, which classifies replies for you. That turns "positive reply rate" into a real, countable metric you can report on and optimize against. Campaigns stop for a lead when they reply, so no one gets a follow-up after they have already responded.

The bottom line on open tracking in cold email

Open rates are unreliable because the tracking pixel fires for machines and fails for real readers. Turning off open tracking improves deliverability and reply rates. The metrics that matter for cold outreach are replies, positive replies, and meetings booked. Build your campaigns around those signals, and you will spend less time interpreting bad data and more time in real conversations.

Build your first sequence in lemlist.

Build cold email campaigns that work

Find verified leads and launch personalized cold email campaigns.

Frequently asked questions

Not reliably. Privacy features and security scanners inflate them, and image blocking hides real reads. Use them only to spot sudden drops that signal a deliverability problem.

Yes. Campaigns with open tracking on have higher bounce rates than campaigns with it off. In Europe, bounce rates are 7.57% with tracking on vs 6.46% with tracking off. In the US, 7.40% vs 5.59%.

For cold campaigns where the goal is a reply, yes. Campaigns with open tracking off reply at 4.02% in Europe (vs 2.67% with it on) and 2.47% in the US (vs 1.70% with it on). You lose a vanity metric and gain better deliverability and more replies.

Track reply rate, positive reply rate, and meetings booked. Add bounce and unsubscribe rates to watch list quality and deliverability.
Rémi Kokabi

Rémi Kokabi

Co-founder @GTM Club

Share this post