Use UTM parameters on plain, unrewritten links. You keep campaign level attribution in your analytics, the link points where it says it points, and nothing redirects. If you need per-recipient click data, use a custom tracking domain rather than your vendor’s shared one, and accept the residual cost.
Published by emailcampaign.ai, which runs cold sending infrastructure and also offers tracking. The recommendation here argues for using less of that feature, not more.
What link tracking does to your message
When tracking is on, every URL is replaced. A link that reads yourcompany.com/pricing becomes something like track.vendor.com/c/a8f3k2, which redirects on click.
Three things follow, and they compound on a young domain:
| Effect | Why it matters |
|---|---|
| The visible URL no longer matches the destination | This is the defining structure of a phishing link. Filters score it, and so do recipients who hover. |
| The redirect host does not match your sending domain | A domain mismatch between the From address and the links inside is a pattern associated with compromised or bulk senders. |
| On a shared host, you inherit others' reputation | If the vendor's tracking domain lands on a list such as SURBL because of another customer, every message you send carries a listed domain in its body. |
The third is the one you cannot see and cannot control, which is why it is the most worth avoiding.
The approach that keeps attribution
Put the campaign information in the URL itself rather than in a redirect:
https://yourcompany.com/pricing?utm_source=outbound&utm_medium=email&utm_campaign=q4-agencies
The link points at your own domain, matches what the recipient sees, and requires no external host. Your analytics attributes the session, the signup and the revenue to that campaign.
What you give up is knowing which individual recipient clicked. For cold email that is usually an acceptable trade, because the decision you make from click data is rarely different from the decision you make from reply data.
When per-recipient clicks genuinely matter
There are real cases: a long sequence where you want to prioritise follow-up by engagement, or a sales team triaging who to call first. If that is you, the compromise is a custom tracking domain, which is a single CNAME on a subdomain of your sending domain.
It fixes two of the three problems above. The redirect host becomes yours, so it matches the sender and you no longer inherit anyone else’s reputation. It does not fix the first: a hovering recipient still sees a URL that is not the destination.
If you do this, keep the tracking subdomain on the sending domain, never on your main company domain, and never proxy the record.
Open tracking is a separate and worse trade
Open tracking embeds a one by one remote image in every message. That is a pattern filters have scored since long before any of the current tools existed, and it is paid on every send.
What you get for it is unreliable. Apple Mail Privacy Protection has pre-loaded images on behalf of recipients since 2021, registering opens with no human involved. A meaningful share of any B2B list reads mail in Apple Mail, so the open rate is partly a measure of Apple’s infrastructure. Teams still optimising subject lines against that number are tuning against noise.
A setup that works
- Open tracking off. Removes a remote image from every message and a metric you cannot trust.
- Link tracking off. Links point where they say they point.
- UTM parameters on the destination URL. Campaign attribution preserved in analytics.
- One link per message where possible. Fewer links reads as more personal and gives a filter less to evaluate.
- Measure replies and meetings. The outcomes you actually care about, and the ones no proxy can inflate.
If you need per-recipient clicks, add a custom tracking domain and keep opens off. That is the middle position, and it is a reasonable one.
Keep this in proportion
Tracking is a second-order factor. If mail is landing in spam, the first four things to check are whether authentication resolves, the hard bounce rate, the complaint rate against Google’s published 0.1 percent threshold, and sends per mailbox per day. Turning off tracking will not rescue a campaign failing on any of those, and checking authentication takes about thirty seconds.
Questions, answered straight
- Does link tracking hurt deliverability?
- It can. Tracking rewrites every link to point at a redirect host, so the visible URL no longer matches the destination or the sending domain. That mismatch is a phishing pattern, and on a shared tracking host you also inherit the reputation of every other customer using it.
- How do I track clicks without rewriting links?
- Put UTM parameters on the real destination URL. The link still points at your own site, your analytics records the campaign, and nothing is redirected. You lose per-recipient click attribution and keep campaign level attribution, which for cold email is usually the trade worth making.
- Is a custom tracking domain enough?
- It removes the shared-reputation risk and makes the redirect host match your sending domain, which is a real improvement. It does not remove the redirect itself, so a recipient hovering still sees a URL that is not the destination.
- Should I turn off open tracking?
- For cold outbound, usually yes. Open tracking embeds a remote image in every message, a pattern filters have scored for two decades, and since Apple Mail Privacy Protection began pre-loading images in 2021 the resulting numbers count opens that never happened.
- How do I measure cold email without tracking?
- Replies and meetings booked, which are the outcomes you are paid for and cannot be inflated by a proxy fetch. Add UTM parameters if you want to see which campaign produced site visits and signups.
Published 20 September 2026. Updated 20 September 2026. Written by the team that runs the infrastructure; numbers come from the platform's own provisioning and sending, and from the providers' published documentation at the time of writing.