- Published
- Byline
- GoHighLevel Partners Editorial
- Filed under
- GoHighLevel Automation
Outbound clicks are the last thing that happens before the outcome you care about, and the thing analytics is least likely to record correctly.
GA4 track outbound links: what leaves the site and where it goes
An outbound link is any click that takes a visitor to a domain you do not own: a booking tool on another host, an affiliate destination, a client’s main website linked from a campaign page, a payment provider, a phone or email link. Those clicks are usually the last thing that happens before the outcome you care about, and they are the ones analytics is least likely to record correctly by default.
The reason is structural rather than a bug. Analytics runs in the page, and a click that navigates away can unload the page before the request describing the click has been sent. Everything below is about closing that gap and knowing what is still missing afterwards.
What is collected without any work
Modern analytics collects a set of interactions automatically once the tag is present, and outbound clicks are usually among them. That default is genuinely useful and genuinely limited: it records that a click left for another domain, along with the destination URL, but it knows nothing about which element was clicked, where on the page it sat, or what the visitor was doing beforehand.
So the first question is not how to add tracking. It is whether the tag is present and firing on every page that matters. tracking code that is not firing is the most common reason a report is empty, and on a funnel it usually comes down to code placed in the page header existing on one step and not on the next.
Where the default falls short
| Situation | What the default does | What to do instead |
|---|---|---|
| Two links to the same destination on one page | Records both identically | Send your own event with a parameter naming the placement |
| Telephone and email links | Often not counted as outbound at all | Track them as their own event type |
| A link opened in a new tab | Usually recorded | Nothing — this is the reliable case |
| A link that navigates in the same tab | Can be lost if the page unloads first | Ensure the request is sent in a way that survives unload |
| A button wired with a script rather than a link | Not seen, because there is no link to observe | Fire the event from the same script |
| A link inside an embedded frame | Belongs to the other document | Accept the limit, or measure at the destination |
The unload problem, in plain terms
When a click navigates in the same tab, the browser begins tearing the page down immediately. A tracking request that has not finished can be cancelled mid-flight, and the click never appears in the report. This is why outbound numbers are often lower than reality rather than higher, and why the shortfall is worst on slow connections — precisely the visitors you would most want to understand.
Two mitigations are worth knowing. The first is to send the event using a mechanism designed to survive page unload, which every current analytics library supports and most use by default. The second is to open genuinely external destinations in a new tab where that is appropriate for the visitor, which sidesteps the problem entirely because the original page is never unloaded. Do not apply the second one indiscriminately: forcing new tabs everywhere is a worse experience than losing a small share of measurements.
Name the event so the report answers a question
A report full of identical click events tells you a link was used. A report with a parameter naming the placement tells you which link was used, which is the version that changes a decision. Decide a small vocabulary before you start — placement, link text, destination host — and use the same names everywhere, since inconsistent naming is harder to fix later than missing data.
Keep the vocabulary short. Every parameter is something that has to be maintained across every page, and a scheme nobody can remember decays into a mixture of conventions within a quarter. Writing the agreed names into a log kept outside the platform takes a few minutes and stops the drift.
Connecting the click to what happened next
An outbound click is the end of what you can see and the beginning of what you care about. Bridging the two is a matter of carrying identity across the boundary: UTM parameters appended to the destination let the receiving system attribute the visit, and where the destination is a tool you also control, a webhook payload can report the completed action back.
Where the destination is a booking or checkout tool, the honest measure is the completed booking rather than the click, and the click is a diagnostic for the gap between the two. That framing is the same one behind pipeline conversion, and it is what the thank-you step exists to capture on the pages you do own.
What the numbers cannot tell you
Outbound tracking counts clicks, not intentions and not outcomes. A rise in clicks with no rise in bookings can mean the destination page is failing, or that the link is being clicked out of confusion — a phone number tapped by people who could not find the form. Read a change in outbound clicks alongside what happened at the destination before concluding anything.
Be careful with the comparison periods too. Clicks respond to traffic mix, so a week of a paid social funnel produces a different pattern from a week of organic visitors, and comparing them measures the campaign rather than the link. The same discipline that governs declaring a split test winner applies here: a difference is only a finding once it has held across a full cycle.
Consent, privacy and the honest version of the number
Where consent is required before analytics runs, some proportion of visitors will not be measured, and the report describes those who agreed rather than everyone. Treat the numbers as a consistent sample rather than a census: useful for comparing one link against another, unreliable as an absolute count.
Keep the record of consent itself separate from analytics — consent recorded at the form is the durable evidence, not a tag. And check the obvious blockers before investigating anything subtle: a not-secure warning on the destination will suppress clicks regardless of how well the tracking is configured, and a page speed score on the source page decides how many visitors reach the link at all.
A short check that finds most problems
- Load each page and confirm the tag is present, not only on the first step.
- Click each outbound link yourself and confirm an event arrives.
- Check the same links on a phone, where new-tab behaviour differs.
- Confirm telephone and email links are producing events, since they are the ones most often missing.
- Compare a day of outbound clicks against what the destination reports; a large, stable discrepancy points at the unload problem.
Frequently Asked Questions
Are outbound clicks tracked automatically?
Usually yes, once the analytics tag is present, and the automatic event records the destination URL. What it does not record is which of several links was clicked, where it sat on the page, or anything about the visitor’s path to it, so a custom event is needed the moment you want to compare placements.
Why are my outbound click numbers lower than the destination reports?
Because a click that navigates in the same tab can unload the page before the tracking request finishes, cancelling it. The loss is worst on slow connections. Sending the event through a mechanism that survives unload, or opening genuinely external destinations in a new tab, closes most of the gap.
Do telephone and email links count as outbound?
Often not, because they are not links to another web domain and the automatic rules may skip them. On a service business page these are frequently the most valuable clicks on the page, so track them explicitly as their own event rather than assuming they are included.
Can I track clicks on a link inside an embedded frame?
Not from the surrounding page. The frame is a separate document and its clicks belong to whoever controls it. Either measure at the destination, or replace the embed with a link you own that redirects, which brings the click back into a page you can observe.
Should every external link open in a new tab so it tracks reliably?
No. It fixes the measurement and worsens the experience, especially on a phone where tab management is awkward. Reserve new tabs for destinations a visitor is expected to return from, and solve the measurement with an event that survives unload instead.
How do I connect an outbound click to what happened at the destination?
Carry identity across the boundary. Campaign parameters on the destination URL let the receiving system attribute the visit, and where the destination is a tool you also control, a webhook can report the completed action back so the click and the outcome can be matched.
Does a consent banner make the numbers unusable?
It makes them a sample rather than a census. Comparisons between links stay valid because the same proportion of visitors is measured either way, but the absolute counts understate reality. Report them as relative figures and keep the consent record itself outside analytics.
We build your GoHighLevel account for you
CRM, pipelines, funnels, SMS and automation — configured for your business, not left for you to figure out.
- Setup (one-time)$850
- Ongoing (per month)$38.80
No obligation quote · reply within 1 business day
This site is independently operated and is not officially affiliated with, endorsed by, or sponsored by HighLevel LLC. We do not receive referral commissions — this is our own flat-fee implementation service.