- Redirects are permitted. What is checked is whether the chain converges on the same endpoint for every visitor, and whether it is short and stable enough for every hop to open within a checking fetch's timeout budget.
- Put the page people actually land on in the final URL, put tracking hops in the tracking template and route them back with {lpurl}, and put the parameters the landing page needs in the final URL suffix.
- Under parallel tracking the visitor goes straight to the final URL while the tracking URL is requested separately in the background. Setups that relied on parameters or cookies added along the redirect lose data quietly.
- A final URL that redirects to another domain is the classic destination mismatch. A chain whose endpoint depends on who is asking falls under the circumventing systems policy.
- Before launch, trace every hop: with no parameters, with test parameters, and from other networks and devices, confirming a consistent endpoint, intact parameters and certificate coverage for each hostname.
Redirects are allowed; the path has rules
An ordinary ad click often passes through several steps: a third-party measurement service records the click, http is upgraded to https, the bare domain moves to www, and a language version is selected. These are everyday components and policy has no objection to their existence. What policy cares about is whether the destination the ad declares is the place the visitor actually reaches, and whether the platform can open that destination reliably when it checks it.
Most redirect-related disapprovals do not come from anything deliberate. They come from URL fields filled in the wrong place, a hop that drops parameters, or a certificate that does not cover an intermediate hostname. Understanding how the fields divide the work is more useful than memorising any single prohibition.
Four fields, one destination
Google Ads separates where the visitor goes from how the visit is recorded, using different fields. Each can be set at account, campaign, ad group or ad level, and the lower level takes precedence.
| Field | What it controls | Common way it goes wrong |
|---|---|---|
| Final URL | The page the visitor lands on and the destination the platform checks; the display domain is taken from it | Filled with an intermediary link that redirects to another domain |
| Tracking template | The URL used to record the click, usually starting on a measurement domain and routing back to the final URL through {lpurl} | Missing {lpurl}, or hard-coding a page different from the final URL |
| Final URL suffix | Parameters appended to the landing page URL, such as UTM tags, which arrive with the visitor directly | Parameters still appended inside a tracking redirect, which under parallel tracking never reach the landing page |
| Custom parameters | Variables referenced as {_name} in templates or suffixes, so one template can be reused across levels | A lower level defines a parameter with the same name and silently overrides the higher value |
One detail is often missed: depending on its position in the template, the value inserted by {lpurl} may be URL-encoded. The measurement service has to decode it correctly before redirecting, otherwise the endpoint becomes a malformed address that occasionally opens in a browser and fails when checked.
What parallel tracking changed
Under sequential tracking, a click passed through each tracking redirect in turn before reaching the landing page, and any hop could append parameters or set cookies. Parallel tracking separates the two: the visitor loads the final URL directly, together with the suffix and the auto-tagged gclid, while the browser requests the tracking URL separately in the background.
The result is a faster landing, with three consequences to adjust for:
- Parameters appended during the redirect no longer appear on the landing page. Anything the landing page or analytics needs has to move into the final URL suffix.
- The measurement service must handle background hits. It is no longer part of the visitor's path and cannot assume a real page load follows each request.
- Redirects on the landing domain are still in the visitor's path. If the server rebuilds the URL during an http-to-https or www redirect and drops the query string, gclid and the suffix disappear together and conversion attribution breaks.
Assuming a sudden drop in conversions means traffic quality has changed. The first thing to check is whether parameters still reach the landing page — a single change to a server redirect rule can make attribution disappear on paper while the traffic itself is unchanged. Investigating invalid clicks is a separate path, covered in invalid clicks in Google Ads.
Redirect patterns and how they are read
| Pattern | How it is usually read | What to watch |
|---|---|---|
| Tracking template through a measurement domain, back to the final URL via {lpurl} | Standard practice | Correct decoding and a fast measurement response |
| Canonical redirects within the domain (https, www, trailing slash) | Normal | Keep the query string; ideally enter the canonical version as the final URL |
| Final URL redirecting to a different domain | Destination mismatch | Enter the real landing page as the final URL |
| Language or region version selection | Acceptable when the result is consistent for comparable visitors | It must still open when checked from outside the targeted region, not end on a blank or error page |
| Endpoint depends on who is asking (whether the request looks like a check, which network it comes from) | Circumventing systems territory, with account-level consequences | Serve the same endpoint to everyone; keep personalisation within the same URL |
| Chains of three or more hops, or link shorteners | Prone to destination not working | Compress to one or two hops and confirm every hostname is covered by a certificate |
The second-to-last row deserves a note. An endpoint that depends on the requester has the same character however it is built — a 302, server-side rendering or a proxy — and that is what cloaking means in an advertising context. How platforms identify the difference from the outside is covered in how ad platforms detect cloaking, and the policy itself in the circumventing systems policy. The trade-offs between delivery methods are covered in redirects, static pages and proxies.
Disapprovals that trace back to the path
Destination mismatch
The domain the ad declares does not match the domain the visitor reaches. The usual causes are a final URL that itself redirects elsewhere, or a tracking template that does not return to the final URL. The fix is almost always the same: enter the real landing page as the final URL and move the intermediary back into the tracking template.
Destination not working
The page returns an error, times out, has an invalid certificate, or fails to open in certain regions or on certain devices when checked. Longer chains have more places to fail, and it opening for you is not evidence, because checking requests come from different networks on a tighter timeout budget.
Circumventing systems
When the endpoint of a chain changes with the requester, the issue is no longer a disapproved ad but action at the account level. What follows in those cases is covered in why Google Ads accounts get suspended, and the organic search counterpart in doorway pages and sneaky redirects.
Testing the chain before launch
- Use the test function in URL options to confirm the tracking template resolves to the final URL. It is the minimum, not the whole test.
- Open the final URL with no parameters and confirm it returns 200 on its own rather than depending on parameters to show content.
- Open it again with a test gclid and the suffix, inspect the response headers hop by hop, and confirm every redirect keeps the query string.
- Check the certificate for every hostname, including the measurement domain and any intermediate domain.
- Test from other networks, on mobile and from inside and outside the targeted regions, confirming the endpoint is consistent and never an error page.
- Keep the chain to two hops or fewer, and enter the canonical URL directly as the final URL.
- Keep monitoring after launch. Certificate expiry, server rule changes and a slower measurement service can all break a chain that tested cleanly.
Where measurement and personalisation belong
The tracking template records, the final URL lands, and personalisation and experiments happen within that same URL. The page the platform checks is then the page visitors see. Running experiments without changing the URL is covered in same-URL A/B testing, and the equivalent path rules for video campaigns in YouTube ads landing page requirements.
Frequently asked questions
Can a redirect be used as the final URL in Google Ads?
A final URL that redirects is not automatically a problem, provided the redirect stays on the same domain, resolves quickly and ends at the same page for every visitor. A final URL that redirects to a different domain is the classic cause of a destination mismatch. The cleaner setup is to enter the page people actually land on as the final URL and put any tracking hop in the tracking template.
Does the tracking template have to be on the same domain as the final URL?
No. The tracking template can start on a third-party measurement domain. What has to hold is that the template ends at the final URL, which is why it normally contains {lpurl}. A template that finishes on a page other than the final URL creates a mismatch between what the ad declares and where the visitor arrives.
Why is gclid or a UTM parameter missing when the landing page loads?
The two common causes are a redirect on the landing domain that rebuilds the URL without its query string, and parameters that were being added by a tracking redirect, which no longer sits in the visitor's path under parallel tracking. Loading the final URL with test parameters appended shows the first; moving landing-page parameters into the final URL suffix resolves the second.
Are link shorteners allowed?
Shortened links add a hop and hide the destination domain, which makes the declared destination harder to verify. They are best kept out of the final URL field. If a shortener is part of the measurement setup, it belongs in the tracking template and must still resolve to the final URL.
Does changing a tracking template send ads back for approval?
Editing URL options at account, campaign or ad group level generally does not send ads back through approval, whereas changing the final URL on an ad does. Either way the new path is still fetched and checked, so a template change should be tested with the same care as a new ad.
ROAS365 keeps a visit-level record for every visit, serves all visitors the same landing-page URL, and lets every decision along the path be examined afterwards. To see how that fits your campaign structure, contact sales.