Your GA4 Landing Page Report Is Forty Rows of the Same Page

By Jackson Kolinski · November 16, 2026 · 8 min read
Your GA4 Landing Page Report Is Forty Rows of the Same Page

You open the landing page report to find out which pages are earning the traffic, and instead you get the same page forty times. Once with a Google click ID, once with a Meta one, once per UTM combination anyone has ever used, once with a tracking parameter an email tool appended. Each row has a small number next to it. The page that is actually working does not appear to be working, because its performance has been divided into forty pieces.

This is not a misconfiguration. It is what the dimension is defined to do, and knowing that is the first step to getting a readable report back.

Why it happens

GA4’s landing page dimension is, in full, landing page + query string. The query string is part of the value by design, not an accident of collection. Two visits to the same page with different parameters are two different dimension values, the same way two different URLs would be.

That design is defensible. The query string sometimes is the content: /search?q=blue+shirts and /search?q=red+socks are genuinely different landing experiences, and a report that collapsed them would be hiding something real.

The problem is that most query strings on most sites are not content. They are campaign tagging and click IDs, bolted on by whoever sent the traffic. So a dimension built to preserve meaningful distinctions ends up preserving meaningless ones, and the volume of meaningless ones grows with every channel you add.

Paid traffic makes it dramatically worse, because click IDs are unique per click. One Google Ads landing page with auto-tagging can generate a separate row for every single visit. At that point the report is not merely untidy, it is structurally incapable of telling you anything.

Four ways to deal with it

Ordered roughly by how permanent each one is, and they are not mutually exclusive.

1. Know what GA4 does not have

Start here, because most advice on this problem is written for a feature that no longer exists. Universal Analytics had a view setting called Exclude URL Query Parameters, which did exactly what you want: list the parameters, and they stop fragmenting your pages. GA4 has no equivalent.Google’s own migration reference states that the feature is not currently available in Google Analytics 4 properties.

There is a GA4 setting that looks like the answer and is not: data redaction, found under the web data stream in the Events section. It does strip listed query parameters from page_location, but it was built for a different purpose — keeping personal data such as email addresses out of your analytics — and it runs client-side before data is collected, which means it is permanent and applies only going forward.

You can use it for tidiness, and some teams do. Be clear-eyed about the trade: it is irreversible, it does not touch data already collected, and if you later discover a parameter did distinguish real content, it is gone. If you go this route, redact the obviously worthless identifiers rather than everything, and write down what you redacted and why.

One reassurance, since it is the question everyone asks: removing parameters from the page location does not cost you campaign attribution. GA4 reads UTMs into its traffic source dimensions separately from the page path.

2. Override page_location in your tag

The code-based route: set page_location yourself, without the query string, when the page view fires. This gives you full control over which parameters survive, and unlike redaction the logic lives somewhere you can read and change.

One sharp edge worth knowing before you try it: the value you assign must still be a valid URL. Assign something malformed and the Page location dimension comes back empty, which is a considerably worse outcome than the messy report you started with.

3. Use page path in exploration reports

If you do not want to change collection, build the report around a dimension that excludes the query string. An exploration using page path rather than landing page + query string gives you one row per page.

The trade-off is that page path is not the same thing as a landing page — it counts every view of the page, not only sessions that started there. For “which pages matter” that is often fine. For “which pages people arrive on” it is a different question with a different answer, and conflating them is its own reporting error.

4. Fix it in the reporting layer

If GA4 feeds a warehouse or a BI tool, normalise there: strip the query string at query time and group. This keeps the raw data intact, which means a decision you regret is reversible, and it is the right answer for anyone already modelling this data downstream.

It only helps the people looking at the modelled view, though. Anyone opening GA4 directly still sees forty rows.

Why paid traffic is the worst case

Worth separating out, because it changes the severity rather than just the volume. Campaign parameters repeat — a thousand visits from one newsletter share one UTM set, so they collapse into one row. Click IDs do not. Google Ads auto-tagging appends a gclid that is unique to the click, and Meta appends an fbclid the same way.

So every single paid visit to a landing page produces its own dimension value. A page with 2,000 paid sessions becomes 2,000 rows of one session each, every one of them rounding to nothing in the report. The page does not look like a weak performer; it effectively vanishes, because no individual row is large enough to appear anywhere near the top.

This is the version of the problem worth fixing first, and it is also the one where stripping is least risky: a gclid carries no reporting value inside the page path. Google Ads attribution does not depend on it surviving in your GA4 page dimension.

The thing to do first

Before changing anything, find out which parameters are actually on your URLs. Open the landing page report, sort by the fragmented page, and read what the query strings contain. Most sites discover at least one surprise: a parameter from a plugin nobody remembers installing, a session ID that should never have been in a URL, or a tracking value appended by a channel partner.

That audit occasionally turns up something more serious than a messy report. Query strings sometimes carry email addresses, order details or identifiers that should not be in a URL at all, since anything in a URL reaches your analytics and frequently your ad platforms too. If you find that, it stops being a reporting problem and becomes the one described in sensitive data in your URLs.

The related mess you will meet next

Fixing the landing page report usually reveals a second problem underneath it, because the same parameters that fragment your pages are the ones that decide your traffic sources. If your UTMs are inconsistently spelled — nine capitalisations of Facebook across four teams — then your channel reporting is fragmented in exactly the same way, and that one does not get fixed by stripping parameters. UTM parameters breaking your reporting covers the convention side of it.

The other neighbour is traffic landing in “not set” or “unassigned”, which is a different failure with a different cause — parameters missing entirely rather than varying — and is covered in why GA4 shows not set for source and medium.

What to do, if you want the short version

  1. Open the landing page report and read the query strings on your worst offender. Ten minutes, and it tells you which of the fixes you need.
  2. Build the recurring report as an exploration on page path. No permanent change, works today, and covers most of what people actually want.
  3. Redact click IDs only — gclidand the equivalents — if the paid fragmentation is bad enough to matter. They carry no reporting value in the page path.
  4. Leave everything else alone unless you have a specific reason, because redaction is permanent and the regret is not recoverable.

Why this matters beyond tidiness

A fragmented landing page report does not just look bad. It actively misleads, because every row individually looks unimportant. A page that drove 400 sessions across thirty rows of twelve each will sit below a page that drove 60 in one row, and anyone scanning the top of the report concludes the wrong thing about which content is working.

That is the worst class of reporting error: not an obvious absence that prompts an investigation, but a plausible-looking ranking that is wrong. The same shape of problem recurs across ad tracking generally, and it is why knowing what size of discrepancy is normal is worth more than it sounds.

Taglertwatches the ad-account side of that — whether your Meta and Google Ads conversions are still arriving and still matching — and emails you when they change. It has no view of your GA4 landing page report, which is a configuration choice rather than a fault, and nothing in this post is something we will catch for you. It is a twenty-minute fix you do once and benefit from for years.

Sources

Stop finding out about broken tracking from your client.

Taglert monitors your pixels and conversion tracking 24/7 and alerts you the moment something breaks.

Start your free trial
  • First scan in ~5 minutes
  • 30-day money-back guarantee
  • No credit card

Related reading

  • Server-Side Tracking on Shopify: Three Routes, and What Each One Costs You

    Shopify pushed everyone toward sanctioned, sandboxed tracking, and the standard answer became 'go server-side'. That is three quite different decisions wearing one phrase: a native channel app, a managed service, or infrastructure you own. Here is what each actually buys, and the four problems none of them solve.

  • Your Tag Governance Tool Will Not Tell You the Pixel Stopped Working

    ObservePoint, Tag Inspector and Feroot all answer the question 'what is running on this page'. A pixel that stopped reporting conversions is still running on the page. That is why a site scanner can give a clean report on the same day your Meta purchases fell to zero, and why the two jobs need different tools.

  • Triple Whale vs Northbeam vs Hyros: Three Different Jobs, Not Three Versions of One

    These get compared as though they were competing implementations of the same idea. Read each company's own positioning and they are not. One exists to send conversions back to Meta and Google, one to give you a number the platforms did not produce, and one to run a commerce business. Pick by which of those you need.

  • Meta Has Two Gateways Now, and One Is Not a Rename of the Other

    If you deployed a Conversions API Gateway and now keep seeing Signals Gateway, the obvious conclusion is that Meta renamed it. It did not. Meta's own developer documentation lists both products side by side, built on the same platform, doing meaningfully different jobs — and the difference decides whether you need to move.