The GA4 "(other)" Row and Data Thresholds: What They Are and How to Reduce Them

You open a GA4 report and a chunk of your traffic sits in a row called “(other)”. Or you add a filter, and rows disappear, with a small warning icon at the top of the report. Neither is a bug. GA4 deliberately groups or withholds data in certain situations, for two quite different reasons: to keep reports fast, and to protect user privacy. Once you can tell the two apart, you can usually get a fuller picture.

Three different things that look alike

GA4 has three mechanisms that change what you see in a report. They are often confused:

Mechanism	What you see	Why it happens
The “(other)” row	Many values grouped into one row	The table has too many unique values (high cardinality)
Data thresholds	Rows missing, warning icon	Small numbers could reveal individual users
Sampling	Totals based on part of the data	An exploration query is too large

Each has a different fix, so the first step is working out which one you are looking at.

The “(other)” row: too many unique values

GA4 limits how many rows a report table can hold for a given day. When a dimension has more unique values than the limit allows, the least common values are rolled up into a single “(other)” row. The limit is much higher on GA4 360 than on standard properties.

It typically appears with:

• Page paths with query strings or session IDs in the URL, where every visit creates a “new” page.
• High-cardinality custom dimensions, such as a timestamp, a search term or a user ID registered as a dimension.
• Combinations of dimensions that multiply the number of rows, such as landing page by source by city.
• Long date ranges on busy sites, where more unique values accumulate.

How to reduce “(other)”

• Clean up your URLs. Strip tracking parameters and session IDs from page paths where they add no meaning, so one page is one row.
• Be careful with custom dimensions. Register a parameter as a dimension only if you need to report on it, and avoid parameters that are unique per event. Values like timestamps belong in the raw data, not in a report dimension.
• Use fewer dimension combinations in one table, or look at one dimension at a time.
• Shorten the date range when you need detail on a long-tail dimension.
• Use the BigQuery export for full, unaggregated event data. The free export is available on standard properties, and it is the proper home for row-level analysis.
• Consider GA4 360 if you are a large site that consistently hits the limits.

Data thresholds: privacy, not performance

Thresholds are a privacy protection. When the numbers in a row are small enough that someone could work out who an individual user is, GA4 withholds that row. You will see a warning icon in the report, and totals may not match the sum of the visible rows.

Thresholding is most likely when:

• the report includes demographic data (age, gender, interests),
• Google signals data is involved,
• your property has low traffic, or you filter down to a small segment, a short range or a niche page.

How to reduce thresholding

• Check your reporting identity. Under Admin → Data display → Reporting identity, the Device-based option relies only on device IDs. Switching to it can remove much of the thresholding, because data is less likely to identify individuals. The setting changes how reports are presented, not how data is collected, and you can switch back. Be aware that numbers for users can change when you switch.
• Widen the range or the segment. A month instead of a day, a page group instead of a single page.
• Remove demographic dimensions when you do not need them in that particular report.
• Use the BigQuery export, which is not subject to report thresholds in the same way, while staying within your privacy and consent obligations.

Thresholds are a reminder that GA4 is designed to describe audiences, not individuals. On a small business site, very granular reports will often run into them; that is expected.

Sampling in explorations

Standard reports in GA4 are not sampled. Explorations can be: on standard properties, an exploration query covering more than 10 million events is based on a sample. GA4 shows an icon at the top of the exploration indicating whether the data is sampled.

To avoid it, shorten the date range, simplify the exploration, use the standard reports where they answer the question, or move the analysis to BigQuery.

Why this matters for your decisions

All three mechanisms can make the same metric look different in two places. A landing page might show one number in a standard report, a slightly different one in an exploration, and appear only partly in a filtered view. Add GA4’s estimated user and session counts, and it is easy to see why people distrust their reports. We cover the counting side in why your GA4 session numbers don’t add up.

The practical rule:

• For headline trends (total sessions, key events, channels), standard reports are fine.
• For long-tail detail (every URL, every search term), expect “(other)” and plan for BigQuery.
• For small segments (one page, one city, one week), expect thresholds and read the results as indicative.

A quick diagnostic checklist

When a report looks incomplete:

• Is there an “(other)” row? Look for high-cardinality dimensions and long ranges.
• Is there a warning icon about thresholds? Check demographics, Google signals and the reporting identity.
• Is this an exploration? Check the sampling indicator.
• Are filters narrowing the data to a small segment?
• Is the same question answerable in a standard report instead?

Key takeaways

• “(other)” means too many unique values; thresholds mean small numbers and privacy; sampling means a very large exploration query.
• Clean URLs and disciplined custom dimensions prevent most “(other)” problems.
• The device-based reporting identity can reduce thresholding; wider ranges and segments help too.
• Standard reports are not sampled; large explorations can be.
• For full, row-level detail, the BigQuery export is the reliable route.

Making GA4 reports complete and trustworthy is a large part of our Google Analytics and Tag Manager services, and pairs well with clear dashboards, as in automated reporting with Looker Studio. If your reports are full of “(other)” or warnings, get in touch.