GA4 Thresholding and Data Sampling: Why Your Reports Hide Data

Two Different Concepts: Thresholding vs Sampling

When GA4 reports contain incomplete data, it is usually due to one of two mechanisms: thresholding or sampling. They are different problems with different causes and different solutions.

What Is GA4 Thresholding?

GA4 applies thresholds when a report might allow identification of individual users. When a dimension combination has very few users (typically fewer than a few hundred), GA4 removes that row from the report to prevent it from being used to identify specific individuals.

You will see a yellow banner: "This report has thresholds applied." Rows affected by thresholding are simply absent — GA4 does not show them with a "(low volume)" label.

When thresholding happens:

• Google Signals is enabled (which enables cross-device tracking and demographic data)
• You are segmenting by dimensions with few users in specific combinations
• You are using reporting identity that includes Google Signals data

Fix: If thresholding is blocking important data, change your Reporting Identity (Admin then Reporting Identity). Switching from "Blended" or "Observed" to "Device-based" disables Google Signals data in reports and removes thresholding — but also removes cross-device reporting and demographic data.

What Is GA4 Sampling?

Sampling means GA4 analyses a subset of your event data rather than all of it, then extrapolates to estimate the totals. GA4 applies sampling in Explorations (not standard reports) when the query covers a very large number of events.

You will see a coloured dot next to your Exploration (green = no sampling, yellow = some sampling, red = heavy sampling). Standard reports are not sampled — only Explorations.

When sampling happens in Explorations:

• Large date ranges with high event volumes
• Complex queries with many dimensions and filters
• Free-tier GA4 properties (GA4 360 has significantly higher sampling thresholds)

Fix options:

• Reduce the date range in your Exploration (shorter range = less data = less sampling)
• Simplify the query (fewer dimensions and filters)
• Use BigQuery for sampling-free analysis of raw event data

GA4 vs UA Sampling Comparison

Universal Analytics applied sampling to standard reports for high-traffic properties. GA4 does not sample standard reports — only Explorations. This is a significant improvement for standard reporting. The downside is that the complex, flexible analyses (Explorations) are more prone to sampling than UA's equivalent (Custom Reports).

Working Around Sampling in Explorations

When sampling is significant (red indicator), there are several approaches:

1. Narrow the date range: instead of 6 months, analyse 1 month at a time and combine manually
2. Use BigQuery: export data to BigQuery and run the same analysis there on 100% of events with no sampling limit
3. Use standard reports for high-level: standard reports are never sampled — use them for big-picture metrics and Explorations only for deeper analysis where sampling is manageable

Summary

Thresholding and sampling are two distinct GA4 limitations. Thresholding hides rows in standard reports when user counts are too low (caused by Google Signals). Sampling uses data subsets in Explorations for very large queries. Both can be worked around — change reporting identity for thresholding, narrow date ranges or use BigQuery for sampling.

See our GA4 Setup & Audit service for data completeness review.

GA4 reports hiding data you need? Contact Adslytics.