AnalyticsSampling

Data sampling and thresholding

Analytics platforms estimating results from a subset of data, or suppressing small values for privacy, both of which distort reporting if unnoticed.

In full

GA4 applies sampling in some exploration and ad-hoc queries above data thresholds, and applies thresholding that hides rows when they could identify individuals. Search Console applies its own filtering and anonymises rare queries, so query-level totals never sum to the reported overall figure. The remedies are BigQuery export for analytics, the Search Console API for search data, and always checking for the sampling indicator before quoting a number.

Example

An exploration reports 62% sampling, and the same query run against BigQuery returns figures 9% different.

Related terms

Google Analytics 4

Google's event-based analytics platform, which replaced Universal Analytics in July 2023 and models sessions and users from event streams.

Google Search Console

Google's free property-level reporting and diagnostics service covering search performance, indexing status, enhancements, manual actions…

Sessions

A group of user interactions within a time frame; in GA4 a session begins with a session_start event and ends after 30 minutes of…

Impressions

In Search Console, the number of times a site's links appeared in results for a query, counted when the link is scrolled or expanded into…

Incrementality testing

Measuring the additional outcomes caused by an activity, by comparing a treated group against an untreated control, rather than inferring…