Data Sampling
A technique where analytics tools analyze a subset of data and extrapolate results to the full dataset. Introduces estimation error that grows with the degree of sampling applied.
How sampling works in GA4
When you create exploration reports in GA4 that query large volumes of data, Google may analyze only a portion of events and use statistical methods to estimate the full result. The free version of GA4 has lower sampling thresholds than GA360, meaning sampling activates sooner for most businesses.
Google indicates when sampling is active through a shield icon in exploration reports. Standard reports use a different methodology Google calls “blended data,” which combines observed and modeled data.
Why it matters
For high-level trends, sampling may be acceptable. For specific analyses — campaign performance by segment, conversion path analysis, revenue attribution by creative — the margin of error can lead to wrong conclusions and misallocated budgets.
The alternative: full-resolution analytics
Cookieless analytics platforms like SealMetrics record every session individually with no sampling applied. When you see 72,847 visitors, that number represents 72,847 actual sessions — not a statistical estimate.
Related concepts
- Data Loss in AnalyticsThe gap between actual website traffic and what analytics tools report. Caused by consent rejection, ad blockers, browser restrictions, and data sampling. Typically 70-87% in the EU.
- Event TrackingThe method of recording specific user interactions on a website beyond pageviews — clicks, form submissions, video plays, downloads, and eCommerce actions. GA4 uses an event-based data model where every interaction is an event.
- Bounce RateThe percentage of sessions where a visitor views only one page before leaving. In GA4, bounce rate is the inverse of engagement rate — a session is a bounce if it lasts less than 10 seconds, has no conversion, and has no second pageview.
- Revenue AttributionConnecting revenue events (purchases, subscriptions) to the marketing channels that drove them. SealMetrics uses last-click on 100% of observed events — no per-user journey tracking, no multi-touch models.