Analytics reporting reference: reports, explorations, dashboards

Explorations are GA4's ad-hoc analysis workspace, separate from the fixed standard reports. They offer techniques — free-form tables, funnels, path exploration, segment overlap, cohorts — for slicing data by your own dimensions and segments. The trade-off: explorations can sample and apply data thresholds, so small segments need care.

A segment is a saved subset of your data — users, sessions, or events that match conditions — applied to a report or exploration. The crucial detail is scope: a user-scoped, session-scoped, and event-scoped segment of the 'same' condition return different rows, because they include different units. Misreading scope is the classic segmentation error.

Standard reports are GA4's fixed, pre-aggregated reports — grouped into collections like Life cycle and User — that load fast because they read from aggregate tables. Unlike explorations they are not generally sampled, but they apply (other) row grouping and can differ from exploration numbers, which query event-level data with their own scope.

The Realtime report surfaces events and users from approximately the last 30 minutes, refreshing continuously. It is built for spot-checking that tracking fires after a deploy or campaign launch — not for analysis. Its short window and live nature mean its totals will never reconcile with processed historical reports.

Funnel exploration is a GA4 technique that charts how users move through an ordered sequence of steps, showing completion and abandonment at each one. You choose open vs closed funnels and trended vs standard views. Reading it well means matching the funnel's scope and step conditions to the journey you actually mean.

Data blending in Looker Studio combines fields from up to several sources into one logical table by joining on configured keys. It supports join types (left outer, inner, full outer, cross). The common failure is join-key cardinality: a one-to-many key fan-out multiplies metric rows, so blended totals can silently overcount.

A good dashboard answers a specific question for a specific audience at a glance. The durable principles — single purpose, clear visual hierarchy, minimal chart junk, and built-in comparison or context — come from data-visualization practice. This page frames them as design constraints, with no benchmark numbers attached.

GA4's analytics intelligence builds a statistical model of expected values and flags points that fall outside its forecast as anomalies. You can also create custom insights that email you when a condition is met. The judgment call: a flagged anomaly is a deviation from a model, which can be a real event, seasonality the model missed, or a tracking break.

GA4's acquisition collection has two reports: User acquisition attributes by the channel that first brought a user, and Traffic acquisition attributes by the channel of each session. They answer different questions and rarely sum the same way, because one is keyed to first touch and the other to per-session source.

The Engagement collection reports on what users do: events, key events (conversions), pages and screens, and landing pages. Its metrics rest on GA4's engagement model — engaged sessions and engagement time — which replaced the old bounce-centric view, so reading them means understanding what 'engaged' counts.

The Monetization collection reports purchase revenue, item performance, in-app purchases, promotions, and publisher ad revenue. Every figure depends on the ecommerce event schema being implemented correctly — view_item, add_to_cart, begin_checkout, purchase and their item arrays — so most monetization gaps are instrumentation gaps.

The Retention report summarizes how well the property keeps users coming back: new vs returning users, user retention and engagement by daily cohort, and lifetime value. It is a pre-built overview; for custom retention windows and acquisition cohorts you move to cohort exploration.

User explorer is an exploration technique that drills from aggregate down to individual app-instance or user IDs and their event stream over time. It is pseudonymous by design and bounded by retention and thresholds. It is for debugging instrumentation and understanding journeys — not for identifying people.

Path exploration is a GA4 technique that visualizes the branching sequence of events or pages users take, starting or ending at a node you pick. Forward paths show what happens next; backward paths show what led here. It reveals unexpected routes and loops, but node ordering and the start/end choice shape what you see.

Segment overlap is a GA4 technique that compares up to three segments and visualizes their intersections as a Venn diagram. It reveals how much audiences share — for example, mobile users who are also converters. The reading hinges on segment scope: overlap of user-scoped segments means different things than session-scoped ones.

Cohort exploration groups users by a shared starting event (the cohort inclusion criterion) and follows a return criterion across time windows. Unlike the fixed retention report, you choose the inclusion event, return event, granularity, and calculation — so the same data yields very different curves depending on those choices.

User lifetime is a GA4 technique that aggregates metrics across each user's full history rather than a single session or date range — lifetime value, lifetime engagement, first/last touch. Because it spans the whole lifespan, its numbers don't map to a date-range report, which is the most common misreading.

Comparisons let you split a standard report into side-by-side subsets defined by dimension conditions — for example, mobile vs desktop. They are the standard-report counterpart to explorations' segments, but they are simpler, evaluated inline, and limited to dimensions available in that report.

Adding a secondary dimension cross-tabulates a report by a second attribute — channel by device, page by country. It is the fastest way to add context to a table, but it multiplies row cardinality, which can push rare combinations into an (other) row and increase the chance of thresholding.

Through the Library, editors can create custom detail and overview reports, then bundle them into collections that appear in the left navigation. Changes are staged until published, and only users with the right role can edit — so reporting structure is governed, not ad-hoc.

Scheduled email delivery sends a Looker Studio report as a PDF to chosen recipients on a recurring schedule. It is the simplest way to push reporting to stakeholders who won't log in. The caveat: the PDF is a snapshot rendered at send time, with the filters and freshness state then — not a live link.

The Google Analytics Data API (Data API v1) returns GA4 report data programmatically — you specify dimensions, metrics, date ranges, and filters and receive rows. It powers custom dashboards and pipelines, but it shares the UI's quotas, sampling for some requests, and cardinality limits, which must be designed around.

Access to reports is governed by roles: GA4 grants property-level roles (Viewer, Analyst, Editor, Administrator) plus data restrictions for cost/revenue, while Looker Studio shares per report with view/edit and link options. The pitfall is data-source credentials — a shared report can expose data the viewer couldn't query themselves.

GA4 lets you define custom metrics (registered from numeric event parameters) and calculated metrics (formulas combining existing metrics, like revenue per user). They extend reporting beyond the built-ins, but calculated metrics inherit the scope and null-handling of their inputs, which is where formulas go wrong.

A connector is the bridge between Looker Studio and a data source — Google connectors (GA4, BigQuery, Sheets, Ads) and community/partner connectors for everything else. The connector defines available fields, default aggregations, and data freshness/caching behavior, all of which shape what a report can show and how current it is.

Calculated fields let you derive new fields with formulas — arithmetic, CASE logic, text and date functions — at the data-source or chart level. The decisive subtlety is aggregation: a formula's result depends on whether it is computed per-row then aggregated, or on already-aggregated values, which differs between field-level and chart-level calculations.

A KPI dashboard surfaces a deliberately small set of key performance indicators, each shown against a target and a prior period so movement has meaning. The discipline is selection: a KPI must tie to a goal and be actionable, which is what separates it from a vanity metric on a crowded dashboard.

Executive and operational dashboards differ by audience and time horizon, not just polish. Executive views aggregate outcomes against targets over longer periods for strategic decisions; operational views show granular, near-real-time detail for day-to-day action. Mixing the two on one screen serves neither — this page frames the distinction, with no metrics attached.

The right chart follows from the question, not aesthetics. Time series call for line charts; comparisons across categories for bars; relationships for scatter; composition for stacked or 100% bars. Pie charts work only for a few parts of a whole. Matching chart to comparison is what makes a number readable at a glance.

Tables and charts answer different needs. Tables excel at exact lookup, many dimensions, and precise values you might export; charts excel at revealing trend, comparison, and outliers fast. The choice follows the reader's task: looking up a specific number, or grasping a pattern across many.

A sparkline is a tiny, axis-light line embedded next to a number to show its recent trajectory. Coined by Edward Tufte, it adds context to a single value at a glance. But because it usually omits scale, an auto-scaled sparkline can dramatize noise, so it shows shape, not magnitude.

Annotations are dated notes pinned to a report timeline — a deploy, a campaign launch, an outage — so that later a spike or dip carries its explanation. GA4 added report annotations to the property; they turn institutional memory into something a chart shows, preventing the recurring 'why did this move' guesswork.

A report filter narrows what a report or chart displays to rows matching conditions — without changing the stored data. This is distinct from GA4 data filters (which permanently exclude events like internal traffic at collection) and from Looker Studio page/chart filters. Confusing display filtering with data exclusion is the core risk.

Interactivity in Looker Studio comes from controls — date range, filter, drop-down, and the cross-filtering that lets clicking one chart filter the page. The decisive detail is scope: a control affects only charts within its scope (report, page, or group), so a control that seems to do nothing is usually scoped away from the chart.

The pivot table is a GA4 exploration technique that arranges one dimension down the rows and another across the columns, with a metric in the cells — a true cross-tab. It answers two-way questions a flat free-form table can't show compactly, but pivoting on high-cardinality dimensions hits row/column caps and grouping.

GA4's demographics and tech reports include geography — country, region, city — shown in tables and on a geo map. Location is inferred from IP address (with IP not stored), so it is approximate, coarser at city level, and affected by VPNs, mobile networks, and IP anonymization. Read it as a regional signal, not precise location.