A dashboard is a product with users. Most GA4 dashboards fail the same way: forty charts nobody asked for, quota errors every Monday morning, and a business owner who still asks "so are we doing well?" This guide covers the tools, the technical setup including the quota problem, and a layout that answers the questions people actually have.
Explorations: Your Analysis Workbench
GA4 Explorations live inside GA4 under Explore. They are free on every property and are the right tool for one-off questions:
- Free form for a table or chart with any dimensions and metrics, with segments.
- Funnel exploration for step conversion, for example product view to cart to checkout to purchase, open or closed funnels, with breakdowns.
- Path exploration for what people did before or after an event.
- Segment overlap for how audiences intersect.
- User lifetime and cohort for retention questions.
Explorations use the data you have, respect data retention, and apply thresholding when Google Signals is on. They are not shareable outside GA4 beyond exporting, which is why the weekly dashboard lives in Looker Studio.
Looker Studio: The Shared Dashboard
Looker Studio connects to GA4 and Google Ads directly and lets you build pages anyone with a link can open. Two ways to connect GA4:
- GA4 connector. Quick, no setup, uses the Analytics Data API with its quotas.
- BigQuery connector on top of the GA4 export. More setup, no API quota, raw event data, unlimited retention.
Start with the GA4 connector for a small dashboard. Move to BigQuery when quota errors appear or when you want history beyond 14 months.
The Quota Problem, Explained
Every chart on a GA4-connected page sends a request. The Data API limits each property to ten concurrent requests and to a pool of hourly and daily tokens. Date comparisons, many filters, high cardinality dimensions such as page path, and multiple simultaneous viewers all cost more. When the pool is empty you see an error about exhausted quota.
Reduce the cost:
- Fewer charts per page; one date range control instead of per-chart comparisons.
- Avoid high cardinality dimensions in large tables; use top 20 with a filter.
- Use data extracts for stable pages, refreshed daily.
- Turn off report caching only if you must.
- Move heavy pages to BigQuery.
Step 1: Link GA4 to BigQuery
In GA4 Admin, then Product links, then BigQuery links, choose a Google Cloud project, enable daily export, and optionally streaming. Tables appear the next day as events_YYYYMMDD. Standard properties export for free; you pay only BigQuery storage and query costs, which are within the free tier for most small sites.
Step 2: Prepare the Data Sources
For the GA4 connector: add the data source, choose the property, and in the field list rename metrics to plain language, for example "Key events" to "Leads" if that is what it measures. For BigQuery: write a scheduled query that flattens the events table into a daily table of sessions, users, key events and revenue by source, medium, campaign and landing page, and connect Looker Studio to that table rather than to the raw events.
Add the Google Ads connector for spend. Blend with GA4 on date and campaign name where you need cost per key event.
Step 3: The Layout
Page 1: Overview, one screen, no scrolling.
- Four scorecards with period comparison: Sessions, Key events, Key event rate, Revenue or Leads.
- A time series of key events by week.
- A channel table: channel group, sessions, key events, key event rate, revenue. Sorted by key events.
- A one-line text box that states the main movement in plain words, updated by whoever owns the dashboard.
Page 2: Acquisition.
- Source and medium table with sessions, engagement rate, key events.
- Campaign table blended with Google Ads cost: cost, clicks, sessions, key events, cost per key event.
- Landing page table: landing page, sessions, key events, key event rate.
Page 3: Behaviour and funnel.
- Funnel bar chart: product views to add to cart to checkout to purchase, or page view to form start to form submit for lead sites.
- Top pages by engaged sessions.
- Custom events that matter: phone clicks, chat starts, calculator completes.
Page 4: Ecommerce or leads detail, depending on the business.
- Products by revenue and quantity, or leads by form and by lead source.
- Returning vs new customer share.
Page 5: Data health.
- Daily key events and sessions as sparklines, so a tracking failure is visible within a day.
- A table of events with zero counts in the last 7 days.
Step 4: Make It Readable
- One date control at the top, defaulting to last 28 days with comparison to the previous period.
- Consistent colours: one colour for sessions, one for key events, one for revenue, used on every page.
- Numbers formatted: thousands separators, one decimal on rates, currency on revenue.
- Chart titles that state the question: "Which channels bring leads?" not "Sessions by channel".
- Nothing below the fold on page one.
Step 5: Share and Maintain
Share with view access to the people who need it and schedule an email delivery on Monday mornings. Review the dashboard quarterly: remove charts nobody mentions, add the question people keep asking in meetings.
Common Mistakes
- A dashboard built from every available metric.
- Quota errors ignored until the client sees a blank page.
- No cost data, so conversions have no context.
- Thresholded data from Google Signals making small segments vanish without explanation.
- No data health page, so broken tracking goes unnoticed for weeks.
Outcome
A dashboard that opens quickly, fits on one screen for the overview, shows cost next to results, and tells you within a day if tracking breaks. Explorations stay your workbench for the questions the dashboard raises.