GA4 + BigQuery consultant: export, first SQL views, dashboards

The free GA4 → BigQuery export is offered by Google up to 1 million events per day, covering most sites up to 10 million monthly pageviews. Activation takes 5 minutes in the GA4 interface. Exploitation, though, takes a few days of consulting to produce usable SQL views, reliable Looker Studio dashboards, and the documentation for your data team to keep exploiting. My GA4 + BigQuery package covers activation, modeling of 3-5 main SQL views, 3 Looker Studio dashboards and handover training for €8,500 (~14 days), delivered in 6 weeks.

When to activate BigQuery on your GA4

Cardinality limits: your GA4 custom dimensions fall into (other) beyond 500 unique values. BigQuery exposes the exact value per event.

Cross GA4 with CRM or ERP: BigQuery lets you join GA4 events_* table with your CRM tables (HubSpot, Salesforce export) or ERP (orders, margins, stocks). Access to revenue + margin pilot by channel, impossible natively in GA4.

Building Looker Studio or Power BI dashboards: GA4 API has cardinality limits, time-window limits (sampling beyond 14 months), dimension-combination limits. BigQuery bypasses all that.

Custom attribution model: GA4 natively offers data-driven, last-click, first-click, linear. For Markov, Shapley, or custom hybrid, it’s BigQuery.

Real-time data: BigQuery export has daily and streaming modes (intra-day with minutes latency). Streaming makes GA4 data into a usable real-time source.

Activation flow

Step 1 — Activation in GA4

GA4 Admin → Property → Product Links → BigQuery Links → Link. Project, region (EU-west default for FR sites), dataset, mode (daily, or daily + streaming).

First table appears next day (daily) or within the hour (streaming). No historical data before activation — activate as early as possible on new setups.

Step 2 — events_* schema discovery

BigQuery creates one table per day: events_20260518, etc. Each row is an event. Main columns: event_name, event_date, event_timestamp, user_pseudo_id, event_params (nested RECORD), user_properties, geo, device, traffic_source.

Step 3 — First useful SQL views

Three views I build on most missions, covering 80% of needs:

sessions_clean: deduplicates session_start, computes session duration, bounce, exposes first and last event of session. Base for behavior reports.

conversions_with_attribution: extracts conversion events (purchase, generate_lead), joins with session source (first-click AND last-click), computes attributed value per model. Base for acquisition performance reports.

users_with_segments: aggregates events per user on 30/60/90 days, computes behavioral indicators (multiple visits, product page views, cart abandons), enables audience construction exportable to Google Ads or Meta.

Step 4 — Looker Studio dashboards

3 typical dashboards on BigQuery:

• Acquisition: sources, campaigns, ROAS per channel, CPA, attribution
• Funnel: conversion rate per step, drop-offs, performing vs weak segments
• Retention: cohorts, return rates, LTV per segment

Step 5 — Handover and documentation

Shared Notion: SQL view documentation (description, business question, query, dependencies), new-dimension-add procedure, BigQuery monthly cost estimate, GCP project access. 2-hour training with your internal data analyst (or maintenance retainer if no internal resource).

Real BigQuery costs

GA4 export free in Google free tier (1M events/day). Costs that appear afterward:

• Storage: ~€0.02/GB/month after free tier (10 GB free). For a 5M events/month site, annual storage cost is marginal (under €100/year).
• Queries: ~€6/TB scanned. Good SQL hygiene (date partitioning, early filters) keeps query costs under €50/month for most sites.
• Streaming export: ~€0.05/GB streamed. Significant above 50M events/month.

In practice, on missions I do, sites between 200K-5M sessions/month pay BigQuery between €30-200/month.

Concrete cases

Media platform 30M pageviews/month: BigQuery activation + 3 SQL views + Looker Studio dashboard crossing editorial/acquisition. 14 days, 5 weeks. Monthly BigQuery ~€180/month. Long-tail analyses across 200K articles enabled, ad-server audiences based on real editorial behavior.

Shopify ecommerce ~€10M revenue: BigQuery activation + cross with Shopify exports for product margin tracking + net ROAS dashboard. 8 days, 4 weeks. BigQuery ~€60/month. 30% of marketing budget reallocated to truly profitable categories.

Frequently asked questions

Time to functional dashboard?

Export starts next day. First usable SQL views: 3-5 days. Production-ready Looker Studio: 8-14 days. Advanced multi-source: 3-4 weeks.

Need internal data engineer?

For setup mission, no — I handle solo. For ongoing maintenance, ideally yes — a data analyst comfortable with SQL is enough. Without internal resource, monthly retainer possible.

Volumes beyond free tier?

Above 1M events/day, export becomes proportionally paid. Sites this size (~30M pageviews/month+) typically pay €500-2000/month BigQuery — still far below the value of exploitable data.