What the f*ck is the GA4 data model?

Understanding GA4's event-based data model is the skeleton key that will unlock your love of a misunderstood product.

Jonathan has been working on Google Analytics since it was called Urchin and, along the way, he's learnt how to use GA4 the hard way. This talk will help you understand the new data model and how to get maximum value from it by showing you things like custom events, parameters, and dimensions.

This talk is perfect for anyone who "hates" GA4 but has been forced to use it. Whether you're an experienced user or someone new to the industry, this talk aims to help you come away energised and armed with practical "cheat codes" you can use on your clients straight away.

Transcript

1. Speakerdeck.com/nathanless Linkedin.com/in/moorejonathan @nathanless
2. Many of us are forced to use GA4
3. You’re not alone
4. Today we will unlock GA4
5. A quick comparison
6. Universal Analytics was shit
7. UA was session centric User Session Pageview Pageview
8. Pageview Event Ecommerce Social There were multiple hit types
9. Events were constrained Category Action Label Value
10. GA4 reports focus on users User Event Event
11. Flexibility wins
12. Events and Parameters
13. Events measure actions
14. Parameters define context WHO WHAT WHERE MEASURE
15. Example: Click events
16. WHO WHAT WHERE click_text first_user_source page_location MEASURE event_count
17. Example: View Item
18. WHO WHAT WHERE product_name first_user_medium page_location MEASURE event_count, price
19. Every event has a scope User Session Event
21. SCOPE IS WHY NUMBERS DON'T MATCH! When acquisition reports don’t match Explorations or Looker Studio, it’s usually a scope mismatch.
22. User scoped dimensions
23. Session dimensions
24. Event scoped dimensions
25. Parameters are everywhere
26. Event Sources and Measurement
27. Automatic Enhanced Recommended Custom Four Types of Event
28. Key events = your objectives
29. Search Product Purchase Home Add to cart Checkout Map user journeys before choosing what to implement
30. Then prioritise Now Next Future Don't
31. Work out parameters https://support.google.com/analytics/table/13594742
32. Reuse names https://support.google.com/analytics/answer/14240153?hl=en
33. User Interface GTM DataLayer Gtag() Common ways to implement
34. Screaming Frog is always helpful https://www.screamingfrog.co.uk/seo-spider/user-guide/configuration/#custom-javascript
35. As are LLMs
36. Find table structure
37. Explore what is available
38. Analysing Data
39. Standard Reports are rigid
40. Explorations unlock flexibility
41. Don’t forget to register parameters!
42. Scope shapes your explorations https://data.ga4spy.com/?search=landing%2520page
43. event_name and event_count are your friends
44. Build depth with rows
45. API = Time Machine https://developers.google.com/analytics/devguides/reporting/data/v1/rest/v1beta/DateRange
46. API = 9 dimensions https://developers.google.com/analytics/devguides/reporting/data/v1/rest/v1beta/Dimension?hl=en
47. Play with connectors https://workspace.google.com/marketplace/app/ga4_reports_builder_for_google_analytics/589269949355
48. Check out API wrappers https://github.com/locomotive-agency/googlewrapper
49. BigQuery = No Limits https://support.google.com/analytics/answer/9823238?hl=en
50. Choose your weapon wisely Easier Harder One-off Regular BigQuery Pipelines Standard Reports Explorations Third-party API connectors API BigQuery Queries
51. Plan your Analytics journey Standard Reports Explorations Third Party Integrations API BigQuery
52. 1. Map user journeys and parameters 2. Don’t tag everything 3. Register parameters 4. Review key events 5. Check scopes match 6. Lean into event_name + event_count 7. Use the API for historic data 8. Set up BigQuery link What we have learnt
53. Thank You!