GA4 for dental practices: the questions it can answer, and the ones it can't
On this page
GA4 for dental practices is good at one job: showing how people arrive on your website and what they do there before they enquire or book. It can't tell you whether they turned up, whether they were new patients, or what treatment they went ahead with. Those answers live in your practice management system (PMS).
GA4 answers some questions well, each with an analysis below, and stops at a clear point. GA4 analysis is one part of tracking ads to booked treatment. The full method, stage by stage, is in measuring marketing against booked revenue. Every GA4 feature here was checked against Google's own documentation on 1 October 2026.
What GA4 can and cannot tell a practice
GA4 (Google Analytics 4) is Google's website analytics. It records events: a page view, a button tap, a form sent. It knows the visitor's device, the page and the source that brought them, such as a Google ad or a search result. It never sees the appointment book.
| Owner's question | Can GA4 answer it? | The analysis |
|---|---|---|
| Which sources send visitors to my implant or Invisalign pages? | Yes | Standard acquisition reports, or a free-form exploration |
| Where do visitors drop out between a treatment page and an enquiry? | Yes, if each step is an event | Funnel exploration |
| What do people read before they enquire? | Yes | Path exploration, run backwards from the enquiry |
| Which pages lead to taps on the phone number? | Yes, the tap only | Free-form exploration on the phone-click event |
| Did that caller book? | No | Call tracking plus the practice system |
| Did the patient attend, and were they new? | No | Practice system |
| What treatment did they accept, and what was it worth? | No | Practice system |
| Would they have booked without the ad? | No | A controlled test, not analytics |
The top half of that table is GA4's territory. The bottom half is where most practice owners actually want answers, and no amount of GA4 configuration will reach it.
Key events worth setting
A key event is an event you tell GA4 is important to the business, so it gets reported as a result rather than just activity. You mark an event as a key event by clicking the star next to it in the Events section. A standard property allows up to 30 key events.1 Key events can then be used to create conversions in Google Ads, which is the word Google Ads uses for its results.1
Thirty is a ceiling, not a target. A practice website needs these:
| Key event | What it records | Watch out for |
|---|---|---|
| Completed online booking | A slot requested or taken in the booking tool | Often invisible to GA4 without extra work |
| Enquiry form sent | A form submission with contact details | Check it fires only on a real submission |
| Phone number tap | A tap on a click-to-call link | A tap is not a call, and a call is not a booking |
| Email link click | A tap on your email address | Low volume, still worth having |
Google lists generate_lead as a recommended event name for a form or information request, and it's the natural name for an enquiry.2 GA4's enhanced measurement also collects form_start and form_submit by default.3 Test an automatic form event with a real submission and a failed one before you mark it as a key event.
Booking tools are the hard part. Whether a booking can reach GA4 at all depends on the tool, and that setup is covered in tracking online bookings in GA4.
Explorations: funnels and paths
Explorations are GA4's analysis workspace, separate from the standard reports. Google lists seven techniques: free-form, cohort, funnel, segment overlap, user explorer, path and user lifetime.4 A new exploration is private to you until you share it.4 For a practice, two of them do most of the work.
Funnel exploration
A funnel exploration shows how many visitors complete each step of a sequence, and where the rest drop out. You can define up to 10 steps. An open funnel lets people enter at any step; a closed funnel only counts people who start at step one. Each user is counted once, for their first pass through the funnel in the date range.5
A typical practice funnel:
You can break the funnel down by a dimension such as device or source, and turn on elapsed time to see how long people take between steps.5 If mobile visitors fall away at the booking button and desktop visitors don't, you have found a page problem, not a marketing problem. How to read that behaviour alongside heatmaps and recordings is covered in how patients use your dental website.
Path exploration
A path exploration draws the routes people take as a tree. Run it forwards from a starting point, such as landing on your implants page, or backwards from an ending point, such as the enquiry event, to see what people did just before. Each step shows the top 5 nodes by default, expandable to 20; the rest are grouped as "Others".6
The backward path is the useful one: it shows which pages people read before they enquired.
One limit to know: explorations may switch to sampled data when a single query covers more than 10 million events.4 A sampled answer is an estimate from part of the data.
The BigQuery export
BigQuery is Google Cloud's data warehouse: a database you query with SQL, a standard language for asking questions of tables. GA4 can send every raw event to BigQuery each day. Standard GA4 properties can use it, and you can export into the BigQuery sandbox free of charge, within the sandbox limits.7
Why a practice would bother: the export keeps event-level history that GA4's own retention setting would delete, and the data can be queried alongside other tables you control.
| Limit | What Google documents |
|---|---|
| Daily export on a standard property | Up to 1 million events a day8 |
| Over that limit | The daily export is paused, and earlier days are not reprocessed7 |
| Streaming export | Charged at $0.05 per gigabyte8; not supported in the sandbox9 |
| Sandbox storage | 10 GiB, a lifetime limit9 |
| Sandbox queries | 1 TiB of processed data a month9 |
| Sandbox tables | Expire automatically after 60 days9 |
| Beyond the sandbox | Charged at BigQuery's rates8 |
That 60-day expiry makes the sandbox a trial, not an archive. Keeping history means adding a Google Cloud billing account. Google's setup page doesn't say whether data from before the link is exported, so I treat the export as starting on the day it's switched on. That is a reason to link it early, even if you don't query it for a year.
Data retention and consent
Data retention is how long GA4 keeps event-level and user-level data. A new property keeps 2 months by default. A standard property can change that to 14 months.410 Retention affects explorations and funnel reports, not the standard aggregated reports.10 Left on the default, a funnel exploration can't look back further than two months. Changing it is a single setting, and I change it on day one.
Consent shapes what GA4 sees in the first place. Since 5 February 2026, UK law has allowed analytics without consent if it's used only to improve the site, visitors are told clearly and given a simple, free way to object, and the provider uses the data only on your behalf. Most practice sites link GA4 to Google Ads for conversions or audiences. That is advertising, so it needs consent, and visitors who decline are not recorded in the normal way. How the banner and Google's consent mode work together is covered in consent mode on dental websites.
Consent mode can estimate the visitors who declined, using behavioural modelling: Google's machine learning fills the gap based on similar visitors who accepted. It only switches on when a property meets all of these:11
| Requirement | Google's threshold |
|---|---|
| Consent mode | On every page, using the advanced implementation (Google tags load for every visitor, not only those who consent) |
| Declined traffic | At least 1,000 events a day with analytics storage denied, for at least 7 days |
| Accepted traffic | At least 1,000 daily users with analytics storage granted, on at least 7 of the previous 28 days |
| Reporting identity | Set to Blended to see modelled data |
Google applies further quality checks even then, and modelled data isn't in the BigQuery export.11 If your site doesn't reach 1,000 consenting users a day, there's no modelling, and GA4 shows only the people who said yes. Where GA4 runs on consent, read every GA4 number as a count of consenting visitors, not of all visitors.
Where the practice system takes over
GA4 ends at the enquiry. Everything that makes a practice money happens after it.
Joining the two halves is the job of dental marketing attribution: the click identifier captured on the website is matched to the appointment and invoice in the practice system, and the result goes back to Google Ads. GA4 supplies only the first half.
For the owner's monthly view, GA4 feeds a dashboard in Data Studio, Google's free reporting tool, which was called Looker Studio until Google renamed it in April 2026.12 Its GA4 connector matches the standard GA4 reports but not explorations, and it can't use GA4 segments or comparisons.13 What that dashboard should show is set out in what a dental marketing report should show.
Where GA4 stops, the next read is what a dental marketing report should show. To find out which of your questions GA4 can answer, email Fayez@imfayez.com; the first look is free. Any unfamiliar term on this page is in the dental marketing glossary.
Sources
-
Google Analytics Help, "Mark events as key events". https://support.google.com/analytics/answer/13128484, accessed 1 October 2026. ↩ ↩2
-
Google Analytics Help, "[GA4] Recommended events". https://support.google.com/analytics/answer/9267735, accessed 1 October 2026. ↩
-
Google Analytics Help, "Automatically collected events". https://support.google.com/analytics/answer/9234069, accessed 1 October 2026. ↩
-
Google Analytics Help, "Get started with Explorations". https://support.google.com/analytics/answer/7579450, accessed 1 October 2026. ↩ ↩2 ↩3 ↩4
-
Google Analytics Help, "[GA4] Funnel exploration". https://support.google.com/analytics/answer/9327974, accessed 1 October 2026. ↩ ↩2
-
Google Analytics Help, "[GA4] Path exploration". https://support.google.com/analytics/answer/9317498, accessed 1 October 2026. ↩
-
Google Analytics Help, "Set up BigQuery Export". https://support.google.com/analytics/answer/9823238, accessed 1 October 2026. ↩ ↩2
-
Google Analytics Help, "BigQuery Export". https://support.google.com/analytics/answer/9358801, accessed 1 October 2026. ↩ ↩2 ↩3
-
Google Cloud documentation, "Use the BigQuery sandbox", page last updated 24 September 2026. https://docs.cloud.google.com/bigquery/docs/sandbox, accessed 1 October 2026. ↩ ↩2 ↩3 ↩4
-
Google Analytics Help, "Data retention". https://support.google.com/analytics/answer/7667196, accessed 1 October 2026. ↩ ↩2
-
Google Analytics Help, article 11161109 on behavioural modelling for consent mode. https://support.google.com/analytics/answer/11161109, accessed 1 October 2026. ↩ ↩2
-
Google Cloud Blog, "Data Studio returns as new home for Data Cloud assets", 11 April 2026. https://cloud.google.com/blog/products/data-analytics/looker-studio-is-data-studio, accessed 1 October 2026. ↩
-
Google Cloud documentation, "Connect to Google Analytics" (Data Studio), page last updated 24 September 2026. https://docs.cloud.google.com/looker/docs/studio/connect-to-google-analytics, accessed 1 October 2026. ↩