GA4 properties fail in consistent, predictable ways. Data retention set to two months by default. Staging traffic leaking into production reports. Payment processors stealing attribution from real conversions. UTM parameters fragmented across dozens of casing variants. Consent Mode v2 absent in EU-facing properties. These problems don't announce themselves — they're silent, they compound, and they make every report that touches them quietly unreliable.
A GA4 health score is a structured way of measuring how many of these problems are present in a property — and how severe they are. Rather than a list of individual findings that require interpretation, a score gives you an immediate, comparable, trackable signal of overall data quality.
What goes into a health score
GA4 Health Check runs 47 automated checks across seven distinct modules, each covering a different dimension of property health. Every check either passes or fails, and failures are weighted by severity and impact. The weighted results are normalised to a 0–100 scale.
Not all modules carry equal weight in the overall score. Configuration and conversion tracking issues have a larger impact than audience and reporting setup, because a broken conversion event corrupts data in ways a misconfigured audience definition does not.
| Module | Weight | |
|---|---|---|
| Conversion Tracking | Broken conversions corrupt campaign spend and CRO decisions | |
| Data Integrity | Phantom sessions suppress every metric expressed as a rate | |
| Attribution | Misattributed conversions misdirect budget | |
| Privacy & Compliance | Missing Consent Mode makes EU data invisible | |
| Configuration | Retention and timezone errors affect historical comparisons | |
| Integrations | Missing links reduce usefulness, not accuracy | |
| Audience & Reporting | Setup issues affect usability, not data quality |
What different scores mean in practice
Data cannot be trusted for decisions
Properties in this range have multiple active problems with direct, measurable impact on reported metrics. Conversion counts are likely wrong. Attribution is significantly distorted. Reports that look normal are built on a foundation that isn't. Any decision made from this data — budget allocation, CRO priorities, campaign optimisation — is being made against unreliable numbers. Fix before using the data for anything.
Significant issues affecting key metrics
Several high-impact problems are present. The data is usable for directional decisions but not for precise analysis or confident budget allocation. A property in this range typically has one or two critical issues — staging traffic contaminating sessions, or payment processor attribution theft — alongside several medium-severity configuration gaps. Prioritise the highest-impact fixes immediately.
Usable data with known gaps
The major structural problems are absent, but medium-severity issues are present that limit the precision of analysis. A score in this range often reflects a property that had a thoughtful initial setup but has accumulated configuration drift over time — UTM inconsistencies, missing Consent Mode v2, integrations not fully connected. Data is reliable enough for most decisions, with caveats.
Data is reliable and trustworthy
The property is well-configured, data quality problems are absent or minor, and reports reflect what's actually happening. A score in this range means you can use your GA4 data confidently — for campaign optimisation, CRO decisions, budget allocation, and stakeholder reporting. Remaining findings are typically minor improvements rather than active problems.
Why a single number is more useful than a list
A list of 47 individual checks produces 47 individual findings, each requiring interpretation. Is a misconfigured data retention setting more or less urgent than a missing payment processor exclusion? How do you prioritise four medium-severity issues against two high-severity ones? How do you communicate the overall state of the property to a client or stakeholder who isn't a GA4 specialist?
A score answers these questions immediately. It aggregates the severity-weighted findings into a single comparable number. It tells you whether the property is in critical condition or broadly healthy without requiring the reader to synthesise a 12-page audit report. And it creates a baseline you can track over time — a property that improves from 54 to 81 after a round of fixes has a concrete, documentable measure of that improvement.
For agencies managing multiple client properties, the score is particularly valuable. A single number per property, updated quarterly, immediately surfaces which clients need attention and which are stable. It turns GA4 health monitoring from an ad hoc activity into a systematic one.
Two example scores — what they look like
E-commerce site, £2M annual revenue — typical findings at this score
- Data retention set to 2 months — 10 months of historical data permanently inaccessible
- Staging subdomain sending 14% of all sessions — every rate metric suppressed
- PayPal and Klarna in top referral sources — purchase attribution corrupted
- Consent Mode v2 absent — all EU visitors invisible, Smart Bidding degraded
- UTM casing split across 23 source variants — acquisition report fragmented
- Google Ads linked and conversion importing correctly
SaaS company, 40k monthly sessions — typical findings at this score
- Data retention at 14 months — full historical comparisons available
- All conversion events active, parameters complete and correct
- Consent Mode v2 implemented via certified CMP
- No hostname contamination — staging on separate property
- 3 UTM source variants for the same email newsletter — minor fragmentation
- Search Console not connected — organic keyword data unavailable in GA4
How to improve your score
The highest-impact fixes are almost always in the conversion tracking and data integrity modules, because these carry the most weight and their problems have the most direct effect on the metrics you make decisions from. In order of typical impact:
- Fix data retention — change from 2 months to 14 months immediately. Takes 30 seconds, is not retroactive, but prevents further data loss from today forward. Guide →
- Remove staging traffic — exclude non-production hostnames. Immediately improves the accuracy of every session-based metric. Guide →
- Fix payment processor attribution — add payment domains to the referral exclusion list. Restores conversion attribution for a significant percentage of purchases. Guide →
- Implement Consent Mode v2 — via a Google-certified CMP. Recovers EU user modelling and restores Smart Bidding signal. Guide →
- Fix UTM inconsistencies — standardise casing and naming conventions across all campaigns. Defragments the acquisition report. Guide →
Each of these fixes corresponds to a specific module score improvement. A property that resolves all five typically moves from the 50–60 range into the 80+ range — from unreliable data to data you can confidently use for decisions.
