Analytics & Attribution
GA4 Can Block Unknown Hostnames. Measurement Protocol Still Takes Another Route.

Updated: October 29, 2026
Build and test a complete hostname allowlist before activating a GA4 Include filter. Include production, checkout, booking and approved test domains; investigate empty hostnames; and govern Measurement Protocol separately because Google says those events bypass the hostname Include filter.
What Google released
On September 21, 2026, Google Analytics announced Include data filters for hostnames. The filter allows approved domains instead of requiring an endless exclude list for spam. Google also documents two critical boundaries: empty hostnames are blocked, and Measurement Protocol events are not filtered by this control.
Inventory before filtering
| Source | Examples | Decision |
|---|---|---|
| Production web | Main site and regional domains | Allow |
| Transaction systems | Checkout, booking and payment hosts | Allow after verification |
| Development | Staging and preview domains | Separate property or limited allow |
| Third-party embeds | Forms or hosted flows | Confirm actual event hostname |
| Unknown or empty | Malformed or spam traffic | Investigate before blocking |
Use a staged activation plan
- Export hostname distribution for a representative period.
- Map each host to an owner and business purpose.
- Trace critical journeys across domain boundaries.
- Test the filter in the least destructive mode available.
- Monitor event volume, conversions and key-page traffic.
- Activate with a documented rollback path.
- Review new or missing hostnames after deployment changes.
Investigate empty hostnames
Google notes that empty hostnames are often spam or abnormal data, but legitimate collection can also fail to populate expected context. Check server-side tagging, consent behavior, embedded flows and custom event pipelines before assuming every empty value is malicious.
Govern Measurement Protocol separately
Because Measurement Protocol events are not subject to the hostname Include filter, treat their API secrets, event validation, source identifiers and server logs as a separate control plane. Restrict who can send events and alert on unexpected volume or event names.
Connect this audit with MTC’s GA4 KPI-definition QA guide. Cleaner collection does not fix a dashboard whose metrics still answer the wrong business question.
Monitor after activation
- Total event and user change versus baseline.
- Conversion events by hostname and source.
- Checkout and booking completion paths.
- Unknown and empty hostname trends.
- Measurement Protocol volume and validation errors.
- Release dates for domains, tags and consent changes.
Trace every critical journey
Document the hostname sequence for lead forms, checkout, authentication, payment and booking. Some journeys leave the main domain briefly and return. A missing hosted step can remove the very conversion the filter is intended to protect.
Preserve a pre-filter benchmark
Export event counts, users, conversions and hostname distributions before activation. Annotate the effective date in reporting. When totals change, compare by event and journey instead of accepting a cleaner-looking property as proof of better data.
Assign domain ownership
Every allowed hostname should have a technical owner, business purpose and review date. Remove obsolete domains only after confirming traffic has ended. Add a change-management step so new microsites and checkout providers cannot launch without analytics allowlist review.
Frequently asked questions
Will the filter stop all GA4 spam?
No. It controls browser-side hostname collection, while other ingestion paths require separate safeguards.
Should staging be allowed?
Prefer a separate property. If staging must share the property, document and segment it deliberately.
Can deleted data be restored after an active filter blocks it?
Do not assume so. Treat activation as a production change and validate before enforcing it.
