GA4 Source Groups and Hostname filters, and why clean data pays for itself

In June 2026 Google added two small features to GA4 that fix two old, boring, expensive problems. Neither will make a headline. Both change how much you can trust the report you make budget decisions from.

I spend my days inside ad accounts and analytics properties for tour operators and D2C brands. The single most common thing I find is not a clever mistake. It is messy data that everyone stopped questioning years ago. Traffic split across ten names for the same channel. Staging sites quietly mixed into the numbers. Reports that look precise and are wrong.

These two features attack exactly that, so here is what they do and what I would do with them this week.

Source Groups: one channel, one name

If you have run ads for a few years, your traffic sources are probably a mess. The same channel shows up under different names because of UTM tags someone set up in 2021, a tool that tags links its own way, and campaigns named by three different people. Facebook traffic alone can arrive as facebook, fb, meta, paid_social and instagram_feed.

The result is that no single report answers the basic question of where your money actually goes. You end up exporting to a spreadsheet and adding the rows up by hand, or worse, you read the top row and believe it.

Source Groups, which Google shipped on 11 June 2026, let you roll those scattered names into clean channels inside GA4 itself. The grouping is applied retroactively, so your history gets cleaned up too, and you do not have to touch the old tagging to benefit.

That last part matters. Fixing years of UTM habits across booking engines, email tools and old campaigns is a project nobody ever finishes. Grouping them at the reporting layer takes an afternoon.

Hostname filters: keep the junk out

The second feature is less visible and just as useful. GA4 can now filter traffic by hostname, under Admin and then Data Filters. That means you can finally exclude the domains that were never supposed to be in your numbers at all.

The usual suspects I find in real properties.

• Staging and development copies of the site, visited all day by your own team
• Mirror domains and scraped copies of your site that still carry your tag
• Spam hostnames that inflate sessions and wreck your conversion rate

Each one adds sessions that never could have booked anything. Your conversion rate drops, your cost per session looks worse, and someone makes a budget call on numbers that include your own developers refreshing a test page.

Two things to know before you rely on it. The filters are exclude-only, and they are not retroactive. They clean your data from the day you set them, so the sooner they go in, the sooner your history starts being trustworthy.

Why this is worth an afternoon of your time

Clean data is the whole foundation. You cannot make a good call on numbers you do not trust, and most operators I meet are making calls on exactly that.

In tourism this bites harder than elsewhere, because the measurement is already fragile. Bookings complete on other domains, tags miss real sales, and ads lift OTA listings the tag cannot see. I wrote about both of those in the tracking guide and the OTA lift guide. Messy sources and junk hostnames sit underneath all of it, making every one of those problems harder to diagnose.

What to do this week

• Open your traffic sources report and list every name that is really the same channel. That list is your Source Groups plan
• Group them, then check that your paid channels now add up to something that matches what you actually spend
• Find every hostname your tag fires on. Anything that is not your real site goes on the exclusion list
• Write down the date you set the filters. Data before that date plays by the old rules, and future you needs to know that

Nice to see Google fixing the boring stuff that actually matters. The flashy features get the announcements. These two will quietly make more accounts honest than any of them.