Quick answer: In Google Analytics (Universal Analytics), each property can have several views, and each view can apply its own filters and settings to the data. A clean setup uses at least three views: an unfiltered raw data view as a permanent backup, a test view to try new filters, and a main reporting view with filters for internal traffic, hostname and lowercase URLs. Filters only apply from the moment you add them and cannot be undone on past data, so always test first and never filter the raw view.
How accounts, properties and views fit together
Universal Analytics organizes data in three levels:
- Account: usually your organization
- Property: a website or app, with its own tracking ID (UA-XXXXX-Y)
- View: a set of reports for that property, with its own settings, filters and goals
When you create a property, Google Analytics creates one view, often called “All Web Site Data”. Many sites never add more. That is risky, because any mistake in a filter on your only view permanently affects your data.
A property can hold up to 25 views by default. Most business sites need three or four.
Why views matter
Views let you see the same raw hits processed in different ways. Settings that live at view level include:
- Filters
- Goals (up to 20 per view)
- Site search settings
- Default page and excluded URL query parameters
- Bot filtering
- Ecommerce settings
- Currency and time zone
Data is processed separately for each view. If you set up goals in one view and not another, the second view will not show them.
The recommended view structure
1. Raw data view
Rename the original view to something like “Raw Data, do not edit”. Add no filters, and leave most settings untouched. This view is your safety net. If a filter in another view goes wrong, you still have complete data here.
2. Test view
Create a view called “Test”. Whenever you want to add a new filter or change a setting, try it here first. Wait a few days, compare with the raw view, and only then copy the change to the main view.
3. Main reporting view
Call it “Main” or “Master”. This is the view your team and reports use. It has your tested filters, goals, site search and other settings.
Optional extra views
- A view for a specific subfolder or section, such as a blog or a country-specific area
- A view that includes only internal traffic, for testing tracking
- A view for a specific team, such as one including only paid traffic
To create a view, go to Admin, choose the property, open the View dropdown and click Create View. You can also copy an existing view under View Settings > Copy view, which copies settings and goals but not historical data.
Essential view settings
In View Settings on the main and test views:
- Time zone and currency: match your business
- Bot filtering: tick “Exclude all hits from known bots and spiders”
- Exclude URL query parameters: list parameters that create duplicate page entries without adding meaning, such as session IDs. Leave campaign parameters alone, since they are needed for campaign reports
- Site search tracking: switch on and enter your search parameter, as explained in our guide to site search tracking
- Ecommerce settings: enable if you run a store, as covered in enhanced ecommerce tracking
Filters most business sites need
Filters are created under Admin > View > Filters. You can also create them at account level and apply them to several views.
Exclude internal traffic
Your staff, developers and agency visit the site often and behave differently from customers. Exclude them:
- Filter type: Predefined, Exclude, traffic from the IP addresses, that are equal to your office IP
- For several addresses or ranges, use a Custom filter with a regular expression on the IP address field
Remote staff and changing home IP addresses make this harder. Options include a browser extension that blocks analytics for staff, or a cookie-based approach with Google Tag Manager that sets a custom dimension you can filter on.
If you anonymize IPs, check carefully in your test view that the exclude filter still matches as expected, since anonymization changes how the address is stored. Our article on IP anonymization explains that setting.
Include only your hostname
Spam hits and copies of your tracking code on other sites can send fake data. A hostname filter keeps only traffic from your own domains:
- Filter type: Custom, Include, field Hostname, with a pattern matching your domain and any legitimate subdomains or payment pages
Be careful to include every legitimate hostname, such as a separate checkout domain, or you will lose real data. Test this one thoroughly. Our guide to referral spam covers spam filtering in more detail.
Lowercase request URI
Without this, /Contact and /contact appear as separate pages:
- Filter type: Custom, Lowercase, field Request URI
Consider similar lowercase filters for Campaign Source, Campaign Medium and Search Term, since people tag campaigns inconsistently.
Optional filters
- Append hostname to request URI for properties that track several domains, such as with cross-domain tracking, so pages on different domains do not merge
- Remove trailing slashes where pages appear with and without them
- Subfolder include filter for a section-specific view
Filter order and other rules
Filters are applied in the order listed in the view. Each filter works on data already changed by the filters above it. That matters when a filter rewrites a field another filter depends on, for example lowercasing before an include pattern.
Key rules to remember:
- Filters are not retroactive. They affect data from the moment you save them.
- Filtered data is gone. An exclude filter removes hits from that view permanently.
- Typos are costly. A wrong include pattern can wipe out all data in a view until you notice.
- Use the test view first. Compare a few days of test and raw data before changing the main view.
- Document changes. Add an annotation in reports whenever a filter is added or changed, so future you understands sudden shifts.
Segments vs filters
Segments are applied when you view reports, not when data is processed, and can be used on past data. They are good for one-off analysis, such as comparing mobile and desktop visitors. Filters are for permanent rules about which data belongs in a view. If you only need to look at a slice now and then, use a segment.
Checking your setup
Signs that your views need attention:
- Only one view exists
- Your office traffic appears in reports
- The same page appears several times with different capitalization
- Odd hostnames or spam referrers appear in reports
- Nobody remembers which filters are active or why
A periodic Google Analytics audit catches these issues before they distort decisions. Once your main view is clean, it makes a solid base for dashboards and Data Studio reports.
FAQ
Can I recover data removed by a filter?
Not in that view. That is why the raw data view matters. You can still see the unfiltered data there.
Do goals carry over between views?
No. Each view has its own goals. Copying a view copies its goals, or you can share goal configurations from the Admin area.
How many views should a small business have?
Three is enough for most: raw, test and main. Add more only when a team or section needs its own settings.
Will filters affect real-time reports?
Yes, view filters apply to real-time data in that view too, which makes the test view handy for checking a new filter quickly.
A clean view makes every report more trustworthy
Ten minutes spent on views and filters saves months of questionable numbers. Keep one untouched view, test changes, and give your team one clean reporting view to work from. If you want help getting your analytics set up so you can trust what it tells you, see our Know what’s working service.
