Add a Universal Analytics filter excluding your office's IP address, handle staff phones and home computers separately, and keep one view unfiltered. Filters change data permanently, and only from the day they are saved.

The visitor who comes every day

On a small business website, the most loyal visitor is often the business itself.

The owner opens the home page each morning to see that it still works. The office manager copies the address from the contact page. Staff look up their own service pages to read details to a customer on the phone.

Analytics counts every one of those as a visit. On a small site, a handful of staff opening pages every day can make up a noticeable share of the total.

And they behave nothing like customers. They go straight to particular pages, need no persuading, and come back daily, which bends almost every report the business reads.

Where your own visits show up

The last is the easiest to miss. Every test of the contact form records a completed goal, and a month with three tests looks like a month with three more enquiries.

How filters work in Universal Analytics

An Analytics account holds a property for your website, and each property has one or more views: separate ways of seeing the data the property collects.

Filters belong to a view. They decide what that view records, such as leaving out visits from a particular address.

Two rules follow, and both matter. A filter works only from the moment it is saved, never on data already collected. And whatever a filter excludes is gone from that view for good.

Keep one view unfiltered

Because filters cannot be undone, the original view should stay as it is, with no filters at all. It is the record of everything that happened.

Add a second view for trying filters out, and a third as the main view you report from. A new view starts empty on the day it is created, so add them now rather than when they are needed.

If a filter is set wrongly, and one wrong character can exclude every visit, the unfiltered view still holds the whole picture.

Excluding the office

The office is the easy part. Every visit from its network arrives from the same public IP address: the number your internet connection presents to the rest of the internet.

Searching for what is my IP on an office computer shows it. An exclude filter for traffic from that address, added to the test view first, takes the office out of the figures.

Ask your internet provider whether the address is fixed. Many business connections keep one address for years; some change it now and then, and a filter on an old address quietly stops working.

A business with several locations needs each one added, and each one checked.

Phones, tablets and staff at home

An IP filter misses every visit that does not come through the office: phones on mobile data, staff at home in the evening, laptops on a client's Wi-Fi.

Phones in the office can simply use the office Wi-Fi. On desktop computers, Google's opt-out browser add-on stops Analytics recording any visit from that browser, to any site, which suits a work machine.

For phones and tablets used elsewhere, the usual answer is a private link. Visited on each device, it marks that device as internal through Google Tag Manager, and a filter leaves the marked visits out.

Safari on iPhones and iPads now lets that mark expire after seven days, so they need the link weekly.

Many WordPress analytics plugins can also stop tracking logged-in users, which covers whoever edits the site.

A worked example

A closet and storage systems company with a small showroom and three in-home designers used its own website as a sales tool. In customers' bedrooms and garages, the designers opened product pages on tablets to show finishes and layouts.

The owner had excluded the showroom's IP address years before. The tablets, on mobile data, had never been excluded.

So the walk-in closet pages looked like the most popular on the site, busiest on weekday afternoons, in exactly the towns the designers covered. The owner was about to expand them and cut pages that looked neglected.

A private link, bookmarked on each tablet and opened every Monday, set a marker; Tag Manager labelled those visits as internal, and the main view excluded them.

The unfiltered view kept everything, so the designers' use of the pages stayed visible there. That turned out to be worth knowing in its own right.

Test before trusting

Put every new filter in the test view first, and leave it for a few days.

Then compare the test view with the unfiltered one. The difference should look like the office: weekday visits, in working hours, to the pages staff use. If the test view shows almost nothing at all, the filter is too broad.

Only then copy the filter to the main view. A mistake made in the test view costs nothing.

A step in the figures

Since filters are not retroactive, the main view's numbers drop on the day the filter starts. Nothing has gone wrong; the business has stopped counting itself.

Add an annotation in Analytics on that date, so nobody comparing this winter with last winter reads the drop as lost customers.

For comparisons that span the change, use the unfiltered view, which counts staff in both periods and so compares like with like.

The counter-case

On a site with heavy traffic, a few staff are a rounding error. The filter then matters mostly for keeping form tests out of the goals.

A sole proprietor who checks the site once a week barely registers either way, and the office filter alone covers most of it.

Some businesses also want staff use counted, such as pages written for staff to read from on the phone. That is an argument for a separate view, not for leaving the main one unfiltered.

Filter the office, deal with the devices, and keep one view untouched.

This week

  1. Leave the original view unfiltered.
  2. Create a test view and a main view.
  3. Find the office's public IP address.
  4. Ask your provider whether it ever changes.
  5. Exclude it in the test view first.
  6. Mark staff phones and tablets as internal.
  7. Annotate the date the filters began.

Step one protects everything else. Filters remove data for good and a new view starts empty, so the untouched original is the only full record of what really happened.

Reading the cleaned figures fairly against last year takes the same care as Black Friday moved, so compare by week.

Frequently asked questions

Why exclude my own visits from Analytics?

Because owner and staff visits behave nothing like customers. They inflate direct traffic, top pages and returning visitors, and contact form tests are counted as enquiries.

Do filters change my past data?

No. A filter works only from the moment it is saved, and whatever it excludes is gone from that view permanently, which is why one view should stay unfiltered.

How do I exclude the office?

Find the office's public IP address and add an exclude filter for traffic from it, in a test view first. Ask your provider whether the address ever changes.

What about staff on phones?

An IP filter misses mobile data. Use the office Wi-Fi in the office, and a private link that marks each device as internal for phones and tablets used elsewhere.

Why do I need three views?

One unfiltered as the permanent record, one for testing filters safely, and one main view for reporting. A new view starts empty, so create them before you need them.

Why did my visits drop after filtering?

Because the business stopped counting itself from that day. Annotate the date so nobody reads it as lost customers, and use the unfiltered view for comparisons across it.

West Coast Media Solutions Inc. provides web design, web development, hosting, digital marketing, and business consulting to organisations across Canada, drawing on more than twenty-five years in the field.

Not sure your figures leave out your staff?

Add us as a user on your Analytics account. We will set up the views and filters, and test them before they count.

Start a Conversation