Configure an internal traffic rule for your office address and mark it as testing, then activate the filter. On a small site this is frequently a tenth or more of recorded visits.

You are the biggest visitor

On a site receiving forty or fifty visits a day, the people who look at it most are the people who run it.

You check the homepage after a change, look at a page you just published, load it to show somebody, and open it on your phone to find your own phone number.

Add a designer, a supplier, and anybody testing something, and internal traffic can be a substantial share of what is recorded.

On a small site that is not a rounding error. It is frequently a tenth or more of the total, and it is concentrated on exactly the pages you have just changed.

Why it distorts more than the count

Because your own behaviour is unlike a customer's in every respect.

You arrive directly rather than through search, so your traffic sources look better than they are.

You go straight to the page you want, which flatters navigation that may be confusing.

You spend seconds rather than minutes, or you leave a tab open for an hour, both of which distort time measurements.

And after a redesign you look at everything repeatedly, producing a spike that looks like the change working.

That last one is the expensive version, since it can lead somebody to conclude a change succeeded when they are looking at their own visits.

It does not carry over

The specific point for anybody who set this up before.

Filters configured in the old property are configuration in that property, and a new one starts with none of it.

So the exclusion you set up years ago and forgot about is not protecting your new property, which has been recording your own visits since the day it was created.

The mechanism is also different: rather than a view filter, the new system uses an internal traffic rule combined with a data filter.

Which means it is two steps rather than one, and the second is the one people miss.

How to set it up

The third is where this fails most often, because defining the rule does nothing until the filter is switched on, and the interface presents them in separate places.

A worked example

A business with modest traffic redesigned their site and saw visits rise by about a third in the first fortnight, which they took as evidence the redesign had worked.

Enquiries had not changed.

They had excluded their office address in the old property years earlier and had never done so in the new one.

The rise was them: three people checking pages repeatedly across two weeks, on desktop and on phones, plus their designer.

Once the filter was configured and active, the figures returned to roughly where they had been before the redesign.

The redesign may still have been worthwhile. What it had not done was increase traffic, and for a fortnight they had believed otherwise.

The problem with home working

The complication that makes this less complete than it used to be.

An office address covers people in the office. It does not cover anybody working from home, on a phone using mobile data, or on a client's connection.

Most home connections also have addresses that change periodically, so listing them is unreliable.

Which means the filter catches the largest single source and not all of it.

The practical supplement is to be conscious of it: check the live site from a private window when you can, and note in your figures when a burst of testing happened.

For a small business, catching the office is the great majority of the benefit for a few minutes of work.

Ask your suppliers too

A step that costs one email and is regularly worth it.

A designer, developer, or agency working on your site generates real volume during a project, and they know their own address.

Ask for it and add it, particularly during a build or a redesign when their activity is concentrated.

The same applies to anybody doing regular checks, such as a maintenance supplier loading pages weekly.

It is also a reasonable question to ask when a supplier reports traffic figures to you, since a supplier whose own visits are in the numbers is reporting partly on themselves.

Do it before you need the numbers

Because filtering is not retrospective.

A data filter applies from the moment it is active and does nothing about what has already been recorded.

So the period before you set it up is permanently inflated, and there is no way to clean it afterwards.

Which makes this a setup task rather than a maintenance one, and a reason to do it on the day a property is created rather than when somebody notices.

Where a property has been running unfiltered for months, note the date the filter started so that any comparison across that boundary is read with the change in mind.

The counter-case

This matters in inverse proportion to the size of the site.

A business with substantial traffic will not notice its own visits at all, and the filter is a tidiness measure rather than a correction.

There is also a risk in over-filtering: excluding too broadly can remove genuine visitors, and an address range entered carelessly can exclude a whole area including customers.

Filtered data is also permanently gone rather than hidden, which argues for testing the rule before activating it.

Set it up carefully, test first, cover the office and your suppliers, and accept that the rest is noise you cannot remove.

The setup

  1. Look up your office address.
  2. Add an internal traffic rule in the data stream.
  3. Activate the data filter, which is a separate step.
  4. Set it to testing and confirm before going active.
  5. Ask your designer or agency for theirs.
  6. Do it on day one of any new property.
  7. Note the date the filter became active.

Step three is the one that catches people, since defining the rule without activating the filter changes nothing and looks as though it has.

Other ways figures get inflated are covered in a page that loads twice.


Frequently asked questions

How much of my traffic is my own?

On a small site, frequently a tenth or more, concentrated on exactly the pages you have just changed. On forty visits a day the people who look most are the people who run it.

Why does it distort more than the count?

Because your behaviour is unlike a customer's: you arrive directly, go straight to the page you want, and after a redesign you look at everything repeatedly.

Does my old exclusion carry over?

No. Filters are configuration in the property they were made in. A new property has been recording your own visits since the day it was created.

What is the setup?

An internal traffic rule for your office address in the data stream settings, then activating the data filter, which is a separate step in a separate place.

Why is my filter not working?

Almost always because the rule was defined but the data filter was never activated. Defining the rule alone changes nothing and looks as though it has.

Does it clean up past data?

No. A filter applies from the moment it is active, so the earlier period is permanently inflated. Note the date it started so comparisons across it are read carefully.

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.

Created a new property recently?

Check whether your office address is being excluded. The rule and the filter are two separate steps.

Start a Conversation