Analytics Market Chart

Organizing Your Google Analytics Account


Google Analytics reports a lot of data. If you don't take out the irrelevant stuff, you will have a hard time getting anything useful from it.

Almost every company I've seen, large and small, uses Google Analytics unfiltered in the beginning. Trying to make useful analysis out of raw data is difficult. To really make sense of the reports, you need to exclude outliers, consolidate pages and isolate irrelevant data.

What Kind of Cleanup?

The kind of data that is in your Google Analytics profile depends on your site, whether you made any modifications to the tracking code and whether you have added filters. You may see:

  • Your own visits, your developers' visits, your employees' and coworkers' visits
  • A separate instance of pages for each visitor or session.
  • Referrals from your own site

How and why to exclude internal traffic is discussed elsewhere. I will just note that internal traffic clouds the reports with irrelevant data.

Google Analytics reports only what it's told. It doesn't try to decide whether two URLs really point to the same page, for example. It reports exactly what is in the address bar of a visitor's browser and then lists each unique URL separately. This isn't helpful when you're looking for trends. Knowing how well a campaign is doing is hard to do when you can't see aggregate data for it.

The goal is to make the data clean enough that the data is statistically significant.

Query Parameters

A common reason for confusing reports is extraneous query parameters.

If your website uses query parameters to track individual users or orders (e.g., session IDs, order IDs, etc.), you will have the same page show up as if it were multiple pages in your reports.

Likewise, if query parameters are used to tell the server to render the page a unique way or style search results in a certain way, these can probably also be ignored.

Filter out all of these kinds of irrelevant query parameters by editing the profile, clicking edit for the main profile information and entering a comma-separated list into the Exclude Query Parameters Field.

If your website uses an internal search box and the search terms are shown in the query string, we can tell Google Analytics to populate Site Search reports and optionally strip these terms out of the URLs as well. On the same page for excluding query parameters, you can enter a list of query parameters that hold search terms or search categories.

Filters

Filters are used for three reasons:

  1. Reformat data
  2. Exclude data
  3. Include only specific data

Exclude filters get used for internal traffic. We won't discuss include filters here. That leaves us with reformatting.

Google Analytics is case-sensitive. If the same URL on your site might be shown in upper case and lower case, you may want a filter to force every URL to the same case. The same goes for keywords, search terms, campaign names, traffic sources, hostnames, etc. Any field where a difference in case is inconsequential may want a filter.

Are some of your URLs total gibberish? Make them human-readable with search and replace filters!

Want to scare a coworker? Use search and replace filters to arbitrarily put their name into keywords reports!

(No, wait.... Scratch that last one.)

The idea is to aggregate as much data as it makes sense to do. Tools like Google Analytics aren't designed to dig down to individual visitor levels. Format the data to make it easier to see the big pictures.

Set Up for Conversions

If your site is an e-commerce site, make sure that you click the little button in the Edit Main Profile Settings area that says "E-commerce". (You also should have added the e-commerce code during setup.)

Even if it's not an e-commerce site, there are certain events that indicate a successful visit. For example, form submissions, file downloads, etc. Set these up as goals in the profile.

Warning Signs

Finally, look for red flags in your reports. There are a few things you may see that will indicate something about your code implementation is wrong.

  • Referrals from your own site
  • 100% of conversions attributed to a single traffic source
  • Domain names in the hostnames report that don't belong to you, or the lack of hostnames that should be included