Five Signs Your Google Analytics Data is Inaccurate – And How to Fix It

/

five-reasons-ga-data-inaccurate (1)

We want you to feel confident in the data you have in Google Analytics, and that’s why we are here to help! If you find yourself looking at Google Analytics data and you’re unsure how accurate it is, this post will help you look for common errors that occur and will show you how to fix them. The goal is to be able to use the data you have!

I am going to use the public Google Analytics account that uses data from thegooglemerchandisestore.com. If you would like to follow along, follow along with our great guide to using the demo account before reading this post. You are welcome to use your own Google Analytics data if you have that available.

Now that we have some data, let’s start digging.

Views: Too Many or Only One

First, let’s navigate to your Google Analytics Account and go to the admin screen where you see the 3 columns for Account Settings, Property Settings, and View Settings. If you are in the account and property that you want to report from, click on the View drop-down.

Views - Admin Screen

How many options do you have? Just one? Or maybe a bunch that you aren’t familiar with? We can make that better.

The public Demo Account has three: the main view (used for reporting), a test view (a place to test filters and goals for accuracy), and a raw/unfiltered view (for backup). This is the minimum that we recommend. Having too many could also be a problem, especially if you don’t know which one you should be using for daily reporting or if you don’t know how one view is different from another.

The idea is that you should feel confident in the data you are reporting on and it should be easy for you to navigate to the data you want – views can help make that happen.

If there are many views in there that you are not familiar with, you can check out the filters and goals that are assigned to each one and hopefully find some clarity. We also recommend that you ask around the office (if applicable) to see if other people use them. A short fix is to rename the mysterious views something that you can easily distinguish as either important or not, but the long term solution would be to delete the views you no longer need for reporting or historical purposes.

The need to separate views is dependent on how your site and business works, but here are a few examples of views that you may want to implement:

  1. Internal Traffic vs External Traffic: You can have two views that either filter in or out internal traffic. For instance, you can direct traffic based on the IP address of your office (one for include only your IP address, the other view with exclude your IP address) so that your employee traffic isn’t mixed with your customers.
  2. Locations: You may have content that is region specific. You can set up views for each region you are interested in.
  3. Mobile traffic: Yes, you can apply some secondary dimensions to your reports to see the devices people are using, but if this is a large business conversation for you, you can set up a view that is specific to traffic from a certain device category.
  4. Hostnames/Subdomains: I will talk about this one more below, but if you have different subdomains and main domains, you will want to have a clear way to identify where your Google Analytics data is coming from.

Hostnames Report: What are all of these sites?

If you are asking yourself what the hostnames report is, that’s okay! We really don’t use it for reporting purposes, but it is a great place to look for initial account debugging.

The Hostnames Report shows all of the websites (hostnames) that someone was on when a hit was sent to your Google Analytics property. With that in mind, your hostnames report should only show the websites that are applicable to that property and view.

Here Is How You Find It:

Navigate inside of the Google Analytics interface to Behavior -> Site Content-> All Pages

Then at the top of the data table, you will see an option to switch the primary dimension. Click on “other,” type in “Hostname,” and click on the icon that says “Hostname” in a green box.

In the case of the public Demo Account, we only see shop.googlemerchandisestore.com, and www.googlemerchandisestore.com. This is a good thing. If you see many websites that aren’t part of your business, you should add an “include only” filter to your view(s) that only includes traffic from your specific hostname(s). If you have multiple domains going to one property, we suggest having one view for each domain (if not a different property), then one with all of them together.

If you have multiple domains to include, make sure you put both hostnames in the same include-only filter using regular expressions. If you have two separate include-only filters, you will end up with no data because Google Analytics processes data in the order that the filters are listed.

Subdomains: If You Have Them, Show Them

Does your website have subdomains? If not, you can skip this one. If you do, don’t fear! You can still have accurate data.

The reason that this is a concern with Google Analytics is because the page related reports will use “/” as a name for the row, and you will probably see this as the one with the most traffic. This is the home page, shown with nothing after the first trailing slash. In the case of the Google Merchandise Store, they have “/home” in the reports, but the same concept applies.

Home Page Path

In the case of The Google Merchandise Store, they have www.googlemerchandisestore.com/home and shop.googlemerchandisestore.com/home, so which site are we referring to when we see “/home” in the reports? Both! The metrics in that table are consolidated by the traffic across those homepages. We can prove this by applying the secondary dimension “Hostname.”

Hostname Applied to Page

Want to fix it? Let’s do two things. First, let’s set up different views for each subdomain. Then we can set up a prepend hostname filter in your main view for all traffic so you can still see all sites together, and still be clear about which homepage is which.

Here is a link to a LunaMetrics blog that shows you how to make the Prepend Hostname filter.

(note: can’t use the link feature in the pages report or page dimension anymore)

Case Sensitivity: Inconsistent Capitalization In Campaigns or Events

Google Analytics is case sensitive in all of its reports, which means that you need to be careful when you are naming your events and custom campaigns. If you have email links with utm_campaign=december2016 and utm_campaign=December2016 floating around, those are going to be two different lines in your Google Analytics report and the metrics will be calculated and displayed separately, even though they look like the same thing to us. The same thing goes for events, which is why we suggest that you be cautious when using the Google Tag Manager for {{Click Text}} as a Category, Action, or Label. Even without using GTM or Click Text, this is something you should lookout for.

Do you see instances of inconsistent capitalizations? Let’s fix it (moving forward) with, you guessed it, filters!

You can set up a lowercase filter for the dimensions that are affected by the inconsistent capitalizations to consilidate your data. We like to start with lowercase page paths and lowercase campaign / source / medium.

Here is a link to a LunaMetrics blog that shows you how to make the Lowercase filter.

Query Parameters: Do You Need Them in Your Reports?

If you look at your All Pages report, do you see your page paths trailing with a large number of query parameters? Each variation of your full page URL is displayed and reported on individually in Google Analytics, which is a good thing when you are looking at your designated page path, like shop.googlemerchansidestore.com/womens/sweatshirts and shop.googlemerchandisestore.com/mens/tshirts, but not so good when you add all the query parameters in the mix because Google Analytics will split by the parameters included, the order, and the values for each one. Depending on your traffic volumes and use of these parameters, you may even run into a high cardinality issue where Google Analytics will lump all your extra rows into one row labeled “other.”

To fix and avoid this, we can update your settings in each applicable view. Within your admin screen, navigate to View -> View Settings -> Exclude URL Query Parameters. In this box you can type in the query parameters that you see in your GA reports that you don’t need for reporting purposes.

To clarify, this does not remove the parameters from the page as your user navigates your site, nor does it affect any other third party systems you are using that pull data straight from your webpage. Going forward, this will only affect the View that you’ve updated and will only looks at the data once it is in Google Analytics, removing the key-value pairs you aren’t interested in seeing from the URL that shows up in your reports.

Bonus: Self-Referrals

What?? More than 5 signs?? Yup! You’re welcome.

Here Is How You Find It:

Navigate inside of the Google Analytics interface to Acquisition -> All Traffic -> Referrals.

Find Referrals Report

Unlike the Hostnames report, you should not see your own site listed here.

Looking for self-referrals is a good thing to do after you set up Google Analytics or go through some major website changes to make sure your tracking codes are working as expected. Self-referrals are a symptom of other tracking issues that will need to be addressed. What are some of the root causes of self-referrals? There are four main reasons:

  1. Subdomains or cross domains aren’t set up to share the cookies
    1. Here is a link to our post on tracking subdomains
  2. Google Analytics or Tag Manager code is missing from one page on the site
    1. Check 404 pages and search pages to start
  3. Default campaign timeouts
  4. Sessions that start before, and end after midnight

Now we know how self-referrals can occur, and here are a few reasons why we don’t want them:

  1. It will inaccurately skew metrics like: bounce rate, session count, and users (if cookies aren’t being shared).
  2. We can’t as easily attribute conversions to their true source if Google Analytics starts a new session with its source listed as your own site. If this happens, all the hard work you put into different campaigns will look like they aren’t delivering results, even if they are! You can still see that information via the Multi-Channel Funnel reports, but most of the Google Analytics reports won’t show the info you’re looking for.

And that’s all for now folks! If you are interested in other ways to check for accuracy, check out some of our other blogs that focus some specific topics in Google Analytics.

Donovan is a Junior Consultant at LunaMetrics, and an MBA student at the University of Pittsburgh. He enjoys connecting consumer behavior to business strategy, which led him to LunaMetrics and Pitt's program for Analytics and Marketing. Out of the office, Donovan is learning to play golf and he tries to lower his score each time he plays.

  • One more: if you see your bounce rate is below 30% – more likely you have inaccurate data because of wrong GA code implementation.

Contact Us.

LunaMetrics

24 S. 18th Street, Suite 100,
Pittsburgh, PA 15203

Follow Us

1.877.220.LUNA

1.412.381.5500

getinfo@lunametrics.com

Questions?
We'll get back to you
in ONE business day.