Accurate Google Analytics (GA) data is the first step to making smarter business decisions, reducing costs, and learning more about the products and services you sell on your ecommerce site.
But do you trust your current Google Analytics data? If not, it’s time to learn how to perform an audit of your setup.
As someone who has worked on hundreds of Analytics profiles at Inflow, I get to see many different approaches to GA setup, and the problems faulty setups cause when it comes to making informed marketing decisions. As the saying goes, it’s typically garbage in, garbage out.
When Should You Audit Your GA?
In an ideal world, you want to have your data quality checked at least on the following occasions:
- Website launch/initial setup
- Website migration
- After major changes to the site’s structure, design, CMS
- When there are changes in the business logic/site business goals
I usually recommend several audits each year, even if none of the occasions above are relevant to your site. Communication failures can happen between departments, especially in larger organizations. So, make sure to minimize the chance that they ruin your data.
Tools of the Trade
Below are the components of the standard toolkit we use at Inflow.
- Chrome Tag Assistant – Get a quick view of installed tags on any page of the website. The new “Record” function is your best friend.
- Chrome Google Analytics Debugger – A bit intimidating for non-programmers, but extremely useful for watching Google Analytics’ inner workings as you interact with the site.
- Google Tag Manager - Avoid having your changes spend weeks in the development queue and have more control over which tags fire on which pages or events. Here is the best GTM tutorial I’ve found.
- Analytics Toolkit’s Health Status Checkr - If you have to do audits on multiple sites every month, it’s worth considering an automated tool to assist in the task. There are multiple alternatives you can consider, but this one is the most thorough in my opinion.
This site offers some really great tools that ecommerce businesses can use to get the most reliable information they can from Google Analytics.
What to Inspect in a Google Analytics Audit
Below are some of the areas we typically check, and with which we often find problems, when performing a GA audit.
Google Analytics Code on the Site
Whether the GA implementation is done directly or via a tag management solution, make sure that it’s present on all pages of your site. Here’s a tool that should help.
First, check your source/medium data for an unusual percentage of self-referrals, which indicates that traffic from your own domain is being counted as outside referral traffic.
If you still don’t have your own domain listed as a referral exclusion, now would be the time to do that to ensure clean data going forward. Here are the instructions.
Next, make sure all relevant events are being tracked, such as when someone adds an item to their wishlist, submits a form or signs up for an account. It is not enough to just track and report on purchases, as there are plenty of other pre-purchase goals to consider.
Here’s how to see GA reports about your tracked events:
- Sign in to your Analytics account
- Navigate to your view
- Select the Reporting tab
- Select Behavior > Events
Same Version of the GA Code
Make sure to have the same version of the code across the entire site. We’ve had situations where a larger site is divided into many sections, and sometimes these are run by completely separate departments, resulting in one part of the site having Classic GA code, while the rest has migrated to the Universal GA syntax.
What I mean by this is that due to the differences in cookie usage on separate parts of the site, you can end up with broken sessions and an unusual percentage of self-referrals.
You could also use Screaming Frog, or another crawler, to go through the site and check if the Google Analytics or GTM snippets are present on all pages. Seer Interactive offers a great tutorial on this feature.
The easiest way to spot double-tracking is to check your bounce rate report – overall and broken down by landing page (bonus points if you’ve set up content groupings and can look at landing page content groups). When double-tracking is occurring, the bounce rate is usually less than 2-3 percent, so it’s easy to spot. If you spot such an issue, use the previously recommended tools to identify the duplicate tracking and remove it.
Check the Network -> Hostnames report for hostnames that shouldn’t be there. Since the report identifies which domains your Google Analytics is executed, make sure to remove the code from any domains that it shouldn’t be on or exclude that traffic via a view filter.
You’ll also want to exclude traffic from development or staging domains or subdomains so that it doesn’t affect your overall metrics. Even if that traffic volume is a very small percentage of overall traffic on your site, you don’t want those test transactions to influence your most important metrics.
Make sure if you set up any filters that you keep a “raw” profile (with no filters) since sometimes a rogue filter can wreak havoc on your Google Analytics data.
Make sure to set up site search tracking if your site has an internal search feature. It’s quick and easy to do in most cases and it comes with great benefits, as Avinash Kaushik points out in his article Five Questions to Ask of Your Site Search Data.
Site search features are useful in many ways — they adjust product visibility, assist decisions on offering new products or product variants, gauge demand even before it results in purchases and more. Here are the instructions for setting up Site Search reports in Google Analytics.
Look for conflicting or badly configured view filters. This is an extremely wide area to cover, so we can’t go into much detail. However, make sure that your internal traffic (basically traffic from everyone working ON not WITH the site) is excluded by IP/IP range. Here are instructions for excluding internal traffic in Google Analytics with IP filters.
Advertising Campaign Tracking
It’s important to ensure that all traffic sources that are under the control of the site owner are tagged with UTM campaign tracking tags. That way, you can see which sources are performing the best. Start with referring sites reports, and tag each referral source to which your company has access.
Make sure multiple tags aren’t being used to describe the same parameter. In this screenshot, there are three different naming conventions for the “ecommerce Content Audit” campaign we ran earlier this year. This makes your reporting much more difficult.
If you are running Google AdWords campaigns, use the automated linking feature to Link GA and AdWords because it will also import additional information about clicks, cost, ad position, and other metrics.
Ensure Goals and Funnels Are Set Up Correctly
Each of your sites should have at least one macro goal and a couple of micro goals. An example of a macro goal would be the completion of a purchase or lead form. An example of a micro goal would be adding a product to the cart, signing up for a newsletter, engaging with live chat, etc.
Where applicable, make sure to have goal funnels. Even a funnel with one step is better than none. For example, Step 1 could be viewing a product page and the goal completion could be selecting “add to cart.” With that setup, you’ll have the data neatly displayed in one place and won’t need to gather data from two reports to see conversion/drop-off rates.
In addition to ecommerce Tracking, retail sites may want to track “checkout” as a goal in GA.
Most conversion funnels are more complicated than this one. Visualizing this funnel in GA is especially important for sites with multi-page checkout processes and post-transaction upsells.
Payment Gateway Referrals
In many cases, a payment gateway, such as Paypal, is set up in a way that causes traffic coming back to the site to start new sessions with a referral – the domain of the payment gateway, e.g. paypal.com.
You don’t want that to happen because payment gateway referrals keep you from seeing the real domain behind that sale.
This doesn’t happen with Classic GA, but it does happen with Universal GA code, so you need to watch out for this if you’ve just updated your Google Analytics code.
To tell if payment gateway referrals are a problem for your analytics setup, review your traffic sources by the number of transactions, then descending order, and then check if your payment gateway appears in the list as a source of transactions. If it does, the best course of action is to add it to the referral exclusion list in the property-level settings screen.
Basic Ecommerce Tracking
The most important and basic things you should be looking for are the number of orders, number of items sold and revenue, and how well they match backend data. Small (1-2 percent) discrepancies are tolerable and sometimes unavoidable.
Also, make sure your cart backend data and Google Analytics share the same time zone.
You’ll want to compare all orders received, both as seen in GA as well as your cart backend, not just the ones you fulfilled successfully. If there are discrepancies, start reviewing each transaction by transaction until you identify the culprit.
In all cases, be certain to send as much data to GA as you can (e.g. product category, product brand, and more). You’ll thank me later when you want to estimate if your blue widgets promotion actually drove up blue widget sales, or if it also drove up red and pink widget sales.
Enhanced Ecommerce Tracking
If you have a simple enough e-shop or there are no plugins available, then you may be satisfied with just the basics. Otherwise, I’d recommend considering enhanced ecommerce tracking, which gives you so much more insights, including product impressions, clicks, checkout funnel performance, internal promotions performance and even refunds tracking.
Auditing enhanced ecommerce is quite tricky simply because it is so complex and it permeates the entire site. If you notice discrepancies, say product impressions are less than product clicks or product clicks are less than product add to carts, you’d most likely need to resort to the Tag Assistant’s recording function or the Chrome GA Debugger extension in order to debug.
Shopify supports both Basic and Enhanced ecommerce tracking and uses the new Universal Google Analytics code by default. Here are instructions for setting up enhanced ecommerce tracking in GA for Shopify sites.
About the Author: Everett Sizemore is Director of Marketing at Inflow, an ecommerce marketing agency providing SEO, PPC, CRO, and inbound consulting and implementation services to retail brands. Follow Everett on Twitter @balibones for more ecommerce news and tips.