Being “data-driven” is a goal we hear more and more often. When executed well, a data strategy can be a strong business asset and a competitive advantage over other organizations. In order for you to realize the benefits of being data-driven, you first have to be confident in your data. If your data isn’t accurate, being data-driven becomes a risk rather than a competitive advantage.
Keeping in mind that best practices are a starting point and you’ll need to layer in organizational nuances, environmental factors, audience research, and critical thinking, you can follow the steps outlined below to build confidence in your data.This is by no means an exhaustive outline of our process or a one-size-fits-all approach, but it’s a great place to start exploring how your data is being collected and processed - two key elements of a data strategy.
Knowing what data is being collected and how it’s being collected is your first step in building confidence around your data strategy.
How is visitor data on your website being collected? Chances are, you're using Google Analytics. You can look in the page source for your clues - either gtag.js or analytics.js (for Universal Analytics) or ga.js (for Classic Analytics). You can also use the Chrome extension Tag Assist. If you happen to still be using Class Analytics, you’ll want to add upgrading to Universal Analytics to your roadmap to enhance your tracking capabilities.
It is important to ensure the code is on all pages, including 404 pages and your mobile pages. Once you’ve verified that the code is on the site, you can confirm that data is being collected by reviewing Real Time reports in your Google Analytics account.
Cross Domain Tracking & Self Referrals
If you have multiple related domains or subdomains, establishing cross domain tracking could be a good idea to better understand how users flow through your sites. You can confirm the tracking is in place by utilizing the Google Analytics Debugger Chrome extension. Open the console and navigate through all of the relevant sites, confirming the Client IDs and Tracking IDs match. Bonus Tip: if you don’t already have a hostname filter set up in your reporting view, you’ll want to do that to make analysis easier.
I like to test one more, related set of data. Navigate to your channels report in the Google Analytics user interface and confirm that you don’t have self referrals in the referral channel (a referral that originates within your own domain).
If you are using other Google tools like Google Ads or Search Console, we ensure these products are linked to your analytics to include relevant data into your data set. This is confirmed within the property settings of your Google Analytics Account.
Based on the parameters you set, Google will process your data prior to it being displayed in the reports. Understanding how data is processed gives you much needed context to what the data actually means. For example, if you have an overall bounce rate of 85%, it’s probably throwing up some red flags. Only 15% of your traffic is visiting more than one page on your site! But what if you realize that spam traffic isn’t being processed out and is your biggest contributor to the bounce rate. When it’s filtered out of your data all of a sudden your bounce rate could drop to 65%…much better! Without the context of spam traffic you might find yourself on a wild goose chase trying to figure out why so many people are bouncing.
Understanding how the raw data collected from your site is a necessary element of a data strategy you can be confident in. Data processing is a big topic, the following considerations can get you started:
A reporting view in Google Analytics is a subset of an Analytics account property. Each view can have it’s own configuration, allowing you to process the data to best fit your use case. At a very minimum, you’ll want to have a few unique views for some extremely common use cases. An “All Data” view that serves as a back-up; no configurations are added here. This is raw data and should not be adjusted in any way. Name this view in a way that indicates it shouldn’t be updated - you can be as straightforward as “DO NOT UPDATE - All Data view”. If anything goes awry with the other views, this is your back-up. We also recommend that you have a main reporting view that includes only production data, a testing view that is a duplicate of the reporting view, but can be used to test configurations prior to adding them to the main reporting view, and a development view that excludes production data and includes data from lower environments of your website (dev, staging, test, etc.). After those four views, you can use the remaining 21 slots in your free Google Analytics account for use cases unique to your organization.
Once we review the views added to the account from a high level, we review the settings within each view.
Goals are set at the View level and often a consideration in the creation of unique views. In our auditing process, we confirm the goals are configured appropriately and we review the site to ensure we are capturing all of the goal opportunities.
We are often asked when to use a goal versus an event to measure engagement with the site. Typically we prefer to see goal slots used for business level actions, things we are confident lead to conversions (form submissions, newsletter sign-ups, demo sign-ups, account creations, important downloads etc.) while events are engagements with the site (scroll tracking, outbound links, video views, etc.).
Traffic coming from bots, spiders, and crawlers is going to happen, but you can do something about it! Google has a bot filtering solution that uses IAB’s International Spiders & Bots List to combat the issue, but you have to proactively turn it on in every View. When you establish a View, be sure to tick the “exclude all hits from known bots and spiders” option during setup. If you don’t already have this set up, you’ll likely see a decrease in traffic when you turn it on (that’s a good thing!). If you continue to suspect spam issues, you’ll likely need to explore custom filters to remove the bad data from your data set.
If your site has an internal site search, the tracking for this is also configured in the view settings. We ensure it is turned on, the correct query parameter(s) is in place and that data is being collected in the site search report.
The More You Know...
Data is one of the disciplines where Aristotle’s famous quote rings true:
The more you know, the more you realize you don’t know.
There is a fairly low barrier to entry but you can quickly find yourself in the deep-end. Keep swimming, the value of a data driven approach built on a sound data strategy is immense. If you’re feeling in over your head, we’re always here to help.