While Google Analytics is an intuitive platform that can be leveraged by users with little analytics experience, it also offers many advanced tools and features that require extensive, in-depth knowledge to be fully utilized. Furthermore, there are nuanced rules governing how Google Analytics collects its data, which then impact how that data is presented. These rules and assumptions make it so you cannot and should not always accept Google Analytics data at face value, unless you fully understand how that data is being collected. This guide is intended to clarify some of the more confusing aspects of why data might appear to be counter-intuitive or inaccurately reported.
Learn What Causes Growth in “Direct” Traffic over Time
At this point, it has been well established that a significant amount of traffic that is not actually “direct” is placed in the “direct” bucket in Google Analytics. This happens when Google is unable to identify the referring source, which is a common occurrence, as pointed out in a previous blog post of mine:
- Several examples of sources where Google is unable to identify the referrer include:
- Referrals from secure (HTTPS) to non-secure (HTTP) websites
- Referrals from emails
- Links listed in mobile app descriptions
- Social referrals with no custom URL tracking
- Referrals stemming from chat programs such as Facebook Messenger and Google Hangouts
While this is not a new occurrence, it is causing a slow but steady growth in direct traffic for many users of Google Analytics. Here is how “direct” traffic looks for a couple of our own clients over the past year-and-a-half, compared to the prior period:
Why is this increase happening? A few reasons:
- More and more websites are becoming secure (HTTPS), and Google Analytics cannot identify referrals from HTTPS to HTTP, so these are grouped into the “direct” category.
- Increased use of link shorteners and mobile apps, whose links are picked up as direct traffic
- Browser issues on various devices causing organic search to appear as direct.
- A study by Search Engine Land demonstrated that up to 60% of direct traffic may actually stem from organic search
How can you tell where direct traffic is truly coming from? It’s not entirely possible to do a better job of ascertaining the referral source than Google Analytics is already doing, but it can be helpful to analyze the actual landing pages when you are analyzing direct traffic data. If it seems unusual that deeper pages on your site are receiving large quantities of direct traffic (because the user likely wouldn’t enter long and complicated URLs directly into his or her address bar), it’s likely that this traffic is actually coming from referrals or organic search.
“Organic,” “Referral,” and “Campaign” Take Precedence over “Direct” Attribution
One hugely important but not commonly understood aspect of Google analytics is the relationship between direct, organic, referral, and campaign traffic when a client visits the site more than once. This article does a great job of explaining these relationships in detail.
Here are some general rules of thumb:
- If a user starts as an organic visitor, and then re-visits your website via direct traffic*, he or she will be tracked as an organic visitor.
- If a user starts as a direct visitor, but then revisits the site from organic search*, he or she will be tracked as an organic visitor.
- If a user starts from a referral, and then revisits the site from organic search*, he or she will be tracked as an organic visitor. Conversely, an organic visitor who revisits from a referral will be tracked as a referral visitor.
*The timeframe specified here is whatever your “campaign” is set to, which is 6 months by default in Google Analytics. See below for more information.
Understanding Session & Campaign Timeout Settings
For Universal Analytics users, there is the option to specify the amount of time that defines a session and a campaign. By default, a session is 30 minutes (in other words, a user will be counted as a new session after 30 minutes passes from their first visit), and a campaign is 6 months.
It’s a good idea to think about the content of your website when determining how long your sessions and campaigns should be. For a website with a ton of content where the average visit may be more than 30 minutes, it can be helpful to lengthen the session time to get an accurate reading of how much time users spend on your site. Conversely, if your site has a much lower average time spent on site, it can be helpful for getting a better read on actual user behavior to shorten the session time.
User Timeouts and Comparison Shopping Can Affect Landing Page Data
Have you ever looked at your landing pages from organic search and noticed that users are entering in from pages deep on your site that can’t possibly be receiving large amounts of organic traffic? For example, pages in your shopping cart? (See the chart below.)
To view your own highest converting landing pages from organic traffic, click Conversions -> Ecommerce -> Overview on the left sidebar, then click Source/Medium -> View Full Report, set your Primary Dimension to Medium, click organic, and then select Landing Pages under Primary Dimension.
This chart shows examples of top converting organic landing pages. How can hundreds of thousands of organic visitors be landing on our “cart,” “track your orders,” “checkout,” or “order status” pages via organic search? They aren’t ranking for anything and they are blocked by our robots.txt file – so what’s going on here?
The answer lies in the rules we have reviewed in previous section of this post. The rules GA establishes for session timeouts have significant impacts on what Google Analytics perceives as the “landing page” for visitors who have come to the site via organic search, and returned after their session timed out.
When you’re online shopping, it’s normal to get to the point where you are ready to make a purchase, and then decide to do some comparison shopping – clicking around on other sites to find the best deal, or perhaps reading reviews of the product you are about to buy. Let’s assume you leave your shopping cart browser window open while you browse to other tabs, and this process takes more than the specified session time in Google Analytics (i.e. 30 minutes). Or perhaps you leave the shopping cart window open while you step away from your computer entirely for 30 minutes. When you return to the shopping cart of the original site, it generally causes the page to re-load. This is what triggers an organic visitor to “land” on a landing page that wasn’t actually the page they originated from on a search engine; it’s actually the URL of the page they left open on their browser and revisited after their session timed out.
Once you understand the rules listed above, you can analyze your data in a new lens and understand why, with Google Analytics, things are not always what they seem.