How Does Google Discover Report Traffic in Google Analytics?

Google Discover is a feature where Google recommends articles based upon a user’s search history. You might see Google’s recommended articles when you visit google.com from a mobile device.

Another place you might have seen this is in your Google Search Console report- under “performance”. This shows you a graph of impressions and clicks (and CTR) for articles Google has served from your website in Google Discover. Not every site has this data because not every site qualifies to show up in Google Discover.

Does Google Discovery Traffic matter?

When I see graphs like this, I ask one question: how do these visitors help my business? Sure, impressions are nice and clicks are great but is it leading to more business? If they are, then I might want to get more out of this!

To answer this question, I need to turn to Google Analytics. Because I’ve setup goal tracking (or ecommerce tracking on a shop) I know whether visitors become customers. This tells me whether or not that traffic is valuable. Otherwise, impressions (and even clicks) are just vanity metrics.

In order to determine whether or not this traffic is producing business, I need to understand how this traffic is being reported in GA.

Since Google Discover articles are recommended on the Google homepage, you might expect them to show up as the source/medium: Google/referral. After all, it’s a referral link from the website google.com. If you look there, however, you’ll not find any traffic (ever!).

Perhaps this is because google.com isn’t just any website- it’s a search engine. Google Analytics attributes those visits to a special category/medium, “organic”. Now, as an SEO, I get a good amount of traffic from google/organic. How can I distinguish “normal” google/search traffic from Google Discover traffic?

To do this I looked in my Google Search Console. I see a couple articles producing traffic from Google Discover. The graph tells me when those visitors came and to what page they arrived. With a couple clicks, I can download this information and reproduce the graph in Excel. If you download this as CSV you’ll get a zip file. Look in the “Dates.csv” file and you’ll find this data.

To find this data in GA I first went to the Acquisition report and then Default Channel Groupings. Since I don’t know which default grouping I want, I changed the dimension to “landing page”. I took one of the landing pages that Google Search Console said is sending Discover traffic and entered it into the search field- this limits traffic to this page.

From there I set a secondary dimension of “source/medium”. Now I see what the source/medium is for traffic to a specific page. This graph resembled the Search Console Graph- I think I’ve found the source! It turns out Google Analytics is attributing Google Discover clicks to Direct traffic! (Try to repeat this experiment with you own data, and see if you can confirm my theory!)

No, the traffic doesn’t match up exactly. I think there are a few reasonable explanations for this:

• Google Analytics reports data late. It’s not real time. It seems that the Session data is almost always a day after the click data from Search Console. It might be because GSC data is real-time and GA data is delayed.
• Even if you shifted these over a day, they don’t exactly line up. I think you can attribute these slight differences (one or two visits) to factors like the GA javascript not firing (due to an Ad Blocker or technical glitch, for instance). They’re too close to be a coincidence.
• There are several reasons (besides Google Discover) that a page’s source could be Direct (see below) so the data isn’t exactly clear.

There’s so much wrong with this source. As I projected (above) the true source/medium should be Google/referral. It’s not really a search result (because I didn’t manually enter a search to get these Discover articles) but I could understand this being reported as Google/organic. Direct traffic, though? That’s silly.

Direct traffic is really the default setting for traffic in Google Analytics. If GA doesn’t know how to attribute traffic, it reverts to “Direct”. Many of us noticed this when we switched to HTTPS- and our referral source was lost. Apparently, this is what Google Discover is doing in GA, as well.

Why would this report incorrectly?

My hypothesis: Google wants to discourage our attempts to “optimize for Discover.” If there’s clear traffic then we might be able to figure out how to “rank” here. Instead, they want to make this difficult.

The fact is, it’s very strange for an article within a website to see the source of traffic as Direct. There are a couple circumstances for which I can think of that this might occur:

• Someone has bookmarked your article on their browser. In this circumstance, they’re just clicking a link and there’s no real source. It’s a direct visit to an article. This is fairly rare- does anyone use bookmarks these days? (I say that, but then have to admit: I do)
• Someone copies and pastes an article from somewhere. Let’s say you get a link in an email but your email provider doesn’t convert it into a hyperlink. If you copy the link to your browser, it will come in as a direct visit. This is true unless your browser’s address bar allows you to use it as a search engine bar. In that case, Google will provide a search to that URL- and the visit will be recorded as a visit from Google/organic.
• The rel=”noreferrer” parameter in a link will remove referring information to the website receiving the visitor. This will be attributed to Direct traffic in Google Analytics. This is probably why traffic from Google Discover is considered Direct. The latest versions of WordPress add this parameter to links automatically.
• Someone links to your site from the HTTP version of a URL, rather than the correct HTTPS version of that URL. When a link passes between HTTP and HTTPS, the referral source is lost. I experienced this one time when someone used an image from my site on theirs and attributed my site as the source (with a link!). Unfortunately, they linked to the HTTP URL and not the HTTPS. It took me a while to figure this one out!

Why would I care how Google Analytics reports Google Discover traffic?

I believe this is important because it tells us about the quality of our traffic. It’s nice of Google to tell us how many impressions we received from our Google Discover articles. It’s even generous for them to tell us how many clicks we got from them. These are merely vanity metrics, however, unless these clicks because customers.

When I can see the Google Discover data in Google Analytics I learn a couple of things:

• Google Discover traffic performs similar to blog content. By that I mean: people come to the site to find the information they want and likely leave when they’re done. This is clear in the high bounce rates and low time on site- which is pretty typical among blogs.
• With an article being promoted in Google Discover for almost a month, this website only gained one new lead from this traffic. That’s a conversion rate of 0.50%. off of a 6.9% CTR (according to GSC).

I’m not particularly impressed by these numbers. Frankly, I’m not particularly interested in “ranking” for Google Discover in light of this data. I don’t mean to be ungrateful for Google’s traffic- it just didn’t do much for my site. I wouldn’t turn it down but it really doesn’t matter much.