Jump to: What is Google Discover? | Why Discover Matters for SEO | Ranking Factors & Eligibility | Analyzing Discover Performance | Risks & Warnings | Analysis of Over 12,000 URLs Performing in Google Discover | Conference Slides
I presented the following information and slides in my talk at DeepSEO Conference on September 7th, 2021. You can view my full slide presentation at the end of this article.
What is Google Discover?
Google Discover is a product Google launched in 2018 that shows mobile users a highly personalized, AI-driven feed of content from across the web, based on articles and topics they find most interesting. With Google Discover, the user does not need to search for anything. The product is a form of queryless search: the feed of content is displayed based on the user’s interests alone and without any query being entered.
Why Discover Matters for SEO
While Google Discover is not a major discussion point in the SEO industry (for now), it should not be ignored by SEO professionals. Any site can be eligible to appear in Google Discover and it’s not just news publishers who receive Discover traffic; we’ve seen significant traffic from Discover for sites in many verticals, including ecommerce, finance, healthcare, and technology. For some of the clients we work on, Discover drives the majority of organic traffic when compared to organic search traffic from Google.
We have also seen many examples where articles that drive little or no SEO traffic can still perform well with Google Discover. Discover can serve as an additional organic traffic opportunity for topics that can be hard to rank for, such as YMYL (Your Money, Your Life) topics, or where the subject matter is highly specific or esoteric.
For these reasons and more, you can and should consider Google Discover part of your SEO strategy. Although Discover behaves differently than regular organic search, it is still unpaid, organic traffic from Google. With the right strategy, Discover can send significant organic traffic and visibility to your site.
Discover Ranking Factors & Eligibility
Ranking in Google Discover requires publishing content that is highly engaging for users (enough so that Google would determine the content worthy of being displayed in their Discover feed). The article must use a high-quality image of 1,200px or more, must contain a meta tag specifying max image-preview:large (which is done automatically using Yoast/Wordpress), or the article must use AMP.
Currently, any content is eligible to appear in Google Discover. There is no specific “noindex” tag or method of using robots.txt to prevent Google from crawling your site for Discover specifically.
Analyzing Discover Performance
Currently, the only way to analyze Google Discover performance is in Google Search Console (GSC) via the Discover tab in the Performance report. The Discover tab only appears there if your GSC property has received Discover traffic during the specified timeframe. Unfortunately, there is currently no dedicated referral source for Discover in Google Analytics. There is also no Google Search Console API for Discover, so it can be hard to analyze the performance of Discover content at scale.
Risks & Warnings
While Google Discover can serve as an incredible source of organic traffic, it’s important not to create an entire business model out of relying on traffic from Discover. This traffic can be extremely unpredictable and is never guaranteed, as Google has made clear in most of its communications around Discover. Discover traffic is also affected by Core Updates and other Google updates, so it can theoretically disappear overnight.
Google can also send 手动操作 for sites that violate its Discover guidelines. This includes a lack of transparency related to the author, sexually explicit content, violent content, harassing content, or misleading content (including clickbait!).
Be careful to avoid these 手动操作. Google doesn’t always tell you which URLs caused the manual action, which can make it hard to diagnose the root of the problem. To add to this, a significant enough offense can cause your website to be removed from Google Discover entirely.
Analysis of Top-Performing URLs
For this exercise, I analyzed 12,142 URLs from 76 domains that have received traffic from Google Discover within the last 16 months. Below are my findings:
Ideal Headline Length & Word Count
SEOs, editors, and webmasters often wonder what is the ideal word count for an article or the ideal number of words in an article headline to drive optimal traffic.
It is important to remember that word count is not an SEO ranking factor, but in terms of correlation with top-performing articles:
- For Discover articles with over 500,000 clicks in the past 16 months, the average headline word count is 17 words and the average word count is 740 words.
- For top-performing articles in Google Discover driving click-through rates of over 25% over the past 16 months, the average headline word count is 12 words and the article word count is 816 words.
Below are 5 findings related to what it takes to perform well in Google Discover:
#1: Discover Content is highly emotional.
By far the most common feature of articles that perform well in Google Discover is that they tend to be highly emotional and cater to the specific interests of the reader.
Below are some of the topics that received the highest click-through rates in Google Discover for our clients, such as cheating scandals, celebrity meltdowns, celebrity deaths, scandalous celebrity photos, and relatable relationship advice.
Also, some of the wording used in top-performing Discover headlines indicates what type of content and language tends to perform best there:
#2: Clickbait headlines appear to drive traffic, but be careful not to get a manual action.
Google is abundantly clear throughout its documentation that clickbait should not perform well in Google Discover. Below is a direct quote from their article, “Get on Discover:”
Use page titles that capture the essence of the content, but in a non-clickbait fashion.
Avoid tactics to artificially inflate engagement by using misleading or exaggerated details in preview content (title, snippets, or images) to increase appeal, or by withholding crucial information required to understand what the content is about.
Avoid tactics that manipulate appeal by catering to morbid curiosity, titillation, or outrage.– Get on Discover
That said, the data shows that the opposite of this advice appears to be true across many of the sites we analyzed. Below are some of the word structures used in headlines among content with the strongest performance in Google Discover, such as:
- this heartbreaking moment
- summed up in this 1 word
- under this 1 condition
- confirms the rumor
- demanded this 1 major change
- sends a terrifying message
- called out this rapper
Note: the wording has been slightly adjusted/redacted to ensure client confidentiality)
Once again, writing clickbait headlines without delivering the content the user expects to get from the headline can cause Google to send your site a Discover manual action, which should be avoided at all costs. Consider adding the most important information to the beginning of the article to avoid this issue.
#3: What works best in Discover might not work at all in regular organic search (and vice versa).
One of the most interesting aspects of Google Discover is that, even though it is technically a source of unpaid, organic traffic from Google, it behaves a lot more like a social media platform than a search engine. For that reason, many articles that perform well on Google Discover emulate what you might expect to see from top-performing articles on Facebook or Twitter, rather than the topics and questions users type into a search engine.
In some cases, Discover content does well precisely because it discusses a topic the user didn’t even know they were interested in (and therefore, they were not searching for it).
Some examples include pictures of celebrities’ children, celebrities following one another on Instagram, pictures of celebrities with their high school sweethearts, how food impacts sleep, and interesting facts about the world and the environment.
Another fascinating aspect of Google Discover is that, despite Google’s ongoing communications about the importance of E-A-T for ranking in Discover, it appears that E-A-T may be calculated differently for Discover than for organic search.
We analyzed thousands of examples of URLs that received significant traffic from Google Discover, but little to no organic search traffic. In many cases, these stemmed from YMYL (Your Money, Your Life) topics that are generally difficult (or even impossible) to compete for in Google organic search.
Below are some examples of “forbidden” topics where many sites struggle to rank in SEO due to the YMYL nature of the content, but where those same sites saw significant traffic coming to those articles from Google Discover.
For sites struggling to rank due to a perceived lack of E-A-T on YMYL topics, focusing on Discover may be the way to go.
#4: Posing questions in headlines appears to drive high CTRs in Google Discover.
While not every site publishes these types of articles, a trend we noticed among many articles with high CTRs in Google Discover is that the headline is structured as a question, such as:
- Did Trump say ____?
- Does ____ help with anxiety?
- Did ____ get a pay raise?
#5: Content that deeply targets users’ interests performs best in Discover.
The best-performing content in Google Discover speaks to the specific, niche interests of its audience. Topics such as natural phenomena, whether the 2020 election was “stolen,” police brutality, diabetes health, and other highly specific topics drive huge amounts of traffic through Google Discover.
Takeaways & Recommendations
Since there is currently no way to index/not index or crawl/not crawl content for Discover specifically, be careful to not employ any methods to try to drive Discover traffic that could be detrimental to your SEO performance. Remember that Google evaluates the quality of your entire domain for rankings, especially during core algorithm updates, and low-quality content can factor into those evaluations.
That said, consider Discover its own source of organic traffic that behaves more like a social media platform than a search engine. The tactics that tend to work best on Facebook, Twitter, and other social sites may drive more Discover traffic than standard SEO best practices.
Perhaps consider using a multi-pronged content strategy that simultaneously addresses SEO and Discover content using different articles or website sections: one predicated on users searching for that content, and one that speaks to the emotional interests of your audience. Learn more ways to optimize for Google Discover in my latest Search Engine Journal article.
DeepSEO Conference Slides
Below are the slides I presented on Sept 7, 2021 at DeepSEO Conference:
Have a question? Reach out to @lilyraynyc.