SEMrush and Brado analyzed 1 million random SERPs with featured snippets to identify correlations between pages that earn these valuable search positions.
This data can be used by SEOs and site owners to create content that may have a better chance of being chosen for a featured snippet.
First, lets take a look at look at some general stats the study uncovered about featured snippets.
Featured Snippet Statistics
SEMRush and Brado studied 160 million keywords on desktop, and 46.1 million keywords on mobile, to find what percentage of keywords generate featured snippets.
Here’s a summary of their findings:
- 19% of SERPs have featured snippets.
- 7.3% of SERPs have double featured snippets.
- 50% of a mobile screen is covered with a featured snippet.
- 70% of featured snippets are paragraphs (an average of 42 words/249 characters).
- 19.1% of featured snippets are lists (an average of 6 items/44 words)
- 6.3% of featured snippets are tables (an average of 5 rows, 2 columns)
- 4.6% of featured snippets are videos (an average of 6m35s)
- Industries that have the greatest percentage of featured snippets are:
- Computers & Electronics
- Art & Entertainment
Featured snippets appear most often for keywords with a specific search intent, such as long-tail keywords
Out of queries made up of 10 words, 55.5% have a featured snippet. In comparison, only 4.3% of single keyword searches have a featured snippet. The percentage gradually goes up as more words are added to a query, then drops off after 10 words.
Now let’s look at the common factors shared by content earning these featured snippet positions.
Optimizing Content For Google Featured Snippets
SEMRush breaks down the findings of its study and identifies these data-backed methods of optimizing content to earn featured snippets.
Answer Question-based Queries
The study finds 29% of queries triggering featured snippets start with a question-based word, such as “why,” “do,” and “can. Questions that start with “why” trigger the most featured snippets.
Of all “why” queries studied, 77.6% return a featured snippet. Queries starting with “can” have the second-largest percentage of featured snippets at 72.4%.
Date Your Content
Google frequently returns dated content in featured snippets.
Of the different types of featured snippets, these are the percentages that return content with a date:
Google also tends to keep featured snippets current; 70% of articles in featured snippets were published no later than 2-3 years ago.
However, older articles can still earn the featured snippet if they provide the best answer.
Use Subfolders Sparingly
According to the study, long URLs are less likely to earn a featured snippet. The “sweet spot” is a URL with 1-3 subfolders.
Of all featured snippets included in the study, 37.3% link to a URL with 2 subfolders. That’s followed by 1 subfolder at 21.9% and 3 subfolders at 21.2%.
To be clear, subfolders are the parts of the domain listed after forward slashes. As an example:
- domain.com is the root domain. It has zero subfolders;
- domain.com/subfolder has one subfolder; and
- domain.com/subfolder1/subfolder2 has two subfolders.
It’s extremely unlikely that Google will return a website’s home page in the featured snippet position, as only 0.4% link to a domain with no subfolders.
Combine Multiple Questions in One Article
SEMRush recommends creating a “featured snippet hub,” which is be accomplished by answering multiple questions in a single piece of content.
This is said to increase the likelihood of earning multiple featured snippets with one URL.
However, I would be remiss not to mention SEMRush doesn’t provide any particular data to back up its recommendation here.
It would be helpful to see some mention of the percentage of featured snippets linking to a featured snippet hub.
It would also be interesting to know the average number of featured snippets earned per featured snippet hub.
What SEMRush does provide are these statistics about featured snippet hubs:
- 98% use HTTPS
- Content length is upwards of 1,100 words or more
- 66% use structured data
- They contain an average of 14.5 heading tags
- They have an average of 8 images (with ALT tags)
For more on any of the data points included in this article, see the full study here.