Big data shows the volatility of Featured Snippets
Featured Snippets are still a mystery. In this analysis of over 100,000 queries and 1,200 domains, I show how Google tests and adjusts Featured Snippets.
At the beginning of the year, I invested over a hundred hours in analyzing two huge data sets to understand the impact of zero-click searches and SEPR Features. The analysis served as the basis for over 20 presentations I gave over 2020. In this first of an article series, I explain what fluctuations in the number of SERPs with Featured Snippets in 2019 say about how Google tests them.
I've written quite a lot about SERP Features, specifically Featured Snippets, and how to use them to understand User Intent better. Now, I want to present more data at scale to complete the picture.
In this first article of a series, I specifically want to zoom in on Featured Snippets.
- Google tests Featured Snippets and either shows more or fewer based on the results
- More Featured Snippets tend to appear in the mobile than Desktop search results
- The degree of Featured Snippets in the SERPs and the keywords they show up for vary greatly by country
- Google showed a lot more Featured Snippets throughout 2019, except in the UK on desktop devices
- Most countries and devices show Featured Snippets on 5-10% of SERPs
About the data
I was lucky to get two big data sets at the end of 2019.
One was from RankRanger and consisted of four different SERP Features over 100,000 keywords in 2019 on both, mobile and desktop. The four SERP Features are Featured Snippets, Image Thumbnails, Image Packs, and Map Packs. The markets are Germany, the US, and the UK.
The other data set was from SEMrush and consisted of traffic for more than 1,200 of the largest domains in the world on mobile and desktop for 2018 and 2019.
By combining these two sets, I was able to understand the impact of SERP Features on traffic, how Google handles certain SERP Features in general, and how traffic changed over last 2 years.
Featured Snippet growth in Germany, the US, and the UK
In 2019, Google significantly raised the number of Featured Snippets. A good example is Germany, where the number of SERPs with Featured Snippets increased by 86% on mobile and 37% on desktop. As you can see in the chart below, SERP Features saw two spikes throughout 2019: once in August and once in November.
In the US, the number of SERPs that show Featured Snippets grew by 38% on mobile and 45% on desktop in 2019. In the UK, Google showed 11% more mobile Featured Snippets and -14% fewer on desktop (keep this in mind for later). My interpretation from the data is that Google sees a lot of success from showing "the one, true answer".
Featured Snippet differences between Desktop and Mobile
Google shows more Featured Snippets on mobile than desktop. In markets like the US, we see about twice as many Featured Snippets in mobile than desktop search results.
It makes sense to me: people want faster results on mobile devices instead of browsing around for too long. That's why quick bite formats like stories work so well on mobile.
The correlations between Featured Snippets in mobile and desktop search results are strong: 0.92 in the US, 0.72 in the UK, 0.95 in Germany. If Google adjusts the number of Featured Snippets shown on the SERPs for one device, it tends to do so on the other device as well. However, we see larger magnitudes of change for mobile Featured Snippets.
In the table below, I mapped out all permutations of correlations between devices and countries.
Featured Snippet differences between countries
When looking at the data in the table above and the chart below, you notice that correlations between devices are strong but weak between countries. You would expect Google to collect data across all markets and languages, but they test and adjust locally.
Some correlations between markets are even negative, for example between US mobile and UK desktop (-0.52) or DE mobile and UK desktop search (-0.68). In my humble opinion, that speaks more for local differences than relationships between countries and languages. More Featured Snippets in the US don't mean more in the UK. People have different preferences and search behaviors.
In most countries, Google shows Featured Snippets on 5-10% of SERPs. But there are outliers. In the UK, for example, 20-25% of mobile search display Featured Snippets.
Featured Snippet testing
As I mentioned in the first section of this article, Google doesn't always increase the number of Featured Snippets. Many of us can attest to this anecdotally, and the data supports this: Google tests Featured Snippets. In the Uk, the number of SERPs with Featured Snippets on desktop even shrank by -14% throughout 2019 (see the chart below).
These changes don't come gradually, as you can see in the chart below for mid-October 2019. All graphs I show in this article display weekly and monthly fluctuations.
The machine is testing and learning. I'm curious to explore more data about the impact of seasonality and search demand spikes on Featured Snippets. The data shows that Featured Snippets are not stable, whether we talk about their appearance or which site is ranking in it, and we need to better understand that.
The impact of Featured Snippets on traffic
To wrap this analysis up, I analyzed the impact of Featured Snippets on traffic. This traffic and Featured Snippet data is aggregated, so we need to take it with a good portion of skepticism.
When I compared traffic to the 1,200 largest domains with the share of SERPs with a Featured snippet, I noticed negative correlations. In other words: the more Featured Snippets are shown, the lower the traffic to the top 1,200. It's visible with the sheer eye.
Correlation of Featured Snippets in mobile search and traffic: -0.41
Correlation of Featured Snippets in desktop search and traffic: -0.5
But how can that be? Featured Snippets are supposed to bring in more traffic, not less. After all, I found in my analysis of over 2,000 Featured Snippets that the average click-through rate of a FS is 43%.
Well, the answer is that Featured Snippets significantly skew the click-curve of a SERP. The first result gets the majority of traffic, whereas the other results get much less. Search becomes a winner-takes-it-all game.