Before we get into the facts and figures, let's do a quick explanation. Although most people don't phrase their search queries this way, every search on Google is a question... When we search on Google, our intention is to find a solution to a problem or an answer to a question, whether we explicitly formulate that question or not.
Ambiguous queries are implicit questions, and Google's problem is to understand "what question the user is asking."
For example, for the vague search query “home,” the user might be asking:
Where can I find house music?
...and hundreds more
Google’s ability to understand the problem behind the query is vital to its ability to provide the best answer and deliver results that satisfy the user.
Google introduced RankBrain in early 2015. Rankbrain’s role is to better understand the intent of the search query – to more reliably infer the hidden question behind ambiguous searches.
This advanced ability to ‘guess the secret question’ has led to significant changes in Google’s results.
It enabled Google to provide more relevant results
Understanding questions better allowed Google to provide a better list of potential answers. ‘Understanding the question’ brings the capacity to show only relevant answers uk mobile phone numbers database rather than a list of possibilities that encompass multiple interpretations of the query’s intent.
When it can remove ambiguity (guess the implicit question behind the query), Google can provide significantly better results and potentially a single 'best' answer.
It allowed Google to provide more direct answers
If Google has identified a ‘best answer’ (their words, not mine), it can display it as a featured snippet or knowledge panel. But to display a ‘best answer’, it first needs to understand the question correctly.
This is possible and reasonable for open questions (how, what, why, where, who). In an ambiguous query where Google cannot guess the real intent, there is no chance of showing an 'answer' in the form of a featured snippet. As the capacity to 'understand the question' increases, the opportunities to show the 'best answer' also increase.
We first saw Featured Snippets in early 2014, before RankBrain’s official announcement… Even though RankBrain wasn’t the technology that directly enabled Google to launch the feature, it’s safe to say that RankBrain was the driving force behind the rapid expansion of Featured Snippets when a user wasn’t explicitly asking a question.
Allows Google to suggest more relevant follow-up questions to the user
After answering a question, the next logical step is to suggest follow-up questions that include the answer to that question… which is why People Also Ask came along at the same time as RankBrain was announced in early 2015.
Today, as we see below, an open question almost always triggers a list of follow-up questions (aka People Also Ask) – 85% of the time. However, as RankBrain has improved and gotten better at identifying the hidden question behind the query, Google may show PAA / related questions in the SERPs for increasingly vague search queries.
Our numbers for 2019 suggest that this is happening. And common sense tells us that this will likely continue as RankBrain continues to improve over time.
Now for the numbers
We analyzed SERPs for 5 million queries in the United Kingdom between February 2019 and February 2020. We identified 4 SERP types – information, navigation, purchasing, and geographic.
Informational = open questions - how, why, what, when...
Navigation = brand names
Purchase = includes words like buy, price, cheap...
Geo = contains geographical words like near me, where...
Other = all search queries that do not fit into one of the top 4
PAA/related questions presence and growth by device in 2019
It’s notable that the numbers are very similar for desktop and mobile, both in terms of the percentage of SERPs containing PAA and growth over the year.
Half of all SERPs have PAA boxes – representing over 20% growth year over year. Google believes that these suggested questions improve the user experience across all search intents.
This shows that RankBrain is indeed improving and Google is increasingly able to identify the question even if the user is not explicitly asking a question.
SERPs containing Related Question PAA (People Also Ask)
search queries containing Related Question PAA (People Also Ask)
PAA growth by search intent in 2019
Now let's break it down by search intent. The differences are striking. And they also make perfect sense.... From Google's perspective, which aims to please its users in the best possible way.
Informational search queries
Almost all informational search queries (open questions) have PAA boxes (85%). Providing possible follow-up questions to the first question is relatively simple, logical and useful for the user.
The growth here has been small. Even at the beginning of 2019, the percentage of open questions that triggered potential follow-up questions was very high. This makes sense, as the question behind this type of query is, and always has been, the easiest to understand.
As RankBrain continues to improve, this rate is likely to increase slightly through 2020, but it is likely nearing its ceiling.
Mobile and desktop are starting to go neck and neck in this area, and they continue to do so.
