How Does a Search Engine Interpret UX?
With the implementation of new innovations in search, like semantic, more complex relationships and interactions between users and search engines arise. No longer do we have to use quotations, “and,” or other symbols to direct search engines – they have evolved to understand user intent. Along with this greater understanding, comes a desire to go beyond simply comprehending the types of pages with relevant content to serve up, but also which pages and results will provide the best experience for the searcher.
The user provides many signals to a search engine, giving feedback on how engaged they are with a website. Search engines track user behavior data – things like session duration, frequency of visit, CTR, navigational paths, and type of access. Through an analysis based on these factors, along with others, search engines are able to gauge UX.
Let’s dive into these signals, and explore their implications within SEO.
One of the most obvious, yet tricky, behavioral signals is CTR. A successful search query provides results to the user, deemed interesting and relevant enough to their search, to click on. The trouble for search engine providers is, that while CTR is a clear indication of interest, the position of the result itself plays a huge part in the click.
Consider the CTR vs. Page Rank study by Philip Petrescu of Advanced Web Ranking, aka the stuff SEOs’ nightmares are made of.*
(Note* this study was conducted in July, 2014, and specifically relates to organic desktop searches in Google.)
As this study shows, CTR is highly determined by where the result ranks within the SERP.
Think about it, how often do you click on a page 2 result?
The challenge for search engines is to determine how to interpret CTR, a strong user behavior signal, without skewing the metric due to “position bias”.
Microsoft determined that CTR can be considered, even with the problem of position bias, when explored through the cascade model. Google has mentioned CTR as an indicator of how it interprets user intent in relation to ranking.
SEO wiz, Rand Fishkin, has extensively explored the idea of long clicks vs. short clicks being a signal to search engines. Arguing that if a user clicks on a result, and then immediately clicks off of the page, it’s a good indication that the user is unhappy with that page for the search term. For the nay-sayers out there, who say that long vs. short clicks don’t matter enough, Rand is quick to respond:
“Is there any reason that you wouldn’t optimize for a higher click-through rate? Is there any reason you wouldn’t optimize for longer clicks versus shorter clicks? Is there any reason that you wouldn’t optimize to try and get more branded search traffic, people associating your brand with the generic term? No way.”
In a person’s decision to click on a search result, ranking does factor in, and there are other factors that affect the choice as well. Additional elements to inciting a click include: Brand Recognition, Meta Description, Special Results like Google’s “Knowledge Box,” Perceived Quality, Personal Relevance, and others.
Session Duration, Session Frequency, URL Access, and Trajectory
John Mueller, of Google, notes that Google does not take into account user engagement via forms filled out on page, clicks on page, or time on page, as direct ranking signals. At the same time, he’s also mentioned that Google may pursue active learning for future understanding of user intent and experience. See his full response to the questions of user engagement and SEO in this Google Hangout.
While not all user engagement signals may be indexable, they can still play a role in page engagement. Afterall, session duration is a key indication of interest, even with the problem of idle pages. Implicit indicators, like mouse hovers, play a role in user monitoring systems.
Google’s monitoring system – also known as Google Chrome – processes user behavior such as maximized windows, opened tabs, and fetched URLs.
User Behavior Data and Rankings
User behavior data, links, and URLs, all play a role in signaling user engagement to Google. Strong links, to popular pages with long duration visits, can provide a boost to the page that links. This encourages strong content, and return visits. Google monitors return page views and frequency of page views. Implicit interactions with a site, while small, do make a difference.
Google notices when a user logs in/out of a site – think Facebook.
A login page can become an indicator of a commonality among many searchers and many sessions, prompting it to rank higher.
Using an equation called The PageRank Damping Factor search engines are able to hypothesize that each time a user clicks on a link, they are less likely to click on another followed link, and will eventually stop clicking on the search path.
Looking at onpage signals like: link location, font size and style, number of links on a page, position of link on page, etc, search engines can determine the probability of a link being clicked. Google has mentioned the monitoring of clicks to assist in gaining a better view of user experience, as links are inherently more manipulative.
Along with links, clicks, and logins, URL data is also an indicator of user behavior. Looking at the kind of URL, an exact URL written into the nav bar, a URL accessed via a hyperlink, or a URL accessed via an email, are all signals to search engines. Each has a different weight, with typing the URL into the nav bar being the strongest signal of intent.
While user behavior data, and implicit data currently play a smaller role than other SEO ranking factors, they are important to keep in mind as search engines advance and adapt to UX needs. Understanding how users interact with websites, and how search engines like Google interpret those interactions, will improve UX and assist with SEO.