Does Google use engagement signals to rank web pages?
Certainly yes. Google even says so in their official How Search Works documents:
Exactly how Google uses engagement signals (i.e., clicks and interaction data) is subject to endless SEO debate. The passage above suggests Google uses engagement metrics to train its machine-learning models. Google has also admitted to using click signals for both search personalization and evaluating new algorithms.
When pressed for specifics, though, Google typically responds with either forced denials (“We’re not using such metrics“) to carefully worded deflections (“clicks are noisy.”)
While many Googlers no doubt work hard to be helpful to the SEO community, they are also under pressure “not to reveal too much detail” about their algorithms out of caution that SEOs will game search results. In reality, Google is never going to tell SEO exactly how they use engagement metrics, no matter how many times we ask.
Most SEO debate focuses on whether Google uses organic Click-through Rates (CTR) in its ranking algorithms. If you are interested, AJ Kohn’s piece is particularly outstanding, as well as Rand Fishkin’s Whiteboard Friday covering this topic. For a nuanced counter-view, I’d recommend reading this excellent post by Dan Taylor.
To be fair, I believe most of the debate around CTR up to this point has likely been far too simple. Whatever way SEOs think Google uses click data, how Google actually uses clicks is guaranteed to be far more sophisticated than anything we may conceive. This complexity gap gives Google easy deniability and justification for calling otherwise reasonable SEO theories “made-up crap.” (Google may very well say something similar about this article, which is fine.)
Not another CTR debate
At this point, you may think this is another post adding to the CTR debate, but in fact, it’s not. THIS SIMPLY ISN’T THAT POST.
Arguing “if” Google uses click signals leads us down the wrong path. We know Google does, we simply don’t know how. For example, are they direct signals, or used for machine learning training only? Are click signals used in the broader algorithm, or only for personalization?
Instead, lets propose something far more radical, and likely far more helpful to your SEO:
Why you should assume Google uses clicks for ranking
Not too long ago, Google patent guru Bill Slawski posted his discovery of a newish Google patent that described “Modifying search result ranking based on implicit user feedback.”
The patent is fascinating from an SEO perspective because it explains how using click signals can be very “noisy” (as Google often says) but describes a process for calculating “long click” and “last click” metrics to cut through the noise and better rank search results.
To be fair, we have no evidence Google uses the processes described in this patent, and even if they did, it would likely be far more sophisticated/nuanced than the process described here.
That said, the patent is riveting because it supports many of the same best SEO practices we’ve advocated for years. So much so that, if you optimized for these metrics, you’d almost certainly improve your SEO traffic and rankings, regardless if Google uses these exact processes or not. Specifically:
-
More Clicks (“High CTR”): earns you more traffic no matter your rank, and initial clicks form the basis of all subsequent click metrics.
-
Improved Engagement (“Long Clicks”): almost always a positive sign from your users, and often an indicator of quality as well as being correlated with future visits.
-
User Satisfaction (“Last Click”): the holy grail of SEO, and, ultimately, the experience Google strives to deliver in its search results.
We can summarize these principles into 3 tenets of click-based engagement metrics for SEO: First, Long, and Last.
Let’s explore each of these in turn.
1. Be the first click: earning high CTRs
As stated earlier, this isn’t a debate if Google uses CTR. There’s plenty of evidence that they monitor and consider clicks in a variety of ways. (And to be fair, there’s evidence that they don’t use CTR as extensively as many SEOs believe.)
As the Google patent US8661029B1 states:
Even if CTR isn’t a ranking signal, having a higher CTR is almost always good for SEO, because it means getting more clicks and more eyeballs on your content.
Besides the inherent value of earning a high CTR, clicks also form the basis of subsequent click-based metrics, including long clicks and last clicks. So earning that first click is an essential step.
How to earn higher click-through rates
Your ability to earn a higher CTR is almost entirely contained in optimizing your appearance in Google search results. How your snippet stands out and gets noticed for being a likely helpful, relevant answer—in a sea of other competing results—is the name of the game.
You may think your options for influencing CTR in this way are quite limited, but in fact, you have many, many surprisingly powerful levers to pull in your favor, including:
-
Compelling, relevant Title Tags (My Master Class, definitely worth a watch)
-
Compelling, keyword-rich Meta Descriptions
-
Structured Data & Rich Snippet Markup
-
Keywords-rich URLs, which Google may use as breadcrumbs
What about artificially manipulating your CTR, either using bots or one of the many blackhat click services you can find on the web? More often than not, these tactics lead to disappointing results. One possible reason is that Google is very skilled at sniffing out “unnatural” browsing behavior.
So high CTR can be a good thing, but the fact remains—as Google has told us countless times—CTR is a “noisy” signal to use for ranking. Should a result with a flashy title be rewarded simply because users click on it, even if the actual page provides a lackluster experience?
In truth, while earning clicks is one of the primary goals of SEO, the “noise” of the signal is probably why Google avoids using CTR as a direct ranking signal itself.
In fact, earning a high CTR if your content leads to a poor user experience may actually hurt you in the end. More on this below.
So first, we need to figure out if our clicks create a good user experience. Read on…
2. Earn long clicks
So what if you trick people into clicking your URL, but your page actually doesn’t deliver what you promised, or even adequately answer the query?
This isn’t good for users, or for Google. And it definitely isn’t good for you.
One measure of content relevancy search engines can use is weighted viewing time, based on the concept that users typically spend a bit longer time on a site they find relevant, versus a page they find not helpful. Within this framework, “long clicks” can carry more weight than “short clicks.”
The patent explains it like this:
“But Cyrus,” smart SEOs protest, “not every query needs a long click. Many searches, like the weather or the “highest mountains in Europe,” can be answered very quickly, often in seconds. It doesn’t make sense for these pages to have long clicks.”
Those SEOs are right, of course. Fortunately, Google engineers understood not every query is the same and devised a clever solution: click scores can be weighted on a per-query basis, including language and country-specific click data.
“Note that such categories may also be broken down into sub-categories as well, such as informational-quick and informational-slow: a person may only need a small amount of time on a page to gather the information they seek when the query is “George Washington’s Birthday”, but that same user may need a good deal more time to assess a result when the query is “Hilbert transform tutorial”
To dive a little deeper, it’s not so much how long visitors stay on your page, but your ratio of long clicks (LC) to overall clicks (C), weighted on a per-query basis. This LC|C ratio could be used to re-rank queries based on user engagement.
Take this a step further: results with good long-click ratios may rank higher, while results with poor long-click ratios may rank lower.
So consider a situation where you “hacked” your CTR to earn more clicks, but the page itself doesn’t deliver, resulting in more short clicks. In theory, this could actually hurt your rankings, even though you started with a higher CTR!
So be sure to back up your higher CTRs with great user experiences, e.g., long clicks.
How to optimize for long clicks
Many SEOs refer to long clicks as analogous to improving your “dwell time” or simply the amount of time a user spends on your site. The signals associated with improving dwell time are often known as “UX” (User Experience) signals.
The golden rule of getting more long clicks is simply this: provide the most useful, complete, and engaging answer to a user search query, in the most attractive and effective format possible.
A note of distinction: because most pages rank for multiple keywords, and multiple keyword variations, all with possibly varying search intent, it’s often helpful to target for those various search intents all on the same page.
For example, a user searching for information about meta descriptions may also be interested in “meta description length”, “meta description format,” and “how to write meta descriptions.” Optimizing more completely for these varying search intents can improve your long-click metrics.
Pro Tip: You don’t need to optimize for every user intent on the same page. Linking to other resources on your site is fine, and even encouraged! Visitors don’t have to stay on the same page for a search click to count as “long.”
Aside from the quality of the content itself, there are a number of UX factors you can employ to encourage your visitors to engage with your content at a deeper level. While not an exhaustive list, a few examples may include:
-
Have a clean, easy-to-use navigation
-
Make your site easy to search
-
Place important content above the fold, where it’s easy to find
-
Leverage high-quality videos (Moz’s Whiteboard Friday pages have an average view time of nearly 10 minutes!)
-
Strive for 10x Content
-
Use attractive, modern design
-
Prominently link to closely related topics to cover multiple searcher intents. These can be internal links or even external links.
Admittedly, there aren’t a ton of good excellent resources published on increasing engagement and improving long clicks. That said, I believe Brian Dean of Backlinko does an excellent job with this, and his resource on improving dwell time is worth checking out.
3. Be the last click
Yes, being the last click may be the holy grail of SEO.
A user clicks their way through a page of search results, not finding what they are looking for. Finally, they click on your URL, and behold!…. You have the answer they sought.
It means you’ve satisfied the user query.
Put simply, being the last click means searchers don’t return to Google to select another result (e.g., pogo sticking.)
Even if Google doesn’t use this as a ranking factor, you can see how it might benefit your SEO to be the user’s last click as much as possible. Satisfying the user query means users are more likely to browse and share your content, as well as seek you out again in the future.
How to be the last click
In my own SEO, there are fewer things I’ve seen associated with greater success than improving visitor satisfaction, and this is exactly what Google seeks to reward.
It’s also damn difficult to achieve.
Sadly, a typical process in SEO is to give a content brief to a copywriter, expect them to cover all the salient points, hit publish, and hope for the best. But more often than not, do you believe this content truly deserves to rank #1? Is this the first, last, and only result a user needs to click?
Years ago, when working in a successful restaurant, a manager gave me advice about delivering 100% customer satisfaction that I will never forget: “Whatever happens, make sure they want to come back.”
This is how you should treat SEO: make sure every visitor to your site wants to come back.
Exactly how to make sure your visitor wants to come back is going to vary based on each and every query, but generally, it means going the extra mile, answering questions more completely, and offering the user more resources and a better experience.
In short, deliver an experience superior to every one of your competitors.
Beyond this, I recommend these 3 resources when improving your content (all amazingly from Rand Fishkin):
Metrics for click-based engagement signals
To be honest, it’s nearly impossible to accurately measure click-based signals, as Google holds all the data.
(Even if you could accurately measure your long click/click ratio, or last click metrics, calculating their actual value would be meaningless without an accurate account of every other Google search result, let alone on a per-query basis.)
That said, there are metrics that can help you directionally measure any progress you might make. These are all available either through Search Console or Google Analytics:
Keep in mind that there is no such thing as a “good” score for these numbers, as everything is relative to the specific query it appeared for, as well as every single one of your competitors.
Regardless, these metrics can be directionally useful indicators when making improvements to your content. For example, if you see a drop in bounce rate and an increase in session duration after a major content update, you can take this as an indicator that things are moving in the right direction. In fact, it’s not unusual to see an increase in rankings/traffic after such a change accompanied by a positive shift in metrics.
While we can’t directly see what Google might measure in terms of complex click metrics, we can often make educated guesses.
And even if Google isn’t using these metrics exactly the way we speculate, we can still improve our SEO by paying attention to the user click behaviors we have influence over.
Thanks for making it this far. Remember:
-
Be First
-
Be Long
-
Be Last
Get those clicks, and earn them!
Appendix A: Evidence of Google using click-based ranking signals (incomplete list)
-
Google Posts That Local Results Are Influenced By Clicks, Then Deletes That
-
How Google Interferes With Its Search Algorithms and Changes Your Results
-
User Behavior and Local Search – Dallas State of Search 2014
-
Is CTR A Ranking Factor In Organic Results? (Negative result)
-
Queries & Clicks May Influence Google’s Results More Directly Than Previously Suspected
-
Test points to likely influence of click-through rate on search rankings
-
Google Brain Canada: Google Search Uses Click Data For Rankings?
