Whether UX signals like clickthrough rate, time on site, etc. are used as ranking signals has long been a discussion among SEOs – myself included. The idea is basically that if your site gets clicked in a search results it would be viewed as a point in its favor, if people tend to stay a long time on your site – there’s another point, etc.
The specific question asked of Gary Illyes on Reddit was:
“Can you please confirm/deny whether RankBrain uses UX signals of any kind?”
And Gary’s reply was:
“Dwell time, CTR, whatever Fishkin’s new theory is, those are generally made up crap. “
There are two things at play here. Let’s start with the fact that the question is specific to RankBrain. Lyndon had asked if RankBrain took UX signals into consideration. It’s possible that the answer is specific to RankBrain and not the ranking algorithm as a whole.
I tend to think Gary is making a broader statement however, taking the UX component of the question and stating that Google does not use this as a ranking signal. Which brings us to the second issue … they’ve already said they do.
I’m not saying Gary is lying, in fact he’s telling the truth … technically. They do not use these types of metrics for a specific search result relative to a specific site. But they do look at it, and it is relevant to your SEO efforts.
Before we dive into how that works let’s take a moment to listen to Google’s John Mueller answer a question about pogo-sticking (clicking through to a site and then back quickly) back in July that gives us the information we need:
The part of the answer that applies to what we’re talking about here is when he says:
“… that’s something we look at across millions of different queries, and millions of different pages, and kind of see in general is this algorithm going the right way or is this algorithm going in the right way.
But for individual pages I don’t think that’s something worth focusing on at all.”
So what Gary said is consistent with what John is saying. Let’s assume for a moment that both of them are accurate in their assertion and that signals like CTR and dwell time are not taken into account on individual sites or cases, they are clearly looked at from a search results quality perspective as a whole.
So let’s think about it from that perspective instead, which actually leads us to the same conclusion:
If specific sites/pages do not perform in a way that is viewed as favorable by Google, let’s say – of the top 10 there are 8 sites that when clicked through to, have a 3 minute time on site and 2 that result in pogo sticking and some function of the algorithm sitting on top indicating that for this query type, a longer time on site is expected. What do you suppose is going to happen?
I can tell you what’s going to happen, the sites are not going to be devalued based on this metric (if we’re assuming Gary’s statement is accurate). No, the algorithm itself will be adjusted to favor sites that have characteristics that match the 8 that are deemed successful.
Basically, it’s not that Google is penalizing your site for it’s poor user metrics, it’s just going to feel a lot like it. What they’re actually doing is adjusting their algorithm to favor sites it views as likely to satisfy THEIR users.
Remember folks, it not always in specifically what the fine folks at Google say literally that matters, it’s what they’re not saying, not including, or their context and limitations around what they’re including in their statement that often results in the insights we need to do our job.
What If You’re Wrong?
Let’s imagine that John is wrong (he’s not) and my interpretation is wrong (it’s not) – in that world, where are you left?
Well, you’ll have just spent a bunch of time improving your clickthrough rate and servicing your users better to improve what they do on your site. Sheesh … what a waste right?
An Aside: Not All Queries Are The Same
I feel it’s necessary to note that not all queries are the same and not all metrics can be viewed as if they are.
As Beanstalk’s Mary Davies points out:
“Dwell time is weird as in some cases the shorter the better, I click through, I get what I want, I leave and start a new task because the process was seamless. I think it’s all about the action taken when I’m back on Google after a click through. That’s where the goods are. That’s what can tell if where I went offered what I was searching for. But Google knows all that, they’re just trying to confuse people …”
There are queries that would imply a longer stick time is necessary for a successful visitor experience. If someone searches “the cause of ww1” they’re probably not going to pogo stick if they find a successful site.
If they search for “weather in detroit” and assuming that it wasn’t just given to them right in the result, the time on the successful page would be seconds. If it took minutes it would generally be unsuccessful (with exceptions).
So as you assess the metrics on your pages be sure to consider the likely intent of the users based on the queries it’s ranking for and determine what metrics are likely to produce a success signal.
And remember, if Gary and John are being accurate it’s not going to directly impact your rankings … it’s just going to seem a lot like it.