(40)
In that Reddit AMA with Gary Illyes, Gary said out right and clear that the theories around Google using UX and behavioral signals such as CTR and dwell time are "generally made up crap." He added "Search is much more simple than people think."
He also took a jab at Moz's founder, Rand Fishkin, who has often produced studies claiming Google uses these types of metrics for ranking. Gary Illyes wrote "Dwell time, CTR, whatever Fishkin's new theory is, those are generally made up crap."
I know most of you don't believe it but Gary, although very sarcastic, is a straight shooter and wouldn't out right say something like this if he didn't believe it to be true.
Google has said for over a decade that these metrics are not used by Google because they are easily spammable and very noisy. Google's messaging around this has been consistent and clear for well over a decade.
Forum discussion at Reddit.
Comments:
gary bullshit
02/08/2019 01:28 pm
If search is so simple why the need for over 200 signals as they claim? Gary talks crap and takes any opportunity to mock webmasters. Perhaps if it really is oh so simple they should stop talking crap themselves and just communicate exactly what it is they do want from webmasters. Oh wait, they do, "just make great content"!!!! LOL LOL LOL. Search IS simple, LINK LINK LINK. (200 signals my ass)
DAVID S FREID
02/08/2019 01:31 pm
It is all important to me... These metrics can make a difference to our clients bottom line when considered by the respective SEO. Just saying...
chaboi_sora
02/08/2019 01:48 pm
U mad bro?
Bill Lambert
02/08/2019 02:12 pm
Don't believe the lies. CTR, Dwell Time & Other UX Signals are heavily intregrated into the Algo. We have a all hands on deck meeting next tuesday. Prepare for a release.
gary bullshit
02/08/2019 02:35 pm
Do you believe Gary bullshit "Bro"?
PM_me_ur_analytics
02/08/2019 02:39 pm
he mad bro
chaboi_sora
02/08/2019 02:46 pm
That's your username, so I assume we're talking about you. In that case, no.
chaboi_sora
02/08/2019 02:47 pm
https://uploads.disquscdn.com/images/0b14dc873c1070acfba9ca7aa53d3ee741106c9d7d8e8eddb956ab59386d8850.gif
BeauxJack_ClaKBaK
02/08/2019 02:50 pm
https://uploads.disquscdn.com/images/b832fdafec92adcf345d32d4524a7ec53a232eb55aba485d3a96ece586c78698.png
Matt
02/08/2019 02:54 pm
"Dwell time, CTR, whatever Fishkin's new theory is, those are generally made up crap." Ouch.
Matt
02/08/2019 02:56 pm
You make a good point. If it's "much more simple than people think." why all the signals? Seems like a waste of time and money.
gary bullshit
02/08/2019 03:22 pm
I'm glad someone in this thread has the sense to see that ;)
Chris
02/08/2019 03:30 pm
Gary just decides what goes where - Simples!
chaboi_sora
02/08/2019 04:42 pm
Sir, this is a Wendy's.
Ammon Johns
02/08/2019 06:24 pm
How simple is it to smile? But your brain has to send hundreds of signals to the many facial muscles, and synchronise them, to make it happen. Something can be 'simple' without being based only on a number of steps you can count on your six-fingered hands.
Adam Heaton
02/08/2019 07:02 pm
Not sure if just me but what I got from his comment is that those things are not relevant to RankBrain, not that they aren't being utilised in other ways by their algorithm, as this came up during a response to a RankBrain question.
Ammon Johns
02/08/2019 07:30 pm
We'd had a pretty long discussion about CTR previously, of course, when Google's Andrey Lipattsev joined Rand Fishkin, Eric Enge, and I for a Q&A session. In that, we not only covered that CTR wasn't used and why, but also, and perhaps more interestingly, some of the factors that might make it look like CTR was being used where it wasn't (burstiness for one). https://youtu.be/l8VnZCcl9J4?t=624
M Pekala
02/08/2019 08:42 pm
Wonder if Rand Fishkin will take the bait?
Riki
02/08/2019 09:17 pm
I give you credit for doing this kind of conversation for 1+ hour..it's exhausting to wait for a concise sentence...I listened for about 10 minutes and got nothing from this discussion whatsoever except yes and no, maybe, sometimes...
Andrew
02/08/2019 09:29 pm
Interesting takeaway by many commentators regarding this single comment (in a very specific context, and using the qualifier 'generally'). It's clear that many SEOs (and others) seem to desire a simpler outlook, which is understandable and fine. It works in most cases. That doesn't mean that an exploration of unknown factors, and even speculation, is "made of crap" or of no use. It's easy to say that "Click Through Rate", a narrow definition, is not a ranking factor - but there's countless potential user signals that could very well be used. "Noisy" signals have been used as key ranking factors for decades (links being the foremost of them), and Google has always responded with half-truths and misdirection in the details. In any case, Gary's attitude has been consistently rude and disrespectful to people many who are making earnest attempts to do their job well. Even if Moz, or any other publication, are "wrong", there's no reason for sweeping generalizations in an attempt to discredit.
glorious_broccoli
02/08/2019 09:48 pm
I notice he didn’t mention bounce rate as defined by fast back button back to serp.
Beats from the bong
02/08/2019 11:32 pm
Lay off THAT shit gor a while. You may get somewhere...
OMG
02/09/2019 01:41 pm
What the BEEJEEZERS is BURSTINESS?
Tomasz Kwiatkowski
02/09/2019 02:37 pm
Exactly.
@Michael_Martinez
02/09/2019 03:19 pm
If they're not measuring CTR or dwell time then they aren't measuring bounce rate. Most Websites shown in the SERPs are never clicked on, so it would give an unfair advantage to a very small percentage of listings to use these kinds of signals for rankings or relevance.
snuffy
02/10/2019 03:04 am
Build a great website and whatever signals needed to rank will develop on their own.
Fishking is a scam!
02/11/2019 02:00 pm
He is a young proud woman on antidepressants, a true liberal cunt, so he's probably drinking soy milk right now to conmfort himself. Xanax user all the way. Fucking cunt is Rand Fishkin, an anti-male!
Larry Kim
02/11/2019 05:20 pm
yep. CTR is such a crappy user engagement signal that Google built their 100 billion dollar adwords algorithm based on CTR. totally makes sense.
The Truth
02/13/2019 05:09 am
Guys, do your own tests, don't believe anything, nore here, nor Google, nor anyone. I've seen many things that could be easily manipulated and Google denies they are a signal. Google wouldn't confirm something that can be manipulated. Of course it matters what kind of proxies you use, is the profile logged in, and many more. It is funny how they openly say some things in their documentation and in these QA they say "it's not a ranking factor". It's in their docs, you just have to read more carefully, usually from other departments.
The Truth
02/13/2019 05:12 am
Let's see the following scenario. You search for something. You open first URL, you don't like it, you click back. You open third URL, you click on it, you scroll a little, you click again back or do another search in the same window. You go to the third result, you click on it, you read it, you browse around their site and then do an action on their site. Do you really think Google wouldn't count that as a ranking signal if it was done by 30 other people for that same keyword and it's related ones? Do you really believe Google Chrome will all its telemetry wouldn't pass even more signals to "verify" this is a real human, being logged in, being an account from a long time and not being a proxy or a bot and do you think Google wouldn't take that into account. Put some brain here, test it for yourself!
Fucktard
02/13/2019 05:14 am
If you sell scammy herbal crap that doesn't actually do anything other than any regular food, then don't blame Google for that.
@Michael_Martinez
02/13/2019 03:15 pm
Since Google cannot collect that kind of data for the majority of listings in any given query, there is no point to integrating the small percentage of clicks-per-query into the ranking and quality algorithms. People "comparison search" in something like 17-35% of queries (estimates vary across the years), which means they intentionally click on more than one result to compare what they find. Hence, using a bounce as a negative signal would create false value.
Janet
08/10/2019 05:32 pm
Could be that Google's engineers are in a bubble - they look for their efforts to be validated rather than what in fact happens. Research by SEOs probably reveals what is happening, not necessarily what Google intends. But to acknowledge that SEOs are using better models than their gazillion highly-paid engineers would be beyond the pale....
Gail Gardner
10/18/2019 03:40 pm
Why can't they collect that kind of data? And if you "comparison search", you have to stay on each page long enough to make a comparison and that probably requires scrolling.
Gail Gardner
10/18/2019 03:50 pm
If a site appearing on page 1 is never clicked on, the sites getting the clicks would have the advantage of continuing to rank. And they probably do; however, it seems to me that if Google prevents sites from consistently ranking, that is a strong incentive to force them to pay more for ads. If any of us can think of something, someone at Google has probably though of it, too. And what ranks all year is different than what ranks during the prime buying periods. Watch during the upcoming holidays and you'll see what I mean. Do a case study on that. I haven't for years, but I bet it still happens. Small businesses rank all year long for their prime keywords, but during the major spending holiday, corporations rank even for products they don't even sell and small businesses disappear. I'd be surprised if it wasn't still happening today.
@Michael_Martinez
10/18/2019 04:03 pm
That kind of data doesn't exist for most Websites in most queries. There is no data to collect. Hence, Google cannot collect it.
@Michael_Martinez
10/18/2019 04:04 pm
"If a site appearing on page 1 is never clicked on, the sites getting the clicks would have the advantage of continuing to rank."
PhysicsForums.com
11/06/2019 03:59 pm
He used the word "generally". aka, They are used.
Gord Collins
12/15/2022 05:18 pm
Content engagement should be a huge factor, but apparently they don't trust their own AI system to assess and filter out that spam engagement. That gives us insight. That neglect makes no sense however if actual, real quality is the goal. It looks like they are sticking with the evaluation of the path to fulfillment and whether based on their theory, that it fulfills user needs. They're more interested in validating their system, and sticking to content they want to surface rather than content users actually want. I'm not saying its political, but more likely it serves their profit/revenue strategy better. After all, giving the customer full satisfaction ruins your business!
Gord Collins
12/15/2022 05:23 pm
Yes, but all that searching is the real path to fulfillment, what Google wants to replace with its "bypass system" with they think searchers/shoppers should see. This ties back into them using AI to hijack the whole natural search process each individual searcher uses. We learn from our own unique inquiries, so now, we not learning like we used to. Google sees our individual creative investigation and engagement as a waste of time.