1/π§΅
π¨:: Does Google use Bounce Back Rate? ::π¨
*sigh*
It's not even 2 years since I last covered this!
π§΅>>>
#SEO #BounceRate #BounceBackRate #BounceBackToSERP
π¨:: Does Google use Bounce Back Rate? ::π¨
*sigh*
It's not even 2 years since I last covered this!
π§΅>>>
#SEO #BounceRate #BounceBackRate #BounceBackToSERP
2/?
Here's the thread I did at the end of 2021...
But if you don't want to read through that,
I'll do a quick summary here...
>>>
Here's the thread I did at the end of 2021...
https://twitter.com/darth_na/status/1456622801985622022
But if you don't want to read through that,
I'll do a quick summary here...
>>>
3/?
Bounce Rate (BR) is an "Analytics" metric.
Dwell Time is a made up metric.
Click Through Rate (CTR) is a SERP metric.
G have stated none are used for Rankings.
But they "do" use user-behaviour in the SERP,
for various things (such as gauging algo performance etc.).
>>>
Bounce Rate (BR) is an "Analytics" metric.
Dwell Time is a made up metric.
Click Through Rate (CTR) is a SERP metric.
G have stated none are used for Rankings.
But they "do" use user-behaviour in the SERP,
for various things (such as gauging algo performance etc.).
>>>
4/?
So we instantly hit Google's "semantic slide",
where they can answer truthfully/honestly,
and yet mislead at the same time.
Simply having a different name permits it.
Through in that the metrics people keep on asking are simplistic, shallow, limited etc.
>>>
So we instantly hit Google's "semantic slide",
where they can answer truthfully/honestly,
and yet mislead at the same time.
Simply having a different name permits it.
Through in that the metrics people keep on asking are simplistic, shallow, limited etc.
>>>
5/?
There are numerous patents and papers for
"implicit feedback"
(user search/SERP behaviours)
which cover a range of potential uses,
ranging from spotting trending queries,
including algo performance gauging,
through to potential satisfaction proxies.
>>>
There are numerous patents and papers for
"implicit feedback"
(user search/SERP behaviours)
which cover a range of potential uses,
ranging from spotting trending queries,
including algo performance gauging,
through to potential satisfaction proxies.
>>>
6/?
People will likely do better to get the concept of
"search sessions".
This may be a single query, or include multiple variations/refinements,
clicking on "suggestions" etc.
(So long as the primary topic/nature remains the same, it's the same session)
>>>
People will likely do better to get the concept of
"search sessions".
This may be a single query, or include multiple variations/refinements,
clicking on "suggestions" etc.
(So long as the primary topic/nature remains the same, it's the same session)
>>>
7/?
How often do you do a search,
click a result,
decide the page isn't for you,
hit the back button,
click another result
(or click a suggestion etc.)
?
There you go.
>>>
How often do you do a search,
click a result,
decide the page isn't for you,
hit the back button,
click another result
(or click a suggestion etc.)
?
There you go.
>>>
8/?
It's NOT:
* Bounce Rate
(only n% have GA setup, and BR includes closes etc.)
* Click Through Rate
(On it's own, weak/noisy - consider "last click" instead)
* Dwell Time (long/short click)
(Short time + a new search session is not the same as short/long + 2nd click)
>>>
It's NOT:
* Bounce Rate
(only n% have GA setup, and BR includes closes etc.)
* Click Through Rate
(On it's own, weak/noisy - consider "last click" instead)
* Dwell Time (long/short click)
(Short time + a new search session is not the same as short/long + 2nd click)
>>>
9/?
* Pogo Sticking
(The influence would be dependent on whether the search session remains, or is amended etc.)
Each every single metric and shallow reference is flawed, as they can be noisy.
Instead, you need to look at multiple features.
>>>
* Pogo Sticking
(The influence would be dependent on whether the search session remains, or is amended etc.)
Each every single metric and shallow reference is flawed, as they can be noisy.
Instead, you need to look at multiple features.
>>>
10/?
If you've seen the latest tweet by @indexsy
You'll note the phrasing "used to have".
This could be a simple back reference,
or suggesting that it's not used "now".
Personally, I'm willing to bet G still user SERP Behaviours,
but not simply.
>>>
If you've seen the latest tweet by @indexsy
https://twitter.com/indexsy/status/1625049714088611840
You'll note the phrasing "used to have".
This could be a simple back reference,
or suggesting that it's not used "now".
Personally, I'm willing to bet G still user SERP Behaviours,
but not simply.
>>>
11/?
So - do yourself (and everyone else) a massive favour;
don't over-simplify the concept of SERP Behaviours for Ranking.
Terms like "Bounce Rate" have no relevance.
Things like "CTR" are too simple/noisy.
>>>
So - do yourself (and everyone else) a massive favour;
don't over-simplify the concept of SERP Behaviours for Ranking.
Terms like "Bounce Rate" have no relevance.
Things like "CTR" are too simple/noisy.
>>>
12/?
G would have to combine multiple factors to make a strong enough signal to punch through all the noise.
AND ... they'd have to implement protections against click-farms/bots etc.
>>>
G would have to combine multiple factors to make a strong enough signal to punch through all the noise.
AND ... they'd have to implement protections against click-farms/bots etc.
>>>
13/?
This is an age-old concept, and plenty have tried to abuse it ... so G are likely looking at
Quantity over Duration (consistency/trend)
as well as multiple individual factors (returns, position, CTR etc.).
If you want more details:
This is an age-old concept, and plenty have tried to abuse it ... so G are likely looking at
Quantity over Duration (consistency/trend)
as well as multiple individual factors (returns, position, CTR etc.).
If you want more details:
https://twitter.com/darth_na/status/1456622801985622022
β’ β’ β’
Missing some Tweet in this thread? You can try to
force a refresh
γ
