RSS Feed
Twitter
November 27th, 2008 
QuadsZilla After my last 2 posts on bounce rate, Many more questions have been popping up about the bounce effect.
First, let’s clarify what we’re talking about here. Google doesn’t need to use analytics data bounce rates. It’s much more likely that Google would use a metric such as:
SCASI
I don’t think there’s a word for this yet, so I’ll invent the term SCASI (SERP Click After Site Inspection).
If a surfer Clicks SERP A from a list of SERPs and then (10 secs to 4 mins later) clicks any other SERP on that same list of search results it is a negative quality indicator. If the user found what they are looking for, they wouldn’t need to keep clicking on SERPs. If they continue to click, they’re initial find was not satisfying.
The links in search results are not direct links to the sites, but rather a google redirect such as:
http://www.google.com/url?sa=t&source=web&ct=res&cd=14&url=http%3A%2F%2Fwww.holidays.net%2Fthanksgiving%2F&ei=UQIvSfPRHaGievXK5OMK&usg=AFQjCNFNzOBQgfUn_14d33MdAMUPBgYz2Q&sig2=SJhrGbVZlvPXlOT9zYBHgQ
This would allow Google to easily track SCASI . . . and to a great extent “bounce”.
This isn’t technically the “bounce rate” but would be a good corollary in most instances. However, in the event that the landing page actually answers the users questions, the corollary would start to break down.
For example, many people who go to the wikipedia through a Google search will not continue to click in wikipedia. Many of them will “bounce” after reading the article. But if the wiki article answered the user’s question, the bounce might be high but the SCASI would be very low.
As webmasters, we don’t have access to SCASI data . . . unless of course we own all the top 10 SERPs 
What we DO have is bounce rate. In a majority of cases, a higher bounce rate will mean a higher SCASI rate.
It’s clear to me that Google already uses this type of data for ranking sites. If not, they should start sending me a consultancy paycheck for designing their damn search engine . . . after all, I did answer their interview questions. Better yet, let me take over the search department at Microsoft; we desperately need more competition in the search space.
I’ve got more to cover on this – but is it getting any clearer to you?
PS: Happy Thanksgiving!
Posted in 
Dude;
Thanks for clearing this up!!! My main money maker has an average bounce rate of 67% according to GA, but I rank well in the SERPS – so there can be no corollary, unless I’m on the cusp. Ii was starting to get worried my relatively high biunce rate would soon haunt me…. But this explanation makes total sense – especially since my goal is to get users to “bounce” ASAP by clicking an AdSense ad (another G metric). Now I see as long as users don’t perform a SCASI, I’m good, my high GA “bounce rate” is also good – for the pocketbook.
I think you’re definitely on target with SCASI.
Your theory is getting better.
imho there’s a global “user experience quality rank”.
Form time to time, Google moves Ranks to test different website’s UEQR :p
Quite an interesting theory, you’re definitely on the right track. One thought though: what about the scenarios where the user intends to visit more sites in one search result? For example an online-shopper who is looking for a specific product: He fires up google and clicks various results to compare prices, services or shipping fees. In these cases the user will probably find what he’s looking for on all (most) results on the first page, ergo each result will satisfy him. Still, he’ll return to the search results and will click the next result for the purpose of comparison.
In the case of SCASI, the winning result will be the last one clicked, since in Google’s point of view there will be no “bouncing”, eventhough the user eventually returns to another shop.
This is just a thought. I’m not sure how many of those cases there will be and if it’s even statistically relevant.
When researching, I quite oftenl open a lot of SERP links at the same time.
Anyhooo, *starts to build Google redirect links* haha.. ahem.
Great post Quads. That really fills the gaps in what you were saying before.
So you know what a SCASI is?
It’s a SERP Tranny!
See… the user clicks a SERP and he imagines her to be this hot chick who will answer all his questions and make him happy.
But then he sees the metaphorical full frontal nude and is like “OH NO NO! NO! NO! NO! GET ME THE FUCK OUT OF HERE!”
So he goes back to the results page and clicks another link, hoping for an answer to his question without any unexpected extra anatomical parts.
Rnt: Me too, but i’ve watched enough surfers to know that we are not typical users.
envul: I remember this one time I was talking to an Ausie in Thailand. He was talking with a bunch of us and was like “yea, last week I hooked up with this Thia chick that was SUPER hot. She was sooo sexy, but when we went to do it, she would only let me fuck her in the ass. Is that wierd?”
I’m like “Bro, you fucked a dude.”
He’s all “No way, she was HOT”
Me: “did you feel her twat?”
Him: “No . . . but she couldn’t have been a . . . shit . . . I was pretty drunk. “
there are some searches where browsing more than one result is probably a much more normal use pattern.
combining aspects of the Google quality rater definitions, (branded/navigational, informational and transactional) would possibly help mitigate this. navigational searches with a high SCASI may be a bad result, while transactional searches probably have a higher average SCASI (shopping around).
also the average time between SCASI could be used.
… Talking of anatomy, what is yours (or anyones) take on the break down of a SERP redirect link? For example;
http://216.239.59.132/search?q=cache:49ZRotMVTrYJ:en.wikipedia.org/wiki/Guitar_Hero_IV+Guitar+Hero+4&hl=en&ct=clnk&cd=4&gl=uk&client=firefox-a
Broken down to:
http://xxx.xxx.xx.xxx/
You asked “…but is it getting any clearer to you?”
YES it is! Thanks
Agreed that it’s not GAnalytics data. SCASI huh? Reminds me of SCSI drives. Pogosticking is my preference but it’s the same idea.
http://www.blindfiveyearold.com/search-pogosticking-and-seo
A user is presented with search results based on a specific query. The engine captures what result you click on and whether you return to that SERP and click on subsequent results and/or refine your query. (They could even conceivably determine the time between each click as a proxy for satisfaction with that result. This would reduce the chances of penalizing results that did deliver value.) The information can be aggregated for each query and compared to average pogosticking behavior by SERP rank.
[...] Descrierea oferita in respectivul articol este destul de concludenta si nu vreau sa incerc sa reproduc termenii “inventati” acolo, asa ca va pot oferi direct un link spre articolul respectiv. [...]
Quote: “This isn’t technically the “bounce rate” but would be a good corollary in most instances. However, in the event that the landing page actually answers the users questions, the corollary would start to break down.”
The SCASI (come on, there must be a better name for it…how about a Reclick or something?)
In the instance that a user clicks through to a site which answers their question in a single page (after which they move on to something else) this would only register as a bounce on the destination website’s analytics, not on Google’s. Google will only register a bounce is a ‘SCASI’ occurs, but in this scenarios there will be no SCASI.
If Google serves a SERP that a user clicks, and there are no SCASIs, then this is ‘job done’ as far as Google is concerned, and the site concerned will be getting major bonus points for dealing with the user’s query, unless the user bounces back and restructures their query to a related phrase, then Google would recognise that the user’s search term was not ideal, or that Google didn’t ‘understand’ what they meant.
Reclick implies that the same thing is clicked twice. I’m open to better names though
Has the redirect gone now? I don’t see it in the source code anymore. Hmmm. :-/
i just saw that today on some of the datacenters . . .
“If a surfer Clicks SERP A from a list of SERPs and then (10 secs to 4 mins later) clicks any other SERP on that same list of search results it is a negative quality indicator. If the user found what they are looking for, they wouldn’t need to keep clicking on SERPs. If they continue to click, they’re initial find was not satisfying.”
This is not well thought out. I can think of a perfectly reasonable explanation for multiple separate SERP clicks – comparative shopping.
It’s important to recognize that the intent of the shopper cannot be known – ever. One shopper might be looking for price, another looking for brands, but both use the same key phrase “surround sound bars.” There’s know way to know “why”, only (somewhat) “what”.
I don’t disagree that Google will likely “test” efficacy, but I can’t imagine the smart folks at Google (or any other search company) would try and compare apples to oranges…
Drew: Right – if someone is doing comparative shopping, it is in Google’s best interest to rank those sites lower in the SERPs and rank non commercial results above them (think adwords). It remains a strong indicator that the person found exactly what they are searching for if they stop clicking.