homepage Welcome to WebmasterWorld Guest from 23.23.22.200
register, free tools, login, search, subscribe, help, library, announcements, recent posts, open posts,
Subscribe to WebmasterWorld
Visit PubCon.com
Home / Forums Index / Google / Google SEO News and Discussion
Forum Library, Charter, Moderators: Robert Charlton & aakk9999 & brotherhood of lan & goodroi

Google SEO News and Discussion Forum

This 42 message thread spans 2 pages: 42 ( [1] 2 > >     
Technologies Behind Google Ranking - from Google's Amit Singhal
chewy




msg:3700336
 11:28 pm on Jul 16, 2008 (gmt 0)

incredibly good posting today - among other things about synonyms.

I'm amazed that no one else saw this - hopefully my posting this is within TOS. It should be.

[googleblog.blogspot.com...]

 

tedster




msg:3700349
 11:43 pm on Jul 16, 2008 (gmt 0)

Yes, thanks for the thread start on this. I just linked to that article from deep inside another thread as well. It's quite the in depth "bragging" piece, isn't it? But it also drops some very interesting information about where Google is focused.

I thought this might hold a clue for some things we're currently noticing:

Our work on interpreting user intent is aimed at returning results people really want, not just what they said in their query. This work starts with a world class localization system...

So localization is step 1 in understanding user intent. I wonder if the recent geographic chaos in the google.co.uk SERPs has its roots in this work.

Also, as with any official communication from any company, beyond the information they shared I always ask, "why did they say it just now?"

tedster




msg:3700483
 4:19 am on Jul 17, 2008 (gmt 0)

I wonder what the exact implications of this have been in the ranking algo:

Another technology we use in our ranking system is concept identification. Identifying critical concepts in the query allows us to return much more relevant results.

In the July SERP Changes [webmasterworld.com] thread, minapple observed "I have noticed what might be a change in the relational [~] term weight..." That's quite a suggestive connection to me.

Marcia




msg:3700489
 4:26 am on Jul 17, 2008 (gmt 0)

Concepts:

define:Concepts [google.com]

IMHO that's what all the phrase-based analysis is about. It isn't just the words in the particular query, it's the concept it represents in relationship to the larger whole it's part of. The ways I can think of for identifying the concept behind a query is through a lexical topology - and particularly through keyword co-occurrence.

Added:

The paper on his site about query expansion is a good read:

[singhal.info...]

[edited by: Marcia at 4:50 am (utc) on July 17, 2008]

[edited by: Receptional_Andy at 10:52 am (utc) on July 17, 2008]
[edit reason] Fixed sidescroll [/edit]

nomis5




msg:3700569
 7:48 am on Jul 17, 2008 (gmt 0)

There's so much of interest in the article and other posters have homed in on a few. This paragraph grabs my attention:

"Someone searching for [bangalore] not only gets the important web pages, they also get a map, a video showing street life, traffic, etc. in Bangalore -- watching this video I almost feel I am there :-) -- and at the time of writing there is relevant news and relevant blogs about Bangalore."

When I did the suggested search on Bangalore, video aside, it was all there as the article mentions. Unfortunately the general information is supplied by Wikipedia amd the travel / tourist side by Wikitravel, but let's not go down that alley!

For me this has serious implications. For example, if my widget site is third "best" in the world for widget information, I would guess that it will appear well out of the top ten results? Because Google will probably present in the top ten, one or maybe two informational sites, a site that has videos of widgets, another that has lots of widget pictures, a couple of blogs on widgets, a site or two containing recent widget news etc.

This may only apply to single word searches though, because a search for "bangalore travel" brings up a very different set of results (wikipedia aside!).

Miamacs




msg:3700765
 1:41 pm on Jul 17, 2008 (gmt 0)

...

We can do this, we can do that, we're great aren't we?
What's this a shareholder meeting?

I wanted to read about technologies not 'results'... like how most of the stuff described in the later half of the post is taken care of by the *single* great idea of analyzing stuff already on file ...by regions.

Regionalization of the link profiles, behavior patterns, news, etc. all in all: *data*

You know, the backbone for local rankings. Local query data, local user behavior, local CTR, local *trust* ( originating from local seed sites/set or voted 'authorities' ) and *local anchor text profiles*.

It's the users and the sites which give weight to regional interests in regional indexes and not Santa Claus. So that queries get to be 'magically understood' (differently) for various parts of the world.

What made this possible NOW and not sooner?
New search technology?
Hell no, an additional 500,000 servers and a better infrastructure.
Apart of cool ideas on how to regionalize/mine/filter data there haven't been any groundbreaking changes for years.

...

The post has nothing about technology, even though we'd like to see some clarifications.
But coming to think of it... *why* would they talk about such stuff?

If someone is trying to understand the technology from this posts, they're just gonna get confused trying to reverse engineer stuff in an age when it's harder than ever for newbies to get up to speed. I've always had doubts about whether these posts are aimed at the industry at all.

...

ps: sorry Mr Singhal, please carry on posting... perhaps giving some non-exploitable tech info would go a long way for winning over persons like me. Not asking for secrets, just something to keep us occupied. in many ways *heh*

[edited by: Miamacs at 1:50 pm (utc) on July 17, 2008]

bwnbwn




msg:3700786
 2:00 pm on Jul 17, 2008 (gmt 0)

This is good but I see link buying tearing the serps up and rendering this technology useless in many verticals I watch. I see sites that have information out of the yazoo getting sent to the bottom by sites buying anchor links from forms, blogs and link selling networks.

There has to be a time when this comes to an end until it does the above technology can still be over written by links.

Marcia




msg:3700834
 2:50 pm on Jul 17, 2008 (gmt 0)

The post has nothing about technology, even though we'd like to see some clarifications.
But coming to think of it... *why* would they talk about such stuff?

Miamacs, if nothing else the posts are spoon-feeding the concept that there's a technology behind search, with this post in particular giving an introduction to technology that might whet the appetite of some to read further.

If someone is trying to understand the technology from this posts, they're just gonna get confused trying to reverse engineer stuff in an age when it's harder than ever for newbies to get up to speed. I've always had doubts about whether these posts are aimed at the industry at all.

I don't know who they're aimed at, but my impression and best guess is that, like most Googlespeak (if you get used to reading between the lines), the pervading message is, rather than using blunt force reverse engineering, to think more in terms of optimizing for users.

Repeat it and hear it enough times, and it can change the way some people view and approach optimization. Above all, the "philosophy" behind any technology they use is being revealed.

signor_john




msg:3700873
 3:22 pm on Jul 17, 2008 (gmt 0)

I don't know who they're aimed at

The general public, I'd imagine--investors, searchers, the press, and anyone else who might find the article interesting or enlightening. I'd guess that a lot of people might be intrigued to learn that a search engine can distinguish the intent behind a "dr" search (whether it's "Doctor Zhivago" or "Rodeo Drive") or a church in "Times" Square vs. an article in The New York "Times."

The author of the article isn't revealing any secrets or throwing out a challenge to SEOs. He isn't bragging or claiming perfection, either. He's just describing what Google tries to do when it gets a search query.

kaz




msg:3700968
 5:17 pm on Jul 17, 2008 (gmt 0)

For example, our algorithms understand that in the query [new york times square church] the user is looking for the well-known church in Times Square and not for articles from the New York Times.

Ok, I can follow that. But what I don't understand is when I do a search for [well known church in Times Square] why don't I get good results like I do when I search [new york times square church]. limitations.

signor_john




msg:3701017
 5:57 pm on Jul 17, 2008 (gmt 0)

Maybe because they haven't achieved perfection?

IanKelley




msg:3701052
 6:35 pm on Jul 17, 2008 (gmt 0)

If someone is trying to understand the technology from this posts, they're just gonna get confused trying to reverse engineer stuff in an age when it's harder than ever for newbies to get up to speed. I've always had doubts about whether these posts are aimed at the industry at all.

If you ask me they're giving away too much information even in this limited post. They should have remained quiet about how they do what they do from the very beginning, PageRank included. If they had there would be a lot less spam than there is now.

Don't get me wrong I'm glad they're sharing information... just illustrating that, in a world where limitless numbers of people are trying to take advantage of search systems for profit, it's completely unreasonable to expect a search engine to give specific details about their technology.

kapow




msg:3701122
 8:14 pm on Jul 17, 2008 (gmt 0)

I suspect the philosophy behind their technology is:

1. Keep the world impressed.

2. Sell Ads to every business on Earth.
i.e. an obsession with local (Dominate generic search and you can sell Ads to industry leaders. Dominate local search and you can sell Ads to every business on Earth).

3. Try to make SEOs think they can't keep up.
e.g. Posts from Google presented as a friendly technology update, but really an offensive to presuade SEOs they can't keep up (please don't chase links and mess up our only algo, just buy Ads).

Marcia




msg:3701125
 8:26 pm on Jul 17, 2008 (gmt 0)

The local search ploy is kind of a hide-behind-Momma's-skirts type of smokescreen used by Google, much like they use PR as hiding-behind-Mum's-skirts. It's a diversion that gets webmasters focusing on *that thing* when the picture is much broader.

Example, in Mattspeak:

Webmaster Question: Will doing "this" or "that" affect my rankings?
MC Answer: It won't affect your PageRank.

Of course, PR is only one of 200+ factors; but it's handy skirt to hide behind with a non-answer, and sidesteps the issue very nicely. It neither confirms or denies (or says anything, really).

Such is local search in GoogleSpeak: same premise, same tactic. ;)

Marcia




msg:3701295
 12:50 am on Jul 18, 2008 (gmt 0)

Giving some more thought, aside from IR theory, which is super-interesting. Let's face it, there are certain specifics that can't be revealed. But from all I've read and listened to, IMHO the underlying philosophy behind their technology and algos (and services) is: discovering, creating, maintaining and ensuring a positive user experience.

That's what gave them their start, and all signals point to that being a pervasive frame of mind - and a sensible one, too. Users who have a positive experience come back and use services, it builds branding and user satisfaction and loyalty, and profits grow and multiply.

When you listen to "Mattspeak" in the messages to webmasters, both openly and between the lines, the message conveyed (to me, anyway) is to create sites for users, and for a reason. Sounds overly simplistic (like Koolaid, eh?), but it's chock full of implications. It just happens to coincide with a ton of marketing and conversion literature out there about designing site architecture centered around enabling simple user traffic patterns and usability and increasing conversions. That's not much different from on-page and sitewide SEO, when you really thnk about it, point by point.

That's just the impression I take away from reading and listening, FWIW.

[edited by: Marcia at 12:52 am (utc) on July 18, 2008]

atlrus




msg:3701334
 2:08 am on Jul 18, 2008 (gmt 0)

That was an useless article and as a poster said - spoon-feeding Google ideology that they are superior.

Maybe MR. Amit Singhal should have searched for "kevorkian dr" before he started bragging about "how great and ground-breaking the algorithms are"... or for "dr rodeo"...what a load!

Marcia




msg:3701339
 2:27 am on Jul 18, 2008 (gmt 0)

Sorry, but I don't accept that spoon-feeding was in any way implying anything negative - surely not superior or useless. The posts are beautifully done.

There are people who know so little about how search works that it's sad - or maybe they think all it means is keyword stuffing and/or amassing a bunch of links, without a clue what search is really about.

What Mr. Singhal is doing is presenting an introduction to the meat of IR in a pleasant, simple to digest way, and that's definitely a POSITIVE thing.

atlrus




msg:3701346
 2:46 am on Jul 18, 2008 (gmt 0)

I also appologize, but the posts, while beautiful, are still useless.

They [google] know very well that only people from the indusrty read their blog(when was the last time you visted a blog of a toilet paper manufacturer to inform yourself on the latest two-ply techniques?) and I was personally offended how they titled the article, compared to the actual content...

if "dr"+[query]
then "dr"=doctor
if [query]+"dr"
then "dr"=drive

wow, ground breaking...You can also type all the above examples in Yahoo (the worst SE) and get the same. If Yahoo can do it, it's miles away from groundbreaking.

Now as another poster said before - getting rid of the paid links spammers, that would be a goundbreaking achievement. No, wait:

if "your link here" present
then "link value passed"=0

wow.

Marcia




msg:3701354
 3:16 am on Jul 18, 2008 (gmt 0)

Clarification:

>>as a poster said - spoon-feeding Google ideology that they are superior.

*This* poster most certainly did not say anything of the sort. What *this* poster said (though more politely and in not so many words) was that they're feeding IR theory, in a VERY nice way, to the unwashed, clueless masses who need to GET a clue about how search works.

FYI, there are *professional SEOs* who have been in business for YEARS and haven't got a clue about canonical issues. Their own cross-linked network of sites all resolve with 3 different URLs for the homepages.

Google is doing nothing less than public service by making basic IR information available. If the self-satisfied and unclued (who optimize very nicely for 4 word phrases that no one searches for) don't want to take advantage, that's their problem, not Google's.

potentialgeek




msg:3701401
 5:35 am on Jul 18, 2008 (gmt 0)

Nobody (unless I missed it) commented on a major change in the Google algo:

"Personalization is another strong feature in our search system which tailors search results to individual users. Users who are logged-in while searching and have signed up for Web History get results that are more relevant for them than the general Google results. For example, someone who does a lot football-related searches might get more football related results for [giants], while other users might get results related to the baseball team. Similarly, if you tend to prefer results from a particular shopping site, you will be more likely to get results from that site when you search for products. Our evaluation shows that users who get personalized results find them to be more relevant than non-personalized results.

Behavioral advertising is the future of online ads but Google has applied the concept to personal search. But I don't know how many people use Google search history.

Tedster, we should have a subforum of Google Search News for official comments from Google. This blog post would be a good one to kick it off. It seems like the most revealing comments by Google on its algo in years. Every webmaster should read it.

p/g

P.S. I hope Google takes the personalization concept a step further to allow filtering of results by the user; e.g., to omit results from: i) countries x, y, z; ii) sites a, b, c; iii) media/news results.

P.P.S. I don't mind the bragging if it causes them to share their secrets. ;)

[edited by: potentialgeek at 5:40 am (utc) on July 18, 2008]

Reno




msg:3701402
 5:36 am on Jul 18, 2008 (gmt 0)

I found Amit Singhal's blog post friendly and informative, though the occasional self congratulatory phrasing was perhaps a bit much. The unprecedented growth of their service speaks for itself, so to my eyes the back pats were unnecessary:

"including a best-in-class spelling suggestion system"

"high accuracy is hard, and we do it well"

"and yes, we get that right too"

The article would read just as well without all that, but given the otherwise good overview of IR, that's nit-picking on my part...

................................

idolw




msg:3701450
 6:59 am on Jul 18, 2008 (gmt 0)

I am not sure why everyone here is saying they congratulate themselves too much.
- they built a billion dollar company in less than 10 years,
- 70% of the internauts use their services even though they are not forced to (opposite to M$).

SEOs are just small beatles and usually not their major shareholders, so I'd say an "our technology is so cool that we can do this and that pretty well" type of post is quite a good idea to keep the stock green in rough times.

atlrus




msg:3701605
 12:48 pm on Jul 18, 2008 (gmt 0)

*This* poster most certainly did not say anything of the sort. What *this* poster said (though more politely and in not so many words) was that they're feeding IR theory, in a VERY nice way, to the unwashed, clueless masses who need to GET a clue about how search works.

The article was titled Technologies behind Google ranking and not Technologies behind Google search.

The first would have been really valuable to the industry, while the latter should have been on Wiki :)

Miamacs




msg:3701618
 1:23 pm on Jul 18, 2008 (gmt 0)

They should have remained quiet about how

They've never shared a byte of info above what was absolutely necessary for them to get ahead of the competition, keep webmasters ahead of black hats or inform shareholders.

discovering, creating, maintaining and ensuring a positive user experience

... *that Google can recognize*.

I agree with you on that Google is close to be *able* to identify a good user experience, and we too can now concentrate solely on that. But our standards for web design are not natural anymore.

Google became the supplier of ~20% of ALL web traffic and with them came a systematic distorsion of web design. It's very hard to re-think non-web design and architecture for those not living in this.

designing site architecture centered around enabling simple user traffic patterns and usability

To us, who grew with Google it's obvious to have *text links* for our navigation, to use *AJAX*, *Flash* ( i know I know ), *JavaScript* only for stuff we don't mind never getting fresh traffic to, to use CSS tricks instead of image based graphics and a hundred thousand such little nuances that shaped our inventories. We're so used to this that we don't even see their influence anymore.

In the new age, White-hat SEO ( if you take away link building and site architecture ) is about trying to get everything *right* for Google. Not stuffing keywords, but for example working around solutions that'd effectively cancel out search engine traffic. Though some things might naturally provide a better experience, we can't use them if we want Google users to find the pages.

If people discovered sites through means that don't require the site to be digestible by a (patched) 1998 bot, I'm sure sites would be designed with elements closer to what is 'natural' ( especially with the bandwidth now available ). I'm not blaming anyone, I'm not even sure I wanted to see that. I'm just saying that what is *obvious to Google* (est. 1998) and *Google-SEOs* (est. 1998) it's not so obvious for *Others* (in 2008).

They've shaped ( froze ) web design trends for way too long to be able to just 'let it all happen naturally' from now, and take their hands off of wannabe webmasters. A site that's providing a good user experience, could still be lightyears away from being Google compliant.

It'll happen, someday. But it's too early to 'let go' just yet.

they think all it means is keyword stuffing

I agree. Google needs to communicate its achievements. The hint for 'keywords not on page' was cool, and will keep people from keyword stuffing. But that's only half of the picture. If they told people that they can *see/guess* popularity from [CTR and user behavior after landing] ... and they extract data region by region to match local interests, what bad would had it done? And THAT would have given a clue for newbies to want to create sites that'll provide a better user experience. That single sentence. No need for details, just mentioning the most important facts otherwise you'd think they're ESPers.

And that was my problem...

WE ( veterans ) hear things they don't say, and this post was NOT useful for newbies. It's totally confusing for anyone who's new to this. WE know, but then... this is supposedly not aimed at us.

latter should have been on Wiki :)

*gah* ...totally.

...

[edited by: Miamacs at 1:25 pm (utc) on July 18, 2008]

atlrus




msg:3701634
 1:57 pm on Jul 18, 2008 (gmt 0)

I would have let the bragging by if it was actually worth it. I am not sure how many of you actually tried to look at the examples in depth. In my previous post I showed you how the "dr" example is not worth jack, unless you pick out an example which works (see kevorkian drive vs. doctor rodeo).

The vancouver pc website is even funnier as they pretty much typed in the URL of the website. You can see how useless this is when instead of searching for "cool tech pc, vancouver, wa" you simply search for "cool tech pc" - the same website comes first (duh). But do a search for "custom pc parts vancouver, wa" and that website is gone. Instead, there are websites indeed stuffed with keywords...

I guess the post was made for those who know nothing about search, but then it was posted on the wrong place.

signor_john




msg:3701663
 2:28 pm on Jul 18, 2008 (gmt 0)

WE ( veterans ) hear things they don't say, and this post was NOT useful for newbies. It's totally confusing for anyone who's new to this. WE know, but then... this is supposedly not aimed at us.

Of course it wasn't aimed at SEOs (or "the industry," as atlrus would say).

Just because Google has a spam-fighting team doesn't mean that the main focus of Google Search is building protective walls around a static search technology from 1998. Even if SEOs and spammers didn't exist, search technology would move forward, and search-based companies like Google would be working to make their products more useful and versatile. It's no different from any other technology business: Intel makes its processors faster, hard-drive makers increase the capacity of their disks, camera manufacturers introduce features like face recognition or sensor dust reduction, and so on. General Electric's advertising agency may have invented the phrase "Progress is our most important product," but that sentiment certainly isn't limited to GE.

SEO may have a role to play in the "search industry" (for better or for worse), but it's unlikely to be the first thing that goes through the minds of the programmers, statisticians, semanticists, mathematicians, etc. who work to improve and expand the applications for search.

atlrus




msg:3701747
 3:48 pm on Jul 18, 2008 (gmt 0)

WE ( veterans ) hear things they don't say

Sometimes a cigar is just a cigar...

skibum




msg:3704145
 5:33 am on Jul 22, 2008 (gmt 0)

Our evaluation shows that users who get personalized results find them to be more relevant than non-personalized results.

As with all these things, it would be nice to understand what "our evaluation" means. What was evaluated, what were the metrics, what were the signals, what does "more relevant" mean? Some "research" is curious, self-serving (as it is to some degree with any company that puts out "research" findings), and the findings leave you scratching your head. Sometimes that stuff jumps right out at you but it seems almost impossible to figure out what is really being said. Is the goal to ensure people find what they want in the fewest possible steps, to get them to click on the most ads possible, to do the fewest searches per session on Google, to interact with the most possible results on the page of results presented? When the variety of things that can be called "relevance" is so vast and relevance is not defined, it makes it possible to say anything and absolutely nothing at the same time.
A statement like when we tried xyz, we found that consumers found what we presumed they were looking for as measured [by the time spent on Google search results, the number of times they hit the back button to come back to Google, to hit what looked to use to be a conversion event tracked by G Analytics, the number of videos they viewed to completion] would shed much more light on what is really being said in "studies" and "research" posts, presentations and articles.

Interpreting intent for Adwords does not generally sem to work very well from an advertiser perspective but whatever they do in organic doesn't seem to have much effect one way or the other.

Seems like this statement could use two more words in it:

Our work on interpreting user intent is aimed at returning results we think people really want, not just what they said in their query.

Marcia




msg:3704160
 5:58 am on Jul 22, 2008 (gmt 0)

WE ( veterans ) hear things they don't say, and this post was NOT useful for newbies. It's totally confusing for anyone who's new to this. WE know, but then... this is supposedly not aimed at us.

Yes we do, because we've been watching them and listening to them for a long time - and also have learned how to read between the lines.

SO: on a constructive note then, if what Google does isn't at the level - or scalable enough - to reach the masses who really need to hear it and learn, then here's the question:

What could Google possibly do to make scalable information available that they could reach Average Jane and Joe with, on a broad scale, about how to do a good, "Google friendly site" that's also good for users?

Can they, or is it a lost cause and completely impossible?

tedster




msg:3704196
 6:54 am on Jul 22, 2008 (gmt 0)

I think Google has begun to develop the kind of communication mechanisms they will need to do the job, but it's going to be a long haul. And there will always be webmasters who have no interest.

One way I understood Google's goal is to make it easier to create quality than to imitate quality. They've already made some remarkable progress on that front - more than I expected. So at some point, the average webmaster may not need to get Google's point of view, they'll just build a website and Google will send traffic. I already know of sites like that.

Some of the new HTML 5.0 tags such as <section> <header> <footer> <nav> and <article> would help the average webmaster a lot - as well as the search engines. So maybe in 6 or 7 years when all that permeates through to the average webmaster, and Google has scaled up their efforts to communicate, we'll have a different landscape.

Right now, with this article, they are NOT talking to most members here, that's for sure. It's for a different and more general audience. I wonder if the official Google blog even reaches that audience.

This 42 message thread spans 2 pages: 42 ( [1] 2 > >
Global Options:
 top home search open messages active posts  
 

Home / Forums Index / Google / Google SEO News and Discussion
rss feed

All trademarks and copyrights held by respective owners. Member comments are owned by the poster.
Home ¦ Free Tools ¦ Terms of Service ¦ Privacy Policy ¦ Report Problem ¦ About ¦ Library ¦ Newsletter
WebmasterWorld is a Developer Shed Community owned by Jim Boykin.
© Webmaster World 1996-2014 all rights reserved