There's some discussion about this on some other forums: what do you think, should Google remove the pagerank display?
What do you get out of seeing the pagerank for your sites? Does it give you anything to see it on other sites that you might be browsing? If your grandma's browser had a pagerank-bar, how would you explain it to her and what should she make of it?
I wouldn't mind if it was removed from the toolbar but I would still like to see it available through an API.
But then again, since it is not updated all that often it is not all that much use as a diagnostic tool.
How about removing it from the toolbar and creating some sort of tool for the Webmasters tool set but have it be able to access more recent data? Being able to view it using various means, besides just a list of pages and their respective PageRanks would be useful too. Maybe something like a tree structure that can expand and contract at each of the various nodes so one can check through one's site structure more efficiently.
How would I explain it? I'd explain that it may have had something to do with the "authority" of a given page at some point in the past but means nothing this exact minute because she won't know how far in the past it was updated and things can and often do change.
> There's some discussion about this on some other forums: what do you > think, should Google remove the pagerank display?
> What do you get out of seeing the pagerank for your sites? Does it > give you anything to see it on other sites that you might be browsing? > If your grandma's browser had a pagerank-bar, how would you explain it > to her and what should she make of it?
> ... it may have had something to > do with the "authority" of a given page at some point in the past but > means nothing this exact minute because she won't know how far in the > past it was updated and things can and often do change.
I'm wondering: in what kind of situation would it matter to the user how Google assesses the importance of a page? Would you trust a low- PR site less than a high-PR site?
Ditto to what Craig said about removing it from the bar but keeping it within the webmaster's reach as a general overview gauge.
However, if it suddenly got dumped from the Google-Bar, what would the millions of bar-using sheeple out there make of it? Would they suddenly complain that the evil Darth Google has things to hide because they could see the Green-bar-of-madness a minute ago, but now they can't?
There is so much confusion out there about what the little green bar actually does; but on the whole, it has filtered through to users by now that 'more green means more important' in a vague, non-plussed sort of way.
And don't forget that a Green-bar-of-madness enabled Google Toolbar acts as an information gathering tool since it sends back cute & cuddly and totally non-invasive statistical data to its nest [although I guess that this could easily be retained without keeping the PR bar].
On the whole, I guess that it's one of those 'six one way, half a dozen the other' situations.
As for my grandmother, I guess I'd just let her know that the green bar is a measure of how environmentally friendly a site is and leave it at that.
> There's some discussion about this on some other forums: what do you > think, should Google remove the pagerank display?
> What do you get out of seeing the pagerank for your sites? Does it > give you anything to see it on other sites that you might be browsing? > If your grandma's browser had a pagerank-bar, how would you explain it > to her and what should she make of it?
> I'm wondering: in what kind of situation would it matter to the user > how Google assesses the importance of a page? Would you trust a low- > PR site less than a high-PR site?
I've had to wean a surprising number of clients off the whole 'PR is the most important thing since Oxygen' diet.
I think that part of the problem about this was caused by the many start-up SEOs who started cold-calling business owners a couple of years back and filling their ears with a load of drivvel about PR and how their sites would die without it. Back in 2005, before I moved to Cyprus, I lost track of the number of cold-calling SEO salesmen I had to dismantle on a weekly basis, all of whom were singing the same tune about PR and making the same idiotic promises about 'the power of sitemaps, portal pages, link farms, pyramids and the predictions of 'king Nostradamus'.
And sure as sh*t, those business owners would talk to their business owner buddies over a beer on Friday night and pass what they'd heard from these Yahoo [the Ignoranus*, not the search engine] salesmen, thus propagating the whole game of PR chinese whispers.
*Ignoranus = Someone who is not only stupid, but also an a$$hole
Add this general user confusion to the fact that every SEO crackpot out there has their own idea about PR and the little green bar, and you wind up with an SEO soup so thick, you could trot a donkey across it... So I'd say that a great many people out there today will actually trust in the green without first evaluating the level of information on any given site.
Ah, what the hell... 'Death to the Green-Bar-of-Madness', it's caused enough havoc for one century.
I'm greedy. If I've got something in my sticky fingers it's not that easy to tell me that I must return it. Also I'm used to these funny green pixels so I think the toolbar PR display should stay. I'd like to have a tooltip displaying the float though. And of course I want the real page rank per page, section, and so on in the Webmaster tools, and more stuff like that, for example a tool to decrease the PR of pages I dislike because they sell the stuff I deal with way cheaper.
As for my granny, unfortunately she's dead, but last time I showed her a page about her 100th birthday on my laptop I didn't bother to tell her who's responsible for the massive green and all the love sent by the nice folks far away in california. Not that she was able to spot the green without a magnifier, but if she'd have asked, I would have said that Google uses green pixels to draw the picture of my surfing behavior, and that personalized search, history and all that are essential, so just ignoring the activated widget leads to better search results even without letting Google scan my brain for each search query's context. Sebastian
> There's some discussion about this on some other forums: what do you > think, should Google remove the pagerank display?
> What do you get out of seeing the pagerank for your sites? Does it > give you anything to see it on other sites that you might be browsing? > If your grandma's browser had a pagerank-bar, how would you explain it > to her and what should she make of it?
I think it's a matter of when not if - the PR toolbar will go soon.
Google seems to be moving towards a much less 'open' index, and this, like the supps, is probably just another signal that is considered of relatively little use and/or simply perpetuating counter-productive practices like link exchange etc.
I think when G started out, it made sense to have the PR toolbar to differentiate themselves from the crowd - after all, everyone likes 'tools'. These days, though, I guess Google has become a victim of its own success - rather than modelling link structures on the web, it could be argued that PR (and to a certain extent Google itself) now drives the web - which has eroded the value of PR via something I call 'link inflation'.
Google has become so endemic that being at the top of Google rankings tends to dramatically increase your readership, which by the law of averages also increases your inlinks, which increases your PR. Thus a PR driven index would tend to have no 'middle class' - the rich get richer, the poor get poorer, and you'd now have sod-all chance of ranking well for competitive terms if you were a newcomer.
This 'google as a driver of the web' phenomenon is something I'm particularly certain Google is aware of - hence the move away from PR to a more complex system - with loads of signals. Nonetheless, you can't escape the fact that PR, albeit perhaps a more complex version than the one first postulated by Page et al, will always be a part of the index - things like the introduction of rel=nofollow support my belief that PR is something that has such utility at the moment that even direct behavioral data can't replace it completely.
Google has shown a propensity towards 'security by obscurity' with their index - they're constantly on the move, tweaking the index, acquiring businesses that might give them new data (urchin is one example) - basically, they're a moving target always trying to keep one step ahead of the spammers. Unfortunately as the whole show gets more and more complex I'm feeling that the collateral damage is increasing - it has to, really. Multivariate models have multivariate errors - Google must be aware that the quality of their index is quite central to their success.
Keeping people in the dark about how the index works probably helps reduce the pace of index exploit development - Removing this signal (however out of date and useless it is) would probably be well and truly in line with that trend.
> I think it's a matter of when not if - the PR toolbar will go soon.
It's gone already. The low frequency of pushes contrasted with the volatility of the environment makes it useless for any purpose I can think of.
Google sometimes drops hints. Rarely do they say: "THIS IS A HINT!"
>From the FAQ:
Q: I have a lot of new links pointing to my site but my site's toolbar PageRank has not changed in months.
A: Don't worry :). We actually recalculate PageRank quite frequently, but only push toolbar PageRank updates occasionally. This is our respectful hint for you to worry less about PageRank, which is but one of over 200 signals that can affect how your site is crawled, indexed and ranked.
I'd like to see it dumped. I think PR data should only be available from the webmaster's dashboard. If the company/person that maintains the site wants to give the client access to the area, that's fine but having it private would put an end to a great deal of nonesense. If people can't check it easily, then fake SEO's have one less method to to try and sell you what you don't need.
> There's some discussion about this on some other forums: what do you > think, should Google remove the pagerank display?
> What do you get out of seeing the pagerank for your sites? Does it > give you anything to see it on other sites that you might be browsing? > If your grandma's browser had a pagerank-bar, how would you explain it > to her and what should she make of it?
They'd probably just love to remove it... but doing so would cause so much confusion (just imagine) and disruption that the cost may well outweigh the benefit. I reckon their PR (pub rels) budget would take a serious dent just trying to explain this to the public.
I imagine when they added the pagerank indicator to the webmaster console they may have had an intention of removing TBPR.
As Phil notes, they've been degrading TBPR's usefulness in recent times so not really a concern.
> I think it's a matter of when not if - the PR toolbar will go soon.
> Google seems to be moving towards a much less 'open' index, and this, > like the supps, is probably just another signal that is considered of > relatively little use and/or simply perpetuating counter-productive > practices like link exchange etc.
> I think when G started out, it made sense to have the PR toolbar to > differentiate themselves from the crowd - after all, everyone likes > 'tools'. These days, though, I guess Google has become a victim of its > own success - rather than modelling link structures on the web, it > could be argued that PR (and to a certain extent Google itself) now > drives the web - which has eroded the value of PR via something I call > 'link inflation'.
> Google has become so endemic that being at the top of Google rankings > tends to dramatically increase your readership, which by the law of > averages also increases your inlinks, which increases your PR. Thus a > PR driven index would tend to have no 'middle class' - the rich get > richer, the poor get poorer, and you'd now have sod-all chance of > ranking well for competitive terms if you were a newcomer.
> This 'google as a driver of the web' phenomenon is something I'm > particularly certain Google is aware of - hence the move away from PR > to a more complex system - with loads of signals. Nonetheless, you > can't escape the fact that PR, albeit perhaps a more complex version > than the one first postulated by Page et al, will always be a part of > the index - things like the introduction of rel=nofollow support my > belief that PR is something that has such utility at the moment that > even direct behavioral data can't replace it completely.
> Google has shown a propensity towards 'security by obscurity' with > their index - they're constantly on the move, tweaking the index, > acquiring businesses that might give them new data (urchin is one > example) - basically, they're a moving target always trying to keep > one step ahead of the spammers. Unfortunately as the whole show gets > more and more complex I'm feeling that the collateral damage is > increasing - it has to, really. Multivariate models have multivariate > errors - Google must be aware that the quality of their index is quite > central to their success.
> Keeping people in the dark about how the index works probably helps > reduce the pace of index exploit development - Removing this signal > (however out of date and useless it is) would probably be well and > truly in line with that trend.
> M
> On Aug 11, 7:27 am, IceGiant wrote:
> > > search results even without letting Google scan my brain for each > > > search query's context.
> > Trouble with Google scanning your brain? > > The IceGiant Leadmet-3000 can help!
Who here fully understands the pagerank calculations and can make an educated guess at how Google has modified the calculations and the factors involved for different sites?
I can understand trying to keep it in the webmaster tools, but seriously: what would it tell you?
How about this (maybe I just need sleep :P), instead of "pagerank" (which tells us everything but nothing really), replace it with these items (for private viewing only):
- a measure of the "global importance" of our sites (from total inbound links minus whatever filters): this gives us a metric to work on, to track for changes. Site-based, not page-based. High resolution (not 10 levels).
- list of the pages with the highest number of inbound (from other sites) links, perhaps with the highest inbound "value" (links * importance).
- site-internal link-value distribution statistic: with the given inbound links (from other sites), which pages gain higher than "optimal" value? which ones are lower than optimal? This could help fix internal linking strategies, help find irrelevant content, help push specialized content that we're proud of, etc.
I think this is about what we'd want from a "pagerank"-replacement, right? What's missing?
> I can understand trying to keep it in the webmaster tools, but > seriously: what would it tell you?
It depends on how it is displayed. A simple PageRank per given page wouldn't tell you all that much.
A distribution of how PageRank flows into and through a site might be more informative.
But, since PageRank doesn't appear to do much of anything for SERPs, other than in the past making a given page available for showing up in the SERPs in the first place, i.e. non-Supplemental, will a webmaster gain any useful information from PageRank alone? I can't think of any.
I was thinking that a distribution of PageRank would help tune one's site navigation but on the other hand, how hard is that to do using simple "logic"?
> - a measure of the "global importance" of our sites (from total > inbound links minus whatever filters): this gives us a metric to work > on, to track for changes. Site-based, not page-based. High resolution > (not 10 levels).
I'm not sure I understand "global importance". You mean as far as contributing "value" to the site?
> - list of the pages with the highest number of inbound (from other > sites) links, perhaps with the highest inbound "value" (links * > importance).
I'd rather not go down the link valuation road. I can't exactly explain why, it is just a feeling that it would lead to no good.
One thing though, if one has only one link to a given page or one finds a new link added, one could then pretty accurately determine the actual "value" of a given link and I could think of a number of scenarios where that would not be a good thing even though it would seem to give one useful information.
> - site-internal link-value distribution statistic: with the given > inbound links (from other sites), which pages gain higher than > "optimal" value? which ones are lower than optimal? This could help > fix internal linking strategies, help find irrelevant content, help > push specialized content that we're proud of, etc.
I was thinking along these lines to but how hard is it to do that by hand without any outside data?
Unless one has a link to every page on every page, put links to the more important "deep" pages higher up in one's navigation "tree", less important pages further down or where they would occur naturally in one's navigation "tree".
In any event, this would seem to be the only "value" to PageRank now, as a way to view navigation issues but there has to be a better way. Even a graphical representation of a given site's navigation would seem to give more information.
> I think this is about what we'd want from a "pagerank"-replacement, > right? What's missing?
What's missing is getting rid of PageRank altogether, even though I previously thought it would be useful to keep around. The reason I thought of keeping it around, analysis of site navigation, would seem to be better supported without using PageRank at all.
> JohnMu wrote: > Who here fully understands the pagerank calculations and can make an > educated guess at how Google has modified the calculations and the > factors involved for different sites?
Unfortunately I would imagine that >95% of the population who know about PageRank connect it in some way to quality.
How often do you hear about people wanting to increase their PageRank?
> - a measure of the "global importance" of our sites (from total > inbound links minus whatever filters): this gives us a metric to work > on, to track for changes. Site-based, not page-based. High resolution > (not 10 levels).
Interesting - but might be problematic if relational measure is used? If they stuck to the scalar 1-10 measure then we'd be back to where we started...
> - list of the pages with the highest number of inbound (from other > sites) links, perhaps with the highest inbound "value" (links * > importance).
Too much data to give away methinks, but would be nice. Of course you can take the current external links data and check the PR of those links (hmm... not very useful) [BTW Joost De Valk's greasemonkey script is a nice toy for this]
> - site-internal link-value distribution statistic: with the given > inbound links (from other sites), which pages gain higher than > "optimal" value? which ones are lower than optimal? This could help > fix internal linking strategies, help find irrelevant content, help > push specialized content that we're proud of, etc.
Given that they have removed the supplemental tag I doubt they would give such granular info out...
Your proposal would be bloody fantastic John - I've suggested something similar myself in the past. I guess Richard has a point though - I'm seeing the index becoming more closed rather than opening up, and I suspect the big issue with anything like this, whether you call it pagerank or something else, is that Google is always going to be reticent to give up2date data about a site lest it be used in some unintended way by black hats to get a leg up.
Craig - I'm still not sure that I agree that PR has nothing to do with serp positioning. I think I'm probably out in the wilderness here but from what I can see it is still a major major factor. Its overall effect seems non linear across different niches - I suspect if you're in a highly competitive industry the other factors tend to drown it out - but my experience is that its far more than just a pretty number.
> > JohnMu wrote: > > Who here fully understands the pagerank calculations and can make an > > educated guess at how Google has modified the calculations and the > > factors involved for different sites?
> Unfortunately I would imagine that >95% of the population who know > about PageRank connect it in some way to quality.
> How often do you hear about people wanting to increase their PageRank?
> > - a measure of the "global importance" of our sites (from total > > inbound links minus whatever filters): this gives us a metric to work > > on, to track for changes. Site-based, not page-based. High resolution > > (not 10 levels).
> Interesting - but might be problematic if relational measure is used? > If they stuck to the scalar 1-10 measure then we'd be back to where we > started...
> > - list of the pages with the highest number of inbound (from other > > sites) links, perhaps with the highest inbound "value" (links * > > importance).
> Too much data to give away methinks, but would be nice. Of course you > can take the current external links data and check the PR of those > links (hmm... not very useful) [BTW Joost De Valk's greasemonkey > script is a nice toy for this]
> > - site-internal link-value distribution statistic: with the given > > inbound links (from other sites), which pages gain higher than > > "optimal" value? which ones are lower than optimal? This could help > > fix internal linking strategies, help find irrelevant content, help > > push specialized content that we're proud of, etc.
> Given that they have removed the supplemental tag I doubt they would > give such granular info out...
I've seen static pages rocket in the rankings due to press releases for terms that don't appear in those press releases - more than just an anchor text affect.
Google has no other accurate way of measuring popularity of a site.
You can go around until you're blue in the face adding h1..h3 tags, making text more natural, improving relevance, but the simple fact is that Google likes buzz.
Unless they have good (and universal) behavioral information, or their index is manually reviewed, they have very little alternative but to use the linking structure of the web as their primary determinant of 'buzz'.
Things like analytics and the google toolbar and perhaps even coarse metrics like 'time from clicking on google result to returning to google' (as a measure of time spent on site) are perhaps going to lead to metrics that can be used to compliment PR, but I don't think it will be replaced in a hurry - although they are trying to.
> Unless they have good (and universal) behavioral information, or their > index is > manually reviewed, they have very little alternative but to use the > linking structure of the web as their primary determinant of 'buzz'.
Are you sure they have very little alternative? Also, how long does buzz last? Almost by definition, "buzz" can't last very long otherwise it wouldn't be buzz, it would be noise.
If buzz doesn't last long, or doesn't last forever then a page that was boosted due to buzz should fall back down to where it was before but if it rocketed due to PageRank alone, that could never happen so something beside PageRank must be involved in what Google determines to be "buzz".
Buzz could also be measured other ways beside PageRank. If something comes up that is fairly new or unique but the number of sites its mentioned on increases quickly over time, whether large changes in PageRank are seen or not, that could be another way of ascertaining buzz, or so it would seem.
And, if sites seen the be involved in buzz topics are authoritative, which may coincide with higher PageRank, it could be that they rocketed due to a combination of a new popular topic and their "authority" alone with having high or higher than previous PageRank as a coincidence.
PageRank seems to be such a rough metric to be used from something as short term as buzz is. I would think that if it is involved, it wouldn't be a direct causal relationship otherwise once a page is boosted due to buzz, it would never fall back down to where it existed previously.
Who knows, it is all speculation anyway but again, PageRank almost always becomes the "target" because it is pretty much the only thing we can see.
>If something >comes up that is fairly new or unique but the number of sites its >mentioned on increases quickly over time, whether large changes in >PageRank are seen or not, that could be another way of ascertaining >buzz, or so it would seem.
And so how do you measure that Craig?
How do you 'measure' the popularity of a new topic?
If something is 'mentioned upon' other sites how do you then determine which of those sites is the most important?
Perhaps by the number of sites linking to, I mean mentioning, that site. Starting to sound familiar?
>PageRank seems to be such a rough metric to be used from something as >short term as buzz is.
Rough compared to what? Or are you assuming that the PR google uses is a discrete value from 0 to 10 like what we see in our browser?
>I would think that if it is involved, it >wouldn't be a direct causal relationship otherwise once a page is >boosted due to buzz, it would never fall back down to where it existed >previously.
Sure it is. Buzz would be something like (delta PR) / time, but irregardless, PR would still be the main factor.
Can you name me a better metric than PR to measure the authority / importance of a site without per user behavioral information?
> > Unless they have good (and universal) behavioral information, or their > > index is > > manually reviewed, they have very little alternative but to use the > > linking structure of the web as their primary determinant of 'buzz'.
> Are you sure they have very little alternative? Also, how long does > buzz last? Almost by definition, "buzz" can't last very long > otherwise it wouldn't be buzz, it would be noise.
> If buzz doesn't last long, or doesn't last forever then a page that > was boosted due to buzz should fall back down to where it was before > but if it rocketed due to PageRank alone, that could never happen so > something beside PageRank must be involved in what Google determines > to be "buzz".
> Buzz could also be measured other ways beside PageRank. If something > comes up that is fairly new or unique but the number of sites its > mentioned on increases quickly over time, whether large changes in > PageRank are seen or not, that could be another way of ascertaining > buzz, or so it would seem.
> And, if sites seen the be involved in buzz topics are authoritative, > which may coincide with higher PageRank, it could be that they > rocketed due to a combination of a new popular topic and their > "authority" alone with having high or higher than previous PageRank as > a coincidence.
> PageRank seems to be such a rough metric to be used from something as > short term as buzz is. I would think that if it is involved, it > wouldn't be a direct causal relationship otherwise once a page is > boosted due to buzz, it would never fall back down to where it existed > previously.
> Who knows, it is all speculation anyway but again, PageRank almost > always becomes the "target" because it is pretty much the only thing > we can see.
> Craig
> > M
> > On Aug 13, 10:06 am, cass-hacks wrote:
> > > > but my experience is that its far more than just a pretty > > > > number.
> > > How much of that is based on the fact that it is about the only number > > > we have to look at, pretty or not? :-()
> How do you 'measure' the popularity of a new topic?
The number of new pages on a given topic or, the number of new pages on a new topic.
The hard part would seem to be determining what a "new topic" is.
> If something is 'mentioned upon' other sites how do you then determine > which of those sites is the most important?
Same as pages are always ranked in the SERPs. Once the topic of the buzz is identified, the rest should be able to be handled by other algos normally used in ranking the SERPs.
> Perhaps by the number of sites linking to, I mean mentioning, that > site. Starting to sound familiar?
Of course, you are talking about PageRank. What I am saying is that PageRank needn't be directly involved in the process any more than it already is for any search query. All one would have to do is identify the topic of the buzz and then rank candidate pages as usual.
> >PageRank seems to be such a rough metric to be used from something as > >short term as buzz is.
> Rough compared to what? Or are you assuming that the PR google uses is > a discrete value from 0 to 10 like what we see in our browser?
"Rough" was probably a bad word to use. What I was trying to get at was that once the buzz has died down, unless lots of people start deleting pages/links on the topic, a page that ranked for the buzz topic purely on PageRank alone should never drop, all things being equal.
But if a page rocketed due to buzz, it should also drop when the buzz dies down but if it rocketed due to PageRank and the number of inbound links doesn't drop, what would then cause it to drop as it surely must as there is no more buzz?
On the other hand, if one used a delta of PageRank one might conclude that the buzz had died down or the value of a given page with regard to the buzz has dropped as there was no longer a large influx of new links but that need not necessarily be the case.
It could be just that the page isn't getting linked to as often as before because it is not as new and everyone has already seen it but does that mean it could no longer be the most relevant page on the topic?
> >I would think that if it is involved, it > >wouldn't be a direct causal relationship otherwise once a page is > >boosted due to buzz, it would never fall back down to where it existed > >previously.
> Sure it is. Buzz would be something like (delta PR) / time, but > irregardless, PR would still be the main factor.
The one doesn't necessarily follow from the other. I would agree that PageRank may be the main factor that we can see and know about but I wouldn't agree that it is THE main factor. It has half the equation right but not the other half. It could be a contributing factor to determining "buzz" but it doesn't seem the work well on the down side because a topic could still be considered "buzz" even after various pages have stopped rapidly gaining inbounds.
On the other hand, if "buzz" were determined by how many new pages are created on the topic, that would seem a more direct and reliable vector.
Determine buzz first, relevance second.
> Can you name me a better metric than PR to measure the authority / > importance of a site without per user behavioral information?
How does Google rank any page in a given SERPs, not on PageRank alone. You are assuming "buzz" is determined purely based on PageRank but it could be based on other metrics that have nothing to do with PageRank.
Determining the buzz could come first, sites ranking for relevance for the buzz could come second. If that is the case PageRank is no more a factor than it would be at any other time.
Google has ~200 signals to deal with, why lump everything onto one?
One thing that bothers me about all the things that are claimed that PageRank is the determinant for is an old engineering rule, "the right tool for the right job", the corollary being, "no tool is right for every job". Let PageRank contribute as it does to ranking SERPs but having it do double duty to try to determine buzz would seem too much like having a hammer and everything looking like nails.
Wouldn't the number of new pages on a given topic seem a much more reliable vector for determining buzz?
Pagerank - at least the one that we can't see - does have an influence on the search results (it's one of the things that you can test: link to near-identical sites from pages with different "toolbar pagerank" and you can see how it works).
I think what you guys are saying is that the idea of pagerank is good, but the number that is shown is only taken seriously by those who have no clue, right? (too bad that they're in the majority :-))
PageRank is the basis for Google's meteoric rise, its what made Google relevant when others were not. It's a big part of their history and their current affairs so I don't think they can/should remove it totally from the public. I do think they should make some changes to their documentation. The tool tip, "PageRank is Googles..." should be modified to indicate that the information is old, stale, and not dynamic. People have come to excpect a lot out of google, they return results for some well loved sites in the index within minutes of their publishing, but yet the foundation of their success is outdated by several months?!?!! When I tell that people who may not follow this stuff as close as we do they hardly believe it. I just think a disclaimer would help their own public relations, we know their motivations, why not let the public know. Then again I always thought they should include a disclaimer with their link:, site:, and related: commands as well saying that they are inaccurate by design. If you are trying to portray yourself as the freshest and most comprehensive index, on items that are not fresh nor comprensive should be tagged as such so people understand the difference.
After watching a commercial today that Ask.com wasted some more money on, they had a search for "the algorithm" on ask, so just for kicks I tried it myself.
Funny how they have an obvious hand-job as the top result, they can't even rank for a multi-million dollar advertising campaign! which is another story, but the top hand edited result has this:
""The Algorithm" refers to the search technology that powers Ask.com. This technology was developed by Rutgers University Professor Apostolos Gerasoulis and provides more relevant search results by identifying experts in specific subjects rather than relying simply on link popularity."
It's an off-handed knock on Google's PageRank system. Now if these Ask people had a clue what they were doing it may get some legs and matter, but due to the over whelming failure of ask's attempt at market penatration I don't think Google has to worry too much.
They need to remove some of the confusion behind PageRank and ranking, for the sake of their searchers. We know that PageRank is only influenced by links, but your average joe-surfer doesn't. He sees that he searched for something, clicked the result, then sees the PageRank, associating the PageRank of the page as an endorsement of the quality of the page for the search term. Something that is wrong.
The other thing to consider is that if visible toolbar pagerank was removed, a huge vacuum where it used to exist would be produced. What would rush in to replace it? Alexa? Some automated toolbar that creates its own pagerank based on yahoo's site explorer?
> Pagerank - at least the one that we can't see - does have an influence > on the search results (it's one of the things that you can test: link > to near-identical sites from pages with different "toolbar pagerank" > and you can see how it works).
> I think what you guys are saying is that the idea of pagerank is good, > but the number that is shown is only taken seriously by those who have > no clue, right? (too bad that they're in the majority :-))
> but the number that is shown is only taken seriously by those who have > no clue, right? (too bad that they're in the majority :-))
Right, and RIGHT! :-()
How and what for Google uses it internally, we can talk about till the cows come home to roost, or something like that but as for published PageRank, means little to nothing as soon as it is published.
It has its maximum value just before being published, which is when most people are talking about it. :-()
Craig - I think you're getting too caught up in the particular word I used - Buzz - that's one example I'm choosing to use.
Overall, what I'm trying to say is that pagerank DOES have an important role to play, and, like it or not, the one thing that it covers that the majority of on page ranking factors do not is popularity.
In my terms when I say buzz in that context I'm meaning popularity / relevance - yes, that changes over time - so does pagerank, as it is a relative measure that is normalized to a 0 to 10 scale for our consumption. Pagerank IS a measure of buzz / popularity / relevance - is it perfect? NOPE.. Would it be great to have an alternative - Definitely. Are they working on alternatives - you'd better believe it - but those measures have mega serious flaws.
I don't believe that other quanta meet the cut here YET - PR still has a big place, and you'd better believe it is important. To say it doesn't really affect serps is a long bow that I wouldn't personally be willing to draw.
Is PR losing relevance? Sure - I think it is - as I mentioned again the very success Google has had dictates that it is probably increasingly driving the very linking patterns that PR seeks to model. Unfortunately that happens with successful reward models (I've seen it happen in other industries too) - the solution, as I alluded above, is usually a more complex or pseudo-random model, which usually increases errors (or collateral damage if we're talking about serps).
For that reason alone taking PR out of the public consciousness is prob a good move.
> > How do you 'measure' the popularity of a new topic?
> The number of new pages on a given topic or, the number of new pages > on a new topic.
> The hard part would seem to be determining what a "new topic" is.
> > If something is 'mentioned upon' other sites how do you then determine > > which of those sites is the most important?
> Same as pages are always ranked in the SERPs. Once the topic of the > buzz is identified, the rest should be able to be handled by other > algos normally used in ranking the SERPs.
> > Perhaps by the number of sites linking to, I mean mentioning, that > > site. Starting to sound familiar?
> Of course, you are talking about PageRank. What I am saying is that > PageRank needn't be directly involved in the process any more than it > already is for any search query. All one would have to do is identify > the topic of the buzz and then rank candidate pages as usual.
> > >PageRank seems to be such a rough metric to be used from something as > > >short term as buzz is.
> > Rough compared to what? Or are you assuming that the PR google uses is > > a discrete value from 0 to 10 like what we see in our browser?
> "Rough" was probably a bad word to use. What I was trying to get at > was that once the buzz has died down, unless lots of people start > deleting pages/links on the topic, a page that ranked for the buzz > topic purely on PageRank alone should never drop, all things being > equal.
> But if a page rocketed due to buzz, it should also drop when the buzz > dies down but if it rocketed due to PageRank and the number of inbound > links doesn't drop, what would then cause it to drop as it surely must > as there is no more buzz?
> On the other hand, if one used a delta of PageRank one might conclude > that the buzz had died down or the value of a given page with regard > to the buzz has dropped as there was no longer a large influx of new > links but that need not necessarily be the case.
> It could be just that the page isn't getting linked to as often as > before because it is not as new and everyone has already seen it but > does that mean it could no longer be the most relevant page on the > topic?
> > >I would think that if it is involved, it > > >wouldn't be a direct causal relationship otherwise once a page is > > >boosted due to buzz, it would never fall back down to where it existed > > >previously.
> > Sure it is. Buzz would be something like (delta PR) / time, but > > irregardless, PR would still be the main factor.
> The one doesn't necessarily follow from the other. I would agree that > PageRank may be the main factor that we can see and know about but I > wouldn't agree that it is THE main factor. It has half the equation > right but not the other half. It could be a contributing factor to > determining "buzz" but it doesn't seem the work well on the down side > because a topic could still be considered "buzz" even after various > pages have stopped rapidly gaining inbounds.
> On the other hand, if "buzz" were determined by how many new pages are > created on the topic, that would seem a more direct and reliable > vector.
> Determine buzz first, relevance second.
> > Can you name me a better metric than PR to measure the authority / > > importance of a site without per user behavioral information?
> How does Google rank any page in a given SERPs, not on PageRank > alone. You are assuming "buzz" is determined purely based on PageRank > but it could be based on other metrics that have nothing to do with > PageRank.
> Determining the buzz could come first, sites ranking for relevance for > the buzz could come second. If that is the case PageRank is no more a > factor than it would be at any other time.
> Google has ~200 signals to deal with, why lump everything onto one?
> One thing that bothers me about all the things that are claimed that > PageRank is the determinant for is an old engineering rule, "the right > tool for the right job", the corollary being, "no tool is right for > every job". Let PageRank contribute as it does to ranking SERPs but > having it do double duty to try to determine buzz would seem too much > like having a hammer and everything looking like nails.
> Wouldn't the number of new pages on a given topic seem a much more > reliable vector for determining buzz?
I'm actually pro closing up the index a bit, purely because I think that constantly applying algorithmic patches is not a good way to sustain the index long term.
The more people understand the model, the more they'll prob just increase the rate of blackhat exploit development. The reaction from Google? Patch the algorithm.
Sooner or later you reach the point where the alg is so complex that keeping track of the cause for changes in user experience will become nearly impossible. As I said - a multivariate model has multivariate sources of error. Put differently, it's possible to bugger up a reward system like Google by making it too complex :)
> Pagerank - at least the one that we can't see - does have an influence > on the search results (it's one of the things that you can test: link > to near-identical sites from pages with different "toolbar pagerank" > and you can see how it works).
> I think what you guys are saying is that the idea of pagerank is good, > but the number that is shown is only taken seriously by those who have > no clue, right? (too bad that they're in the majority :-))
