The author seems to have missed the core argument about spam websites: They reduce the quality of search results and therefor devalue google and other search engines as a way to find information.
Not sure I understand your complaint. The idea is that these sites get de-indexed after the grace period. (i.e. they will no longer show up in the search results).
During the grace period they would be decreasing result quality. I don't care if all the junk cluttering up my searches will be gone in a few months, because I'm not going to wait a few months to try to find the information I'm looking for.
Allowing money to influence rankings (selling "mappings" is basically selling pagerank) would also potentially decrease quality; I am much, much more inclined to having websites ranked only by their quality and the quality of information in them, which is the way search should work anyways.
Oh, and a way that this could be broken: One person wants to be at the top of the results for a single, fairly common, search term. They start a bunch of "funnel" sites (hundreds, maybe), and manage to get most of them into the top ~40 pages[1] for that search term. Then, in the marketplace you've proposed, they buy up all of them from themselves, at as low a price as possible. The resulting increase in pagerank (which they're buying from themselves, so only paying google for the privilege) pushes them up to the top of the rankings for that high-value search term.
[1] Random number. Could be higher, could be lower.
With respect to your first point - the idea is that these spam sites are cluttering up things already anyway. A result which has them de-listed after three months is better than the current situation. You could possibly find a way to eliminate the grace period altogether if you could find a way to demonstrate the value of particular keywords to those who would buy them. All google would really need to do is release more information about keywords. And think - they only keep that information guarded now because they want to make life harder for spammers.
With respect to the second point - again, the answer is to keep incentives properly aligned with desired outcomes. In this case I suggested that the price for a mapping be increased if the marketer already has sold to a stable of quality sites (as determined by other aspects of the PR algo).
With respect to your third point - I agree these sorts of issues will have to be sorted out. This particular one would be easily avoided though - just place constraints on how many domains they can register as funnel sites. Yes they can try to game - but now being caught has a real risk - the loss of their legitimate income.
But there is no doubting that there will be real difficulties of the sort you mention. Nevertheless, I think the general gist is good - however the final form, black hats need to be incentivised to work with the system, not against it. And this can only be achieved by bringing them into the fold somehow.
pitches to rank perfectly fine for the search queries before the spam sites push them down the index. The black hat SEO guy therefore isn't adding any value to the unsophisticated end user he targets at all. Better solution: anyone pitching SEO services to webmasters on the basis of their competing sites' higher rankings gets their competing sites booted from the index.
Even if your solution were to miraculously yield better results it would still create a massive PR problem for Google, who have always purposefully kept paid and organic search rankings separately. Frankly I'd probably trust an index that directly sold SERPs rankings more than one that deliberately facilitated spammers acting as intermediaries.
In what sense is putting up a simple website with articles (even poorly written ones) about topic X, with a domain name and meta tags about topic X, considered "black hat" SEO?
Because the content clearly isn't designed to add value. It just gets in the way of what people are actually looking for. Then there is the use to which these sites are put - which is explained in the article - to extort legitimate websites that do try to add value.
The content on those sites is also frequently scraped from other, more legitimate sites. Next time you see a Wordpress spamblog, try doing a (quoted) Google search for some of the post content.
Another tactic I've seen, simpler than the one described in the linked article, has simply been to link to a target site from all of the SEOer's content-scraping blogs. The core principle of what makes this all black-hat SEO is the same, though: they're building sites with the intent to manipulate search engine rankings, rather than sites which anyone would want to visit.
True - this one wasn't though. It was content created by cheap labour. Scraped websites are less of a concern because they are easier to detect algorithmically. Mind you Google doesn't seem to have entirely solved this problem either.
No, it's not a good idea for Google to let business owners buy these pages to increase their own ranking - willingness to pay isn't a useful signal for page relevance.
3.5 is not a part of it at all. They don't need to link to legitimate websites in order to be a nuisance - they just have to rank well and therefore push legitimate sites down.
3.5 doesn't even make sense. Why would google penalise a site for being linked from one that ranks for a particular term? If google thought the page was spam then they wouldn't rank it in the first place. The mostly likely result of linking competitors would be to increase their rankings.
Surely inbound links have no bearing on your search ranking...that would be a remarkably stupid way for Google to empower the worst citizens of the web to hold the rest hostage. It would become trivially easy to outrank any legit site by bombing them with links from your spam farm.
Where YOU link has an effect on your ranking, but surely not who links TO you.
Gain control of a bunch of sites that are hosting viruses, autodownloading exe's, or otherwise performing unlawful hacks on the user. Then start linking to all of the people who rank higher than you. You could pick off all of your pagerank competitors one by one.
(this is theoretical, but I don't see why it wouldn't work in a probabilistic sense)
It's more popularly referred to as "Google bowling." A couple years back, Matt Cutts acknowledged that it was possible, though he also stated they have countermeasures making it very difficult to do in practice.
"that would be a remarkably stupid way for Google to empower the worst citizens of the web to hold the rest hostage"
It is easy to forget that Google is not the Internet. A time goes on, SEO will cause Google to become nearly useless, and people will move onto greener pastures.
There's a lot of speculation around it. On WebmasterWorld you'll hear stories of people who bought no paid links get delisted because of spammy links.
The'll say a competitor bought the links to their website, then reported them to Google for black-hat SEO -- basically setting them up and then turning them in. Of course, this is the first defense anyone would use if they were caught actually link spamming -- so who can tell how common it is. (By the way, I don't see how Google could tell either.)
There are rumors of other bad things going on too. People duplicating your content to try and get your copy removed, bogus DMCA claims, etc.
One of the cool things that Google provided during their SearchWiki experiment was the [x] button. An option to remove a site from further search results.
This could easily give some power back to the users to not get hassled by such spammy site, as well as give Google some inputs to take action against such sites.
Google is already in the process of doing this. It's their company, they have access to all the data, why would they need individuals to ferret out underdeveloped keywords for them?
23 comments
[ 27.6 ms ] story [ 794 ms ] threadAllowing money to influence rankings (selling "mappings" is basically selling pagerank) would also potentially decrease quality; I am much, much more inclined to having websites ranked only by their quality and the quality of information in them, which is the way search should work anyways.
Oh, and a way that this could be broken: One person wants to be at the top of the results for a single, fairly common, search term. They start a bunch of "funnel" sites (hundreds, maybe), and manage to get most of them into the top ~40 pages[1] for that search term. Then, in the marketplace you've proposed, they buy up all of them from themselves, at as low a price as possible. The resulting increase in pagerank (which they're buying from themselves, so only paying google for the privilege) pushes them up to the top of the rankings for that high-value search term.
[1] Random number. Could be higher, could be lower.
With respect to the second point - again, the answer is to keep incentives properly aligned with desired outcomes. In this case I suggested that the price for a mapping be increased if the marketer already has sold to a stable of quality sites (as determined by other aspects of the PR algo).
With respect to your third point - I agree these sorts of issues will have to be sorted out. This particular one would be easily avoided though - just place constraints on how many domains they can register as funnel sites. Yes they can try to game - but now being caught has a real risk - the loss of their legitimate income.
But there is no doubting that there will be real difficulties of the sort you mention. Nevertheless, I think the general gist is good - however the final form, black hats need to be incentivised to work with the system, not against it. And this can only be achieved by bringing them into the fold somehow.
Even if your solution were to miraculously yield better results it would still create a massive PR problem for Google, who have always purposefully kept paid and organic search rankings separately. Frankly I'd probably trust an index that directly sold SERPs rankings more than one that deliberately facilitated spammers acting as intermediaries.
Another tactic I've seen, simpler than the one described in the linked article, has simply been to link to a target site from all of the SEOer's content-scraping blogs. The core principle of what makes this all black-hat SEO is the same, though: they're building sites with the intent to manipulate search engine rankings, rather than sites which anyone would want to visit.
3.5 is not a part of it at all. They don't need to link to legitimate websites in order to be a nuisance - they just have to rank well and therefore push legitimate sites down.
Where YOU link has an effect on your ranking, but surely not who links TO you.
Right?
Gain control of a bunch of sites that are hosting viruses, autodownloading exe's, or otherwise performing unlawful hacks on the user. Then start linking to all of the people who rank higher than you. You could pick off all of your pagerank competitors one by one.
(this is theoretical, but I don't see why it wouldn't work in a probabilistic sense)
It is easy to forget that Google is not the Internet. A time goes on, SEO will cause Google to become nearly useless, and people will move onto greener pastures.
The'll say a competitor bought the links to their website, then reported them to Google for black-hat SEO -- basically setting them up and then turning them in. Of course, this is the first defense anyone would use if they were caught actually link spamming -- so who can tell how common it is. (By the way, I don't see how Google could tell either.)
There are rumors of other bad things going on too. People duplicating your content to try and get your copy removed, bogus DMCA claims, etc.
Judging by the recently quality of their results, they already have
This could easily give some power back to the users to not get hassled by such spammy site, as well as give Google some inputs to take action against such sites.
Bring the [x] button back, Google :)
See recent patent: http://www.itproportal.com/2010/6/17/google-patent-looks-eme...