Penalty For Dead Links? Options

[bradlee27514]

I've read that your site may be penalized for having dead outbound links on it (I guess b/c the search engine would assume that your site is not maintained regularly?). Is there any hard evidence on how much of a penalty you receive?

We have something in the works to check the links on our sites automatically, but I don't know when it will be finished. We have some searchable directory sites that have tons of outbound links. We probably notice around 10-20 a month that need to be changed/removed.

cheers,
brad

[SearchRank]

I have never heard of a search engine penalizing a site for having broken or dead links. The only problem you would run into is human edited directories which if they find a bunch of broken links on your site when reviewing it for inclusion may decide not to list it.

[bradlee27514]

Thanks for the reply searchrank. I'll try and track down where I read that.

cheers,
brad

[OldWelshGuy]

Hi bradlee27514,

I am with David here as well., especially on a directory. If you are cleaning up broken links monthly, then for sure your better than most. One word of advice though, if a site is showing a 404, don't drop it on the first instance, check it again in a day or two, as it might just be a server hosting glitch at the time of checking.

OWG

[Jill]

The search engines only penalize sites that are doing things to deceive them. You can rest assured that they would never penalize a site for broken links.

It's possible that a site like that might not be given as much weigh in the algorithm, which some might twist into thinking that's a "penalty" but it's not a penalty at all, as they would never penalize for something that wasn't some sort of trick.

[tempy]

Indeed. It would be logical to think that a site with dozens of broken links (or over a certain percentage, perhaps) would be 'marked down' as a less useful resource, or something.

However, I recently went through a site that I had almost forgotten I had (it was for a defunct project) and found about a hundred broken (or rather, out of date) links on one page. It was still the page with the highest PR and highest rank in Google, for a number of relevant keywords. The rest of the site was nothing to boast about either. (IMG:http://www.highrankings.com/forum/style_emoticons/default/tongue.gif)

[Michael Martinez]

I don't think they would downweight the page, but certainly the dead links would be diluting the value conferred by other links on the page (in terms of the classic PageRank calculation).

I think someone has finally come up with a valid model that fits the expression "bleed PageRank", but it doesn't work the way people think it works. If you think about it, those dead links are hurting the documents that the active links point to. So, you ARE bleeding potential PageRank away from the active target documents.

That might explain in part why Yahoo! and DMOZ don't seem to help as much as they used to. Both directories have a fair amount of dead links in them. Once you accumulate millions of links in your directory, keeping them all current is a real challenge.

[tempy]

There are indeed a fair few dead links in DMOZ. There is also an extraordinary amount of dross and inaccurate/out of date information there.

The last few times I've had a good wade through, it leaves me the impression of a very shabby operation these days. Not like the DMOZ of old.

[bradlee27514]

It's possible that a site like that might not be given as much weigh in the algorithm...

Here are you referring to a site with broken links, or was this in reference to when I said the site was a directory? Is there something about directory sites theyautomatically hurt their rankings? I think you mean the former, but if it's the latter could you explain?

Thanks for all the responses! Great site Jill!

cheers,
brad

[Alan Perkins]

I think someone has finally come up with a valid model that fits the expression "bleed PageRank", but it doesn't work the way people think it works.  If you think about it, those dead links are hurting the documents that the active links point to.  So, you ARE bleeding potential PageRank away from the active target documents.

It's likely that dead links are ignored for Pagerank calculations, since they are probably treated as dangling links (links to documents that have not been retrieved) and dangling links are ignored.

[Michael Martinez]

Dangling links are ignored in the models discussed in the very old papers. We cannot be sure that they are ignored in whatever Google is doing these days.

[Alan Perkins]

I don't think they would downweight the page, but certainly the dead links would be diluting the value conferred by other links on the page (in terms of the classic PageRank calculation).

I was talking about the "classic PageRank calculation".

I think someone has finally come up with a valid model that fits the expression "bleed PageRank", but it doesn't work the way people think it works.  If you think about it, those dead links are hurting the documents that the active links point to.  So, you ARE bleeding potential PageRank away from the active target documents.

You were saying that with certainty. It is not certain at all. It's likely to be not true, according to the classic PageRank calculation. (IMG:http://www.highrankings.com/forum/style_emoticons/default/smile.gif)

[Michael Martinez]

The Classic PageRank calculation doesn't eliminate dead links. The testbed eliminated them. Big difference there.

[Alan Perkins]

Michael, I'm not going to argue about how PageRank is, was, should be or will be calculated. The reason I made a comment in this thread was your statement:

If you think about it, those dead links are hurting the documents that the active links point to.  So, you ARE bleeding potential PageRank away from the active target documents.

That's complete conjecture, stated as fact. We can't have that. (IMG:http://www.highrankings.com/forum/style_emoticons/default/smile.gif)

[Michael Martinez]

If you think about it, those dead links are hurting the documents that the active links point to.  So, you ARE bleeding potential PageRank away from the active target documents.

That's complete conjecture, stated as fact. We can't have that.

What we cannot have is people snipping out relevant portions of what I write and claiming I am making statements of fact that I am not.

What I wrote was:

I don't think they would downweight the page, but certainly the dead links would be diluting the value conferred by other links on the page (in terms of the classic PageRank calculation).

I think someone has finally come up with a valid model that fits the expression "bleed PageRank", but it doesn't work the way people think it works. If you think about it, those dead links are hurting the documents that the active links point to. So, you ARE bleeding potential PageRank away from the active target documents.

In the Classic PageRank calculation, a document's PageRank is divided equally among its outbound links. The calculation, as I pointed out, makes no distinction between links to existing content (live links) and links to non-existing content (dead links).

In fact, Page and Brin justified allowing for dead links under a number of circumstances in their original paper:

The text of links is treated in a special way in our search engine. Most search engines associate the text of a link with the page that the link is on. In addition, we associate it with the page the link points to. This has several advantages. First, anchors often provide more accurate descriptions of web pages than the pages themselves. Second, anchors may exist for documents which cannot be indexed by a text-based search engine, such as images, programs, and databases. This makes it possible to return web pages which have not actually been crawled. Note that pages that have not been crawled can cause problems, since they are never checked for validity before being returned to the user. In this case, the search engine can even return a page that never actually existed, but had hyperlinks pointing to it. However, it is possible to sort the results, so that this particular problem rarely happens.