Forum Moderators: open
1. I saw a great article [google.com] in the NY Times about How to Optimize for Google.
2. I saw [google.com] a great article in the NY Times about How to Optimize for Google.
3. I saw a great article in the NY Times [google.com] about How to Optimize for Google.
4. I saw a great article in the NY Times about How to Optimize for Google [google.com].
5. I saw a great article here [google.com] about How to Optimize for Google[/url].
I'd guess more often than not, the relevant keywords would show up in the paragraph surrounding the link rather than the link itself.
Does google give as much weight to the surrounding text as the anchor text itself? I used to think anchor was more important but in analyzing a few urls recently I feel that they are indeed giving as much weight to the surrounding text.
Lets take the Google sitemap: [google.com...]
Check the link to: Google Zeitgeist
it has the surrounding text: "Our Company"
This already was in place pre-dominic: [web.archive.org...]
Do a Google search for: Google Zeitgeist Our Company site:google.com
the Zeitgeist page: [google.com...]
does not appear in the results.
One could discuss how "surrounding" the above example is, but examples like this are around enough and easiest checked on ones own site.
As said above, of course Google could decide to only add some weighting to surrounding text if it exists in the target page, or if it has any form of "relevance" to the anchortext or the target page. With the big question what relevance should or could be in Google's eyes...
Not me. However, if this weren't true I'm surprised that I haven't seen anyone offer real world examples which strongly indicated this was the case.
>How about all the other elements on any given page that aren't text that's in immediate promimity or within the anchor text itself? They're all worthless?
I doubt it. One thing that occurs to me is that adding in more than just the anchor text #1) would require a lot of computer resources to do; and #2) likely would have a lot of false hits in terms of relevancy. #2 suggests to me that nobody would think #1 was worth the cost and effort. The biggest problem is that while it makes no sense to use irrelevant anchor text (why use anchor text of "purple penguins" unless the linked to page is relevant for that?), it isn't considered good webmastering practice that nearby text necessarily be relevant. Factoring in the surrounding text to rank the linked to page would end up adding a lot of bad data to the index. So much it might in fact lower overall relevancy of SERPs.
here
article
link
click
next
previous
etcetcera
then check the "paragraph" for other links. If no other links, use the paragraph to detect a "theme" and compare the "theme" to the "theme" of the linked-to page.
If there's one or more other links in the same paragraph, use some words surrounding the text.
Then compare the surrounding "theme" to other backlinks of that url, and you may have something.
Naturally, this takes a lot of processing power, however this can be done in batch mode and does not need to be done on the fly.
I know there are many holes in the idea but I think something "better" than the current SERPS is possible. Do you agree?
Because of the way I conducted linking strategies, my key phrases do not appear as part of the anchor tag, but as surrounding text. For many, many of these (competitive) keyphrases I rank in the top 5 for normal SERPS. And if I do an allinanchor: search then again I appear in the top 5. But my keyphrases don't exist as anchor tag text, only as text surrounding the tag.
So, from my qualitive analysis, it appears that surrounding text does matter. And matters alot.
I also see this!
So I say, anchor text is VERY important.
I didn't say anchor tag doesn't count; I am saying that the text surrounding the anchor does.
It is the allinanchor: search that throws this into sharp focus. Why would I rank well on allinanchor: when the key phrase I use doesn't appear in the anchor tag?
In fact on allinanchor: I am actually #1 or #2 on many of the keyphrases. The only thing that is going to improve my postion in the the normal SERP is a little more Page Rank!
IMO the weight of the surrounding text is trivial (if any) compare to anchor & domain name. The sites below us in the serps have plenty of pertinent surrounding text & incomming links (#5 has 80 PR4+ links but no kw's in domain name). Naturally our 1 page small (3k!) test site was done with very basic title/description/kw/text/image name/alt text/ optimization but nothing over the board. Seems to me that a handful of quality links are much more important than quantity, surrounding link text or age of site with the new Google fresh/permanent-refresh whatever algo.
BTW, just checked and I am better then #5 in the normal SERPS for a couple of really competitive phrases.
I stand by my assertion - mainly because I can observe a real life example.
I'm not "mistaken" about your domain name. I was talking about OUR test. I agreee that there should not be a premium to kw's stuffed domain. We hate it: we will probably have to abandon our branded domain because of it!
do it!
I meant you were mistaken about surrounding anchor text being trivial. If it were trivial I shouldn't feature at all in an allinanchor: search - but I do. In fact I am a little higher in allinanchor than normal SERPs. Without those keyphrases appearing in the anchor tag!
I agree that putting the keyphrases in the anchor tag may also have a positive effect - but I assert that the keyphrases in the surrounding text also has a very postive effect.
Certainly I can obtain some more links with the text as part of the tag - it may improve my position. The problem is that could be down to extra page rank. And I am not going to change my current links because, as Marcia said earlier, if it ain't broken don't fix it. Sure I'd like #1, but #3 is also very good - I'd hate to lose that position for the sake of an experiment. Obviously, if I were on the second page then I would try something different.
BTW, I do see where you are coming from. I'm not trying to have an argument.
[edited by: merlin30 at 8:47 am (utc) on July 23, 2003]
I don't know how it ranks the subset of pages that it finds - but the subset of pages should only include those that have the phrase in the anchor text.
As my pages appear in the subset - regardless of where my page is ranked - Google must consider my pages to be linked with the keyphrase. As the keyphrase is only in the surrounding text it follows that Google must be taking the surrounding text into account.
Or allinanchor is broken and doesn't do what it implies.
You did not answer my question: will you do it? If not: why not? That's friendly advice (that's what we do and we are #1 in both serps & allinanchor for our target kw's). Did'nt you noticed all the smileys I've been putting just for you? Here, what the heck, I'll put an other one :]
[edited by: mundonet at 9:11 am (utc) on July 23, 2003]
Could be - I'm not too sure that this command is functioning as it was prior to Dominic. Looking at a cached page Google now lists allinanchor as one of the terms that only appear in links pointing to the page.
These terms only appear in links pointing to this page: allinanchor
I did an edit to one of my earlier posts (#44) to answer your question.
I appreciate any advice people have. I hope my observations also help other people understand what it happening.
coconutz: Could well be that allinanchor: missed the last update train, just like the PageRank Toolbar and link:? Time for dodo so bye-bye.
So it's important to do both for the future...after all as VitaPlease Pointed out ,there seems to good number of research papers on surrounding text at google.It's good to be prepared for the future after all.
Having said that, I'm not going to stick my neck out and guess what sort of relative bias they use compared with the anchor text, but I would give it equal weight within a limited range such as a short sentence. Google may also have a list of special-cases for dealing with links that read CLICK HERE or DOWNLOAD NOW.
Kaled.
PS
I have considered setting up a site specifically for analysing search engine behaviour. My guess is that several WW members have done this in the past and may have such sites running right now. However, I doubt they are going to admit to this so asking people to give details of the tests they have carried out is pretty pointless. I know I would not answer such questions.
I just did a search for Click here and google is the number 1 result :)
And Google thinks so. The objective of google's PageRank is to determine the quality of a page, not its optimization. If you improve your site for the visitors, your PageRank will increase continuously.
So, once again, I say: think in the surfers before thinking in the SEs
Regards,
Herenvardo
The text of a link must always tell where the link leads. Is better a link like
<a href="article.html">View article</a>
than something like this:
Click <a href="article.html">here</a> to view the article
How true this is rather depends on the visibilty of links. If the links are in a significantly contrasting color/style to the surrounding text then a link that reads View article may be fine. However, on a site where the exact location of links may be less obvious Click Here may be better from the user's point of view.
===============
If Google's algos are biassed heavily towards anchor text, then they are probably not serving the best possible results. They are also wide open to SEOs. This may be the case right now, but I doubt that it will continue.
Most people like to feel good about themselves. For some, believing that they have knowledge or abilities that others lack helps in this regard. Just as some people like to think they can talk to angels (or whatever), I imagine some people will like to think they know the innermost secrets of search engines. Some might, but not as many as think they do.
Kaled.
Regards,
Herenvardö
