is alt text as good as anchor text?

[inertia]

Just a quicky! I've been reading a bit about the way Google treats internal linking and such. Most of the sites that I work with have the company logo as the link at the top of the page. This is fine, I add some decent alt/title text and leave it at that.

My question is... when an image is linked and has an alt and title tag in place do they act as anchor text? Do they swing as much weight as anchor text? Would i be better replacing the image for a textual link (i wouldnt be replacing the logo, just turning the link off and adding a text home link somewhere else high up in the code)?

It wouldnt be too hard to figure this out... but this is the lazy approach first.

[kgun]

Originally Posted by inertia

My question is... when an image is linked and has an alt and title tag in place do they act as anchor text? Do they swing as much weight as anchor text? Would i be better replacing the image for a textual link (i wouldnt be replacing the logo, just turning the link off and adding a text home link somewhere else high up in the code)?

Haven't you asked a similar though not identical question earlier? My fast answer is that

1. Anchor text,
2. alt text and
3. title

are three different animals. What about combining an image link with a text link that I thik was what you asked about in that similar thread?

[inertia]

I think i was asking about nofollowing the image link when the text link is next to it. I may have asked this before, but i cant remember my own age sometimes kgun!

My train of thought on this one relates to some articles I posted about Google ignoring all other anchor text apart from the first link on the page, when there are several links that point to the same page.

The logo at the top of the page is usually the first link back to the home page and (by the logic of the above findings) when you add alt/title text to that logo (does alt/title text replace anchor text completely?) this becomes the first anchor and only counted anchor text....ff you believe these findings that is!

I may not have explained myself well... but do you see what im thinking?

[inertia]

Sod it! Ill setup an experiment and get back to you...

[Chris_E]

Be interested to see your results!

[Webnauts]

Originally Posted by inertia

My question is... when an image is linked and has an alt and title tag in place do they act as anchor text? Do they swing as much weight as anchor text?

The linked images do not have as such weight as a text link.

Originally Posted by inertia

Would i be better replacing the image for a textual link?

Yes.

[Webnauts]

Originally Posted by inertia

My train of thought on this one relates to some articles I posted about Google ignoring all other anchor text apart from the first link on the page, when there are several links that point to the same page.

What you read in the link you posted, that is my position too. I read the original post of Rand a while ago and I had to laugh to be honest.

[kevsta]

so you would agree with the findings on SEO scientist John?

i havent noticed a lot of difference between alt text and anchor text to be honest, but have never tried swapping on for another on something already ranks or setting up proper testing either to be fair

[inertia]

Originally Posted by Webnauts

The linked images do not have as such weight as a text link.

That is exactly what I'm trying to find out. I often think about changing sites image based menus for textual based CSS equivalents but i always think... is this a waste of time? If i add alt/titles will this have the same effect?

Ill think of a way to test this and do so... Guess thats the only way to put my mind at rest.

[kgun]

Intuitively I agree with John and thought of saying what he said above

Originally Posted by Webnauts

The linked images do not have as such weight as a text link.

in my first post.

Links are one of the most important elements in the WWW and at the heart of the Google algorithm. Google computes the link matrix of the web and inverts it, what I personally call:

"The adaptive inverted link matrix model of the web".

Google is the dominant SE and so long I have not seen a better SE model.

<Technical comment>
Note the word adaptive. That takes care of modifications e.g. incorporating trust and human intervention into the system. Google use Bayesian filtering. The Kalman filter is a recursive bayesian filter. There is an adaptive version of the Kalman filter where it is possible to manually modify the filter by disturbing the elements in the system covariance matrix. That is in a nutshell why I use the word adaptive above.
</Technical comment>

[2fk]

The alt acts as the anchor text when you use image as a replacement for an text anchor..The alt attribute do have importance for seo..Robots cannot read images and alt attributes are very important for image search..

[inertia]

Originally Posted by kgun

<Technical comment>Google use Bayesian filtering. The Kalman filter is a recursive bayesian filter. There is an adaptive version of the Kalman filter where it is possible to manually modify the filter by disturbing the elements in the system covariance matrix.</Technical comment>

Erm... come again?

Ive read the Stanford paper from page and brin. I found it quite interesting that the whole idea behind Google was pre-conceived and (like many .coms) didnt just "happen". They planned for it to be "massive" from day one. Clever cats - anyway im digressing!

The alt acts as the anchor text when you use image as a replacement for an text anchor.

This is what ive heard a million times before but does it... i will investigate!

[kgun]

Example code from the home page of the second link in my signature.

<a href="http://www.multifinansit.no"> (This is an ordinary link)
<img src="/images/norge dot gif" width="68" height="50" border="0" (This is an image link)
alt="Norwegian flag" (This is an alt attribute describing the image link.)
>
</a>

[inertia]

So what youre saying is... alt text is an attribute and in no way applies to the link matrix? But doesnt the adaptive nature of googles link model mean that they could place as much emphasis on alt text when no anchor text is present?

[kgun]

Originally Posted by inertia

Erm... come again?
Ive read the Stanford paper from page and brin.

I have not. I have only skimmed the mathematics and the idea of scientific citation. Show me a better model, please.

Originally Posted by inertia

I found it quite interesting that the whole idea behind Google was pre-conceived and (like many .coms) didnt just "happen".

Cryptical. Please explain.

Originally Posted by inertia

They planned for it to be "massive" from day one. Clever cats - anyway im digressing!

Agree on the first part of your statement, but not on the last.

Originally Posted by inertia

So what youre saying is... alt text is an attribute and in no way applies to the link matrix?

Where did I say that?

Originally Posted by inertia

But doesnt the adaptive nature of googles link model mean that they could place as much emphasis on alt text when no anchor text is present?

Of course they can. The Kalman filter is used in hospitals - laser gyros, to cut wood optimally (economically) in the tree industry, to predict the position of space ships etc. etc.

Example:
The new (estimated / predicted) position of the space ship is a weighted average (with the Kalman gain matrix as weights) of the old position and the new measurement. You can modify that in any way you want. If you wan't to hit the target that modification may be non optimal.

Note: I do not say that Google use the Kalman filter. There are numerous Bayesian and other filters, where the Klaman filter is one (linear) based on (too) restrictive statistical assumptions. So you can modify your algorithm in an (yes) infinite number of ways.

I only use the Kalman filter as an example of Bayesian filtering know to be used by some SE's.

[inertia]

Originally Posted by kgun

I have not. I have only skimmed the mathematics and the idea of scientific citation. Show me a better model, please.


Cryptical. Please explain.


Agree on the first part of your statement, but not on the last.

Like i said, digression on my part. Lets leave the Google history for a related thread.

Originally Posted by kgun

Of course they can. The Kalman filter is used in hospitals - laser gyros, to cut wood optimally (economically) in the tree industry, to predict the position of space ships etc. etc.

Example:
The new (estimated / predicted) position of the space ship is a weighted average (with the Kalman gain matrix as weights) of the old position and the new measurement. You can modify that in any way you want. If you wan't to hit the target that modification may be non optimal.

Note: I do not say that Google use the Kalman filter. There are numerous Bayesian and other filters, where the Klaman filter is one (linear) based on (too) restrictive statistical assumptions. So you can modify your algorithm in an (yes) infinite number of ways.

I only use the Kalman filter as an example of Bayesian filtering know to be used by some SE's.

Not being a mathematician means i dont have a clue what youre talking about. I could spend the next hour, researching your post and deciphering the meaning of all those alien words but i havent got time. Can you dumb it down for me?

[kgun]

Originally Posted by inertia

Can you dumb it down for me?

The alt attribute on images is an important (W3C recommeded) way to explain what is on the image for surfers with text browsers and similar. It is of course a ranking factor. I doubt that an attribute is as important as the link.

One minor argument to support my intuition or speculation if you prefer. What do you think is easiest to spam an URI or an alt (text) attribute?

[inertia]

Originally Posted by kgun

I doubt that an attribute is as important as the link.

Originally Posted by Webnauts

The linked images do not have as such weight as a text link.

Originally Posted by 2fk

The alt acts as the anchor text when you use image as a replacement for an text anchor.

Contrasting opinions on this one... Ill do some experimenting.

[kgun]

Exellent observation. Look forward to your conclusion.

[Tig]

I've switched from text links to image links with alt tags on almost all my client sites. I believe it looks more professional, and helps with branding.
I started doing this several months ago, and have seen no detrimental effects of doing so. In fact, I'm a bit higher in my targeted search results than I used to be, but that could be because of other SEO practices I've changed also. I didn't think of doing it on it's own to see how it would change my results, but it definitely didn't hurt anything.

[craigmn3]

Originally Posted by Tig

I've switched from text links to image links with alt tags on almost all my client sites. I believe it looks more professional, and helps with branding.
I started doing this several months ago, and have seen no detrimental effects of doing so. In fact, I'm a bit higher in my targeted search results than I used to be, but that could be because of other SEO practices I've changed also. I didn't think of doing it on it's own to see how it would change my results, but it definitely didn't hurt anything.

Call me a chicken but when I have done this....and I have I have always left a text based link somewhere else on the page. It doesn't matter how good you look if you can't get indexed. Its a belt and suspenders approach

[Andilinks]

Its a belt and suspenders approach...

An approach I used too for years until the recent advent of the "over-optimization penalty."

I have paid very dearly for using too many internal links, it has been an extremely expensive error (which Google requires you to find on your own while your revenue spills down the drain).

I have heard that image links back to your default page should bear an alt tag of "HOME" and have been using this for some time but I'm not sure of its effectiveness. I'd be interested to hear opinions on that.

[datetopia]

Some sites have texts menus while some others use image menus. Theoretically crawlers should be able to manage both situations.

[Orion]

alt should not contain keywords per se.. it is to describe the image to those that can't see it use it properly...

title="" is to provide additional information to a visitor

and yep text links are more powerful than image links

[jordanmcclements]

I agree with Kguns first post, and find that having anchor text directly above the image AND alt text works great for me...

[amxfan]

Along these lines.

I'm running a cart and it links product names above the image, the image itself and under the image a "more details" link to the cart page. I was told it would be better to add a nofollow to the images. Is there a way to make all images a nofollow site wide using .htaccess? and would this do any good since I also have text links going to the same page?

[kgun]

Originally Posted by kgun

Haven't you asked a similar though not identical question earlier?

Here nofollow on image links is that important related thread.

Originally Posted by amxfan

Along these lines.

I'm running a cart and it links product names above the image, the image itself and under the image a "more details" link to the cart page. I was told it would be better to add a nofollow to the images. Is there a way to make all images a nofollow site wide using .htaccess? and would this do any good since I also have text links going to the same page?

Can that thread help you?

[amxfan]

It does and doesn't help.

It did state that only having one link is better then the 3 I have now. But it also stated to add the image folder to the robot.txt file. This only stops the SE from indexing the images and does not stop them from following the link. I can not make the images nofollow though the admin side of the cart. Is there anyway of stopping the SE from following links on images Globally?

[kgun]

Using .htaccess on your root domain and on subfolders, you should be able to get the control

Example

1. On the root folder you restrict (deny) some IP's IP Addresses of Search Engine Spiders (there are a lot of related sites that you find here: DigitalStart.net: The starting point for English speaking surfers and webmasters)
2. On subfolders, you restrict more or allow more.
3. Note .htaccess on the root folder carries its information to the rest of the site, so remeber to allow from all etc. if that is your intention on subfolders.

My priority is this:

1. Use robots.txt if that does the job.
2. Use markup if that is better.
3. Use apache server configuration if that is your preference.
4. Use scripting if the above do not solve your problem.

Note: If you deny an IP, you deny all domains that share that Ip: YouGetSignal.com - Find Other Web Sites Hosted on a Web Server

Related threads:
Blocking IPs for Countries will it affect SEO

Long thread.
How do you know if a competitor is using Negative SEO against your site?

[kevsta]

Originally Posted by inertia

This is what ive heard a million times before but does it... i will investigate!

yes it definitely does, you don't need to investigate. i haven't noticed very much (if any) difference in weighting either and have tested reasonably extensively over time now.

[inertia]

Well, this is what ive setup...

I have put a link on my cash cow site (gets spidered roughly once a week) to page A. Page A has some blurb and a link (the only link on the page) to Page B with a very low competition keyword (i wont put it here coz this page will probably out rank it). Page B is completely isolated apart from this link and also has one sentence which has no mention of said keyword.

Firstly i have linked from page A to page B using a standard anchor with link text. When both pages are indexed ill check the position and then change the link - ill go for an image link with alt next, then an image with title, then an image with title and alt... etc.

I might try and figure out the nofollowing images issue which Kgun cited before but ill have to think about that setup.

Im aware that there are various other factors like the complete irrelevance of the site content and the anchor text but these obvious factors may reveal more about how google handles links.

Im feeling a bit like i left half my brain on the pillow this morning so if anyone can see any errors or has any suggestions then let me know.

Ill post results periodically.

Ill post results periodically.

[inertia]

Is there an echo in here?

[kgun]

Originally Posted by inertia

Well, this is what ive setup...
I might try and figure out the nofollowing images issue which Kgun cited before but ill have to think about that setup.

Im aware that there are various other factors like the complete irrelevance of the site content and the anchor text but these obvious factors may reveal more about how google handles links.

Im feeling a bit like i left half my brain on the pillow this morning so if anyone can see any errors or has any suggestions then let me know.

Ill post results periodically.

Ill post results periodically.

We will wait for results, but I think it will be very difficult to draw a general (scientific) conclusion.

Originally Posted by inertia

Is there an echo in here?

You think of this

http://www.webproworld.com/marketing...tml#post387294

thread?

There are distinctions from yours IMO.