Is the "title" attribute of the anchor tag read?

[storewebmaster1]

The anchor tag <a> has an attribute called "title" that will show a screen tip in some browsers. Does anybody know if the googlebot pays any attention to this tag?

[piskie]

Good question, I would also like to know.

[jomaxx]

It's not that hard to test if such text is indexed. I have done so recently and the result was negative.

[jimbeetle]

Looks like the consensus on this forum is that the title attribute is read:

[webmasterworld.com...]

It doesn't carry a heck of a lot of weight but quite a few folks use it to firm up secondary keyword phrases.

Jim

[jomaxx]

Consensus or not, the link and image title attributes are not read and indexed by Google. If anyone can provide a counterexample, I will happily change my opinion.

Are they read and not indexed, but used somehow to "firm up" a page's ranking for various keywords? Probably unknowable, but I tend to doubt it.

Do they contribute to a page's overall word count or page size and therefore have a *negative* influence somewhere in the algorithm? Again, probably unknowable, but this kind of text would normally be such a small proporion of a web page that any effect would presumably be microscopic.

[Marcia]

It can be used in lieu of keywords in link text when an image is used for a link to a page, or when a generic word such as "Products" is used for a link, the reason being that the page receiving the link benefits from keywords in link text.

The question is whether the title attribute carries any degree of benefit to the linked-to page in the same way that link text does, even a fraction. Sometimes there's no other option.

jomaxx, did your research check into that aspect as well as the page the title attribute was on? I'm wondering how it would be tested - as in when we see at Google that certain terms only appear in links pointing to a page, not on the page itself, when viewing the highlighted keywords in the cache.

[jomaxx]

No I didn't look at the possible indexing/ranking effect on the destination page. That's a good question, and at least the indexing aspect is quite testable.

[jomaxx]

... i.e. testable by creating an link in the form
<A HREF="whatever" TITLE="scoobiedoobie">
and subsequently looking for any pages related to "scoobiedoobie".

[jimbeetle]

jomaxx -- after I hit submit there was your post above. And yeah, the "pages related to" part confuses what I said below.

>>It's not that hard to test if such text is indexed. I have done so recently and the result was negative.

I don't think that's the same as it not carrying any weight as part of the on-page criteria. I'm not sure if I'm going to be able to explain exactly what I mean, so please bear with me.

Not everything on a page that is indexed is searchable on the SE. An img link will be indexed and give weight to something but -- no matter how many keywords in it -- will not show up as part of a Google snippet. Similarly alt tags.

I think what I'm saying is that, depending on SE, most everything on a page is "indexed," and different items carry different weights in the respective algos, but only the visible (hopefully) text is "searchable" by the user.

Does that make any sense?

Jim

[buckworks]

I make frequent use of TITLE tags on one site for usability reasons, but couldn't get any of the TITLE tag text to appear in Google snippets, even searching for "the exact phrase in the TITLE tag."

On the other hand, text in ALT tags DOES show up in Google snippets. What that says about how the two tags are weighted, I don't know, though.

[jomaxx]

jimbeetle, I have the opposite hypothesis - that stuff not searchable like comments, meta keywords, title attributes and so on are probably thrown out. My reasoning is that if Google doesn't think it's valid data to index, then why would they use it to determine relevancy? But since there's no easy way to test our beliefs I stuck all that speculation under "unknowable".

Snippets are something else again. First of all, I would assume that Google predigests pages for speed - compressing them, removing common words and punctuation, removing capitalization, etc.

It looks like snippets are generated on the fly from the cache and NOT from the predigested page that would be used for determining keyword relevance. That suggests to me they aren't connected to the ranking algorithm, but I could be wrong.

[Dumpy]

In their advertising for "Google in a Box", where you can buy a server with Google installed ($20,000 to $250,000)they say:

"For each query,Google factors in 100 variables, including anchor text, URL patterns, fonts and positioning data to calculate relevance."

Now we need the 100 variables and how they are factored!

[jimbeetle]

>>that stuff not searchable like comments, meta keywords, title attributes and so on are probably thrown out

No, no. Every SE uses some other on-page criteria in its algo to some extent or another. Really, really read that thread I linked to above. There are some real heavy hitters talking there who are much better informed than I am about the Google specifics.

Speaking of heavy hitters, where are they when you need them?

Jim