Bolding, Higlighting, Italicizing Text, Does it make any difference? Options

[haylieplum]

If you bold words, italicize them, underline them etc. does it make any difference to the engines? Will it make them take more notice of your keywords, etc?

[qwerty]

I believe so. Bolding and italicizing were specifically mentioned in Page and Brin's original paper that led to Google. Matt Cutts has recently stated that <strong> and <em> are treated exactly the same way that <b> and <i> are, and that certainly implies that they are treated in some particular way.

Basically, it just makes sense. If you do something in your HTML to set certain words apart, it's because you want to stress those words. They're more important, and you want the reader to see that. Of course, that's only if you use HTML elements that are understood to do this. Sticking them in a span and setting the style to big, red and bold won't accomplish that, except for people using browsers that will display them visually.

[lyn]

I'm on board with what Bob is saying. I'd like to underline the part about doing things that...

set certain words apart

I have always believed (without having proof) that the SEs recognize differences more so than specific tags. Meaning that, if you simply make everything on your page bold. it's the same as doing nothing at all.

L.

[Jill]

This is something that would be so easy to prove.

So how come nobody can ever show any proof of it?

I've certainly not seen any.

Matt Cutts has recently stated that <strong> and <em> are treated exactly the same way that <b> and <i> are, and that certainly implies that they are treated in some particular way.

Yes, it implies they're treated the same way. Which could mean that all of them are treated as if they weren't even there, no?

[qwerty]

Well, Matt wrote that an engineer showed him the actual code for what the algo does with text in those tags, so that pretty much says that they're not going to be treated the way "normal" text is. If they were, they wouldn't need their own lines of code.

[Jill]

If they were, they wouldn't need their own lines of code.

Good point!

I still never believe a word that scoundrel MC says though! (IMG:http://www.highrankings.com/forum/style_emoticons/default/wink.gif)

[lyn]

So how come nobody can ever show any proof of it?

Jill, I don't think it's quite as clear cut as, say, checking whether a particular tag or attribute gets indexed. You'd be looking for smaller incremental effects, and certainly nothing as dramatic as the effect of Title elements or anchor text. That, I think, makes it harder to measure (and prove/disprove) their value.

As Bob says, though, it makes sense that a clever algo designed to figure out so much stuff to distinguish between key content indicators across different pages would also be able to recognize the differences between key content indicators within a page.

For myself, I use Heading tags and bold face and italics (or strong and emphasis) for normal markup reasons anyway. I've been writing that way since I was on my Smith Corona, so I'd continue doing it whether it helps SEO or not!

(IMG:http://www.highrankings.com/forum/style_emoticons/default/biggrin.gif)

[qwerty]

Actually, the way I normally write is to stress words that aren't really keywords, like "Believe what you like, but I am not in fact envious of your leisure suit."

That's probably why, while I believe that using <strong> and <em> is likely to be helpful for SEO, I don't do it a lot.

[Jill]

Actually, the way I normally write is to stress words that aren't really keywords

Exactly, which is why I personally think it would be silly to give them extra weight. Any time I've ever seen a site that bolds keywords it just looks so dumb and sticks out like a sore thumb. It's why I would never do it, as it rarely ever makes sense in the true context of a page.

Jill, I don't think it's quite as clear cut as, say, checking whether a particular tag or attribute gets indexed. You'd be looking for smaller incremental effects, and certainly nothing as dramatic as the effect of Title elements or anchor text. That, I think, makes it harder to measure (and prove/disprove) their value.

Yes, it is clear cut. Make a page, bold some keywords, see where it ranks for the next few months. Remove the bold see where it ranks the next few months. Take 'em out again, see where it ranks, etc.

I have not bothered to do it because I strongly believe that bolded keywords just looks stupid, but that's how I would do it if I were to test it, and I'm pretty sure it would be fairly clear what the answer is. (You'd want to try it on a few different sites of varying ages and link popularity, of course, as well.)

[jehochman]

Yabbaduffle
Pifflenotch

(This is a test. I am going to post these words somewhere else too, reversing the bold/not-bold states. Then we'll see which ranks better.)

[Jill]

Which ranks better than what?

You have to put it on and then take it off.

[jehochman]

I've posted two unique words (no Google results) on two forum pages here. On one page the first keyword is bolded. On the other page, the second keyword is bolded.

When we run a search for each keyword, if both times the page with the bold keyword comes up first, the test would support the hypothesis that bold provides an extra ranking boost. To be considered proof, the test would have to be repeatable, and provide consistent results.

[Jill]

I don't think that's really the correct way to test that. (I can't explain why this second, but something tells me that's not right.) I guess I'll have to try my way at some point now!

[qwerty]

It depends on where you post the other version. This may be a very deep page, but it's on a pretty authoritative domain.

But I suppose that if after you switch the bolding from one page to the other and the ranking follows it over...

[MaKa]

I'm quite curious how quickly a deep page like this is picked up by the main SEs. I've mainly got experience with opening sites to SEs after being locked (e.g. behind a form) and found that it can take up to a couple of months for most of the new pages to be indexed.

Had a very quick look (looking at the highest post id on Google query) and the most recent deep page from highrankings I could find indexed was from two days ago.

Anyone have another guess to how long you need to run an experiment like this before you can say something sensible?