In perusing a press release of ours as written by our Communications Director and published on PR web, I noticed that there is rotating text in that snippet of the content that appears as wrap-around text in a larger, italicized font (there’s a name for that, I’m sure :dunno: ). So, everytime you click on the link that leads to the press release, a new snippet of text taken from the body appears in this section. And it does appear as text, not an image. It seems to choose any text that is contained in quotation marks. I imagine that this is a javascript function built into the code on the page that allows this feature. It struck me that the reason they set things up this way is to encourage frequent spidering of the press release, given that there is effectively a new version of it every time a user views it in the browser. How close to the mark is this? Does anyone use this technique to encourage spidering? If so, is it effective?

Welcome Rob! :wave2:

Well, it's a good theory, but not the case :)

First, spiders and bots can't read Java so if the snippet is a function of Java (and it most likely is) it will never be seen by Googlebot or any other bot.

But, IF it were seen by Google (just as an example) changing the snippet every time a refresh occurs would not be a good idea, IMO. Here's what happens...

The page changes and the spider visits. It downloads the text into the database where it is sifted and sorted and a decision is made whether it will be indexed or tossed. If it's indexed, next it goes through a process that ranks the page. If the text is constantly changing minute by minute, your position on the SERPs could, theoretically, be in constant limbo - up, down, up, down.

People get carried away with the "the spider needs to visit frequently" thing. It's not that the spider needs to visit frequently. It's that the site needs to stay current and present updated information (not stale, outdated information). Adding press releases is one way to do this, but the Java script function isn't helping. If your CD wants to put snippets on the home page for the sake of the engines, I suggest that s/he choose one, keyword-rich snippet to put up and leave it be.

It struck me that the reason they set things up this way is to encourage frequent spidering of the press release, given that there is effectively a new version of it every time a user views it in the browser. How close to the mark is this? Does anyone use this technique to encourage spidering? If so, is it effective?I spoke to Jon Glick recently about this (he's a former Senior Search Mgr. at Yahoo), and he explained a technique called "shingling" which is in use now with all the algorithms -- what it means is that minor text changes will have no impact on spidering, ranking, etc. Adding a new post to your blog counts as new content; minor changes or rearranging existing text does nothing.

Hope this helps....

Matt

Thanks Matt. Good to know!

Any qualification on what counts as a minor change? How many words or whatnot?

No, he didn't (couldn't?) give any quantifiable measure like that. The main points were that rearranging content wouldn't fool the crawler, and that minor changes wouldn't have any impact. "Search engines are looking to find changes that a human reader would consider significant."

My gut feeling is that it's relative to what else is on the page. A page with only 3 short paragraphs of text could be deemed to have changed if two of those paragraphs are substantially rewritten. That may not work on a page with 1,000 words or more.

that's a very interesting perspective when it comes to seo tweaks. maybe that plays a role when it comes to placing an emphasis on link optimization versus on the page changes/tweaks.

Thanks for the info Matt. I would guess it is a matter of a percentage of what's there and what changes. It probably is an indication of the relative weights of linking and on-page optimization at Yahoo and the other engines who use 'shingling'

It does make sense since search engines want to present the most relevant page and minor changes generally won't make a siginificant difference in the overall worth of a page. More likely just someone trying to optimize it to boost ranking. Or perhaps just fixing typos or rewriting things to make them read better.

Minor changes can certainly make for a better page, but they probably don't increase the relevance of a page to a given query much.

It might also have something to do with search engines relying more on latent semantic indexing. A few changes in words shouldn't affect how a page is viewed to an LSI algorithm.

Welcome Rob! :wave2:

Thanks!


...If your CD wants to put snippets on the home page for the sake of the engines, I suggest that s/he choose one, keyword-rich snippet to put up and leave it be.

Actually, this function is managed by PR web, so our CD didn't really have anything to do with it. He was the one who pointed it out to me, and asked me why they might choose to integrate this function on the page. The above was my best guess.

I somehow knew that many of you would answer in the way that you did. It certainly makes sense to me, being a big believer in reader-first web copy.

Thanks for all of the responses in general, guys.