Published: Nov 09, 2007 - 01:26 pm
Story Found By: DanThies 1656 Days ago
Category: SEO
This may be a little bit heavy for some readers. Take your time. Read it twice if you have to.
2 Comments
2 Comments
Search Engine Land produces SMX, the Search Marketing Expo conference series. SMX events deliver the most comprehensive educational and networking experiences - whether you're just starting in search marketing or you're a seasoned expert.
Join us at an upcoming SMX event:
Learn more about search marketing with our free online webcasts and webinars from our sister site, Search Marketing Now. Upcoming online events include:
Comments
Its a good discussion piece - I remember when you suggested TSPR as a reason for Florida, and I argued on a few forums that Hilltop features seemed to resonate more with what we saw. But Ive still kept an eye on TSPR because it does hold a number of ideas you can really see Google leveraging.On the keyword scoring issue - I would seriously avoid underestimating Googles information storage capcity. I believe Ive already read Matt Cutts implicitly state that he regards the information storage Google can leverage as without limit (sorry, cannae find quote). Additionally, too often people have said certain things would require too much data storage or processing to implement in search, only to be clearly proven wrong later.In that regard, while I wont argue for Google being able to assign a PageRank value on a single keyword basis, I think if youre not already planning SEO copy with text block analysis in mind then youre already in danger of heading straight to supplemental.2c, and thanks, Dan, for some signal. :)
Thanks Brian, and I think Im with you 100%.We preserved a bunch of SERPs from Florida, and when we got hold of Personalized search, we could easily produce nearly identical SERPs simply by picking the wrong topics - slap Computing and Education on it, and the "laptop rentals" SERP looked exactly like it did during Florida - lots of university computing centers laptop rental policy pages and such.I have no doubt that there are better ways to do it, perhaps as yet undiscovered, but LSI and Topic Sensitive PageRank just dont look like they can scale up to the point where theyre worth doing. TSPR might be possible with 50 topics, but probably not with 53,000. Its not just about storage capacity, its processing time and all that. The web is really really big.I agree - if you arent expecting some kind of visual page segmentation and text block analysis, youre probably going to be very sad very soon.