The prediction is based on a tool I created (my OWN algorithm - ha, take that Google!), but it has tracked pretty well with some of the minor algo tweaks and tests that the Google engineers have been running. Read about my prediction on my DOFOLLOW Blog. (Feel free to post a comment with your DOFOLLOW links in them also - just don't spam the crap out of me ) On Edit: this is NOT a PR update thread - I don't care when the toolbar PR update happens.
The tool runs as a Windows Application that connects to a SQL Server 2005 database that I created. Right now, I'm manually entering data, but if I can find a way to automate the data delivery process, I might put it up for sale. Until then, it's proprietary to me and me only. I started my career by getting a Computer Science degree and learned to code. I then got a business degree and started a web design company (which I knew nothing about, LOL). But, I use my programming skills to create applications useful to me. If I decide to webify the tool, I might even put it up for free, but until then, no there isn't a way to see a live example of the tool.
I do something similair to your tool manually on a small scale and have been waiting for an "earthquake" quite a while (far away from daring a prediction), well just went back to my PC for the last control today and saw some of those jumps in certain serps that I know are usually the prelude for major changes before I read your post.
Yep it feels good to be in the top ten....Since I have been preparing for this for awhile hint co occurance and LSI I have nothing to worry about...
I've been testing some techniques to take advantage of the new dynamics presented by co-occurance and LSI, but it remains to be seen if they will become major parts of SEO anytime in the near future. The technology is there - a correct implementation is not (at least that we know of ). I think there's at least 1 full generation of the Google algo separating us and the time when co occurance and LSI actually become enough of a factor for many SEOers to pay attention to.
Nice article. I do hope you are right. I'm seeing improvement for a couple of my sites. As for which weekend, maybe Columbus day. G likes 3 day weekends just in case they need to do a rollback, not that that has ever happened. /*tom*/
Seriously... what have U been smoking? Every week or so some crack pot comes along with a prognostication about what is up at the Plex…. There have been no major infrastructure changes in quite some time and general tweaks to the algos is an ongoing concern… not in a big move a-la Big Daddy, Panama and the other major updates we used to name in the old days Things are far more fluid these days…. You are either one of those knowledgeable in the art ( as G like to say ) and know this already and are creating some link bait, or you haven’t spent enough time in search engineering to understand fully the indexing and retrieval processes. Which is it? As far as Latent Semantic Analysis is concerned, the purchase of Applied Semantics (the guys that coined LSI) was primarily for the AdSense program than the RI (regular index). It most certainly could be at play in the RI, but there s no direct evidence that it is another layer in the IR (indexing and retrieval) process. Ya want to learn more.. sure… try this; http://64.233.179.104/search?q=cach...usage.pdf+applied+semantics+white+paper&hl=en An interesting and related patent recently published by G is also of interest if you like LSA concepts and methods; Document compression system and method for use with tokenspace repository http://appft1.uspto.gov/netacgi/nph...023.PGNR.&OS=dn/20070220023&RS=DN/20070220023 Now, another important element in the relevance world at Google are the 5 patents released late last year relating to ‘ Phrase Based Indexing and Retrieval ‘ which is far more recent then the LSI stuff which was waaaayy back in 2003-04; I have posted links to most of the info around here; http://www.huomah.com/search-engines/algorithm-matters/phrase-based-optimization-resources.html While we’re at it, you may also want to read up on - Method and apparatus for learning a probabilistic generative model for text http://patft.uspto.gov/netacgi/nph-...PN.&OS=pn/7,231,393&RS=PN/7,231,393withMethod These are also more recent methods that Google has put out that is important to understanding how Google is working with relevancy. ..and even; Method and apparatus for characterizing documents based on clusters of related words http://patft.uspto.gov/netacgi/nph-...7,231,399.PN.&OS=pn/7,231,399&RS=PN/7,231,399 I could go on and on.. I am sure U are getting the idea here.... I spend a great deal of time working with some pretty smart minds in the SEO industry picking Google apart every day of the week. Your statements in this post make me think you may have not completely done your home work as to how the IR process works at Google…. NO ONE that I play with ( many reputable, well known SEO folks) would ever make such statements as you have because we understand the complexity behind what Google does and have been around long enough to remember the days of major infrastructure updates ( which is more complex than simple ‘algo updates’ as you call them) …and I haven’t even touched on the Local, Universal and Personalized search aspects to be considered. Then also seem to have missed how spam filter systems affect the ranking mechanisms in the over-all process. Feel free to email me if ya want and I can turn you on to even more info…. Just don’t go around being the SEO Guru that is ‘in the know’ …. There are some of us trolling around here that have considerable search engineering and indexing retrieval knowledge. You are making bold claims that don’t ring true due to the lack of a comprehensive understanding of how things work, the history and evolutionary processes that Google has been implementing. I am not flaming.. I love this stuff and would be happy to talk shop with you, I just get irked by folks thinking they have all the answers… it simply isn’t that easy. Peace………….
Certainly, but as I mentioned, IF it was actually implemented in the RI, it is merely one layer in the IR and ranking mechanisms.... I was a big LSI proponent a year or so ago. and then Bill ( Slawski ) turned me onto some oter goodies... So I learn from those I play with. I am simply suggesting ya'll may want to look into some of the many other technologies G has introduced SINCE LSI ...back some 4 years ago. He has been kind enough to let me play with him on some stuff lately and even he doesn't have all the answers... and I think he's pretty knowledgable in Search ya know? .. don't put all of your eggs in that particular basket as there is much more to this gig.... why limit yourself if you are serious about the studies?
Hey thegypsy, As I pointed out in my blog post, I'm simply reading the tea leaves and making a prediction. A prediction can be right or wrong, but I'm starting to see the signs. First there have been the wild swing in SERPs when there has been obvious testing going on. Second was the declaration by Matt Cutts that an infrastructure upgrade is in the works. Third is the fact that I have a system I use that's in place that is a pretty good indicator of when things are swingin' over at the Google command center. The fact that changes are being tested, then rolled back, points to a larger change coming soon. Especially because of the frequency of those tests since late July. I never said that the change would be on the order of Big Daddy or Panama, but it's all relative - a major change now probably wouldn't come close to the upheavals we saw with those updates. As a software and database engineer, I was actually offered a job at Google in the wild days before the IPO (turning it down was one of the biggest mistakes of my life, at least according to my wife ). My math skills are not the best, but my database skills (even in IR) are pretty good. I wrote my own Windows search utility that could search an 80 GB hard drive in less than .1 second. The client I wrote it for is a design and marketing company with literally millions of files that were in complete disarray and strewn across a network of 30 computers without a centralized server. So I may not be an expert, but I understand the basics of large scale text indexing. My algorithm definitely wasn't as sophisticated as Google's but pretty good for a one man job. I've never called myself an SEO guru, but if you want to label me as a prognosticator, go right ahead. That makes my a "prognosticator" instead of a "Self Proclaimed Prognosticator" . Since starting my web design business, I've become more and more of a webmaster instead of a database guy, and it has given me a perspective that I didn't have before. I'm in the unique position of understanding some of the behind-the-scenes technology and experiencing the trials and tribulations of a webmaster trying to make sense of the Google experience. I've also been a Google user since almost day-one (I even know it's original name ) And please, please don't start the LSI debate again - I read all of James Dean Nash's LSI thread before and it gave me a headache. I also didn't intend this to be link-bait. If I wanted to start a link-bait campaign, I'd Photoshop a picture of Jessica Alba so that she appears naked, then post it on all the "Oh-my-gawd-what-is-that-celebrity-doing-now?" message boards. I am starting a link-bait campaign to attract more webmasters to my blog, but I'm not using posts like this to do it. Whatever you do, don't consider me an expert at any of this stuff. But, I am just knowledgeable enough to be dangerous. BTW, I didn't take this as flame bait at all - a lively discussion is a real discussion. The pats on the back are nice, but someone with real questions gets the juices flowing. (and for those following this thread, Mr. Gypsy is right - it's only a freakin' guess)
Wel, like I said... feel free to PM and I shall pass along my email addy. I don't find much oportunity to get into these kind of goodies around the boards ( here and elsewhere) .... I'd be happy to keep ya in the loop as interesting stuff comes across my desk... Once ya get the swing o patents ( and yer eyes stop bleeding) they get easier to get a grip on. An educated guess is always a little better than a wild one... he he.... .... and I still want to know what yer smokin cause I am dyin here..... bwaa ha ha ha ha ha BTW ... While I enjoy the technical stuff.... I am an avid (rabid?) SEO Ranter... drop by the Blog from time to time.. ( will add yers to my reader); The Fire Horse I have a pretty funny illustration gag of Matt you may get a kick out of as well Matt Cutts New Software ( I started in the web design biz as well back in 1998)
Wow. I wish I had your skills. I could have used my "million dollar idea" to make a million for myself.
My "million dollar idea" COULD have been a grueling job at Google working 18 hours a day on improving retrieval technologies then cashing out after the IPO. Google was interested in me because I have degrees in both Computer Science and Business Administration and I'm a pretty much a nerd in general.
I agree if you study the G patents you can definitely get a good feel for what their doing..Especially when they implemented the G quality score and relevance factors in their adwords...
As far as LSI I'm not going their but just to let you know what do you think adwords quality and relevance score is based on...LSI and Co occurance but nuff said...not going their Well I am not the one who starts some sorda algo that is only a guess on what might happen rather than actually doing something about it...And everyone knows that they test in the summer and their is probably no big update coming considering they update all day every day.. Not to say that what your doing is not very ambitious because I think it's great what your doing but I think I would put my effort in other areas that would bring results not some stock market algo that is not predictable.