Funny Microsoft doesn't mention TrustRank or all the other Google assessments.
Couple different points here -
1. Saying they use 200 signals to rank a page is akin to saying there are thousands of different variables to growing tomatoes.
While both these statements are true, for an SEO/tomato grower, it only takes 4-5 variables to grow good tomatoes and possibly 10-15 to grow "award-winning" tomatoes.
- tomato seeds
- amount of sunlight
- amount of water
- amount of fertilizer
- types of insecticide.
if you want to be the "#1" tomato, you might have to test different nitrogen fertilizer contents, organic vs. man-made insecticides, different watering ratios depending on the sunlight, etc but one still doesn't need to know or even control the 1000's of different variables that go into the photosynthesis and horticulture of tomato growth.
2. I can hear all the people who think Goog is all set to roll-out their analytics algo pointing to this story to prove their point.
But careful research should illustrate exactly WHY MSN could roll out such an algo waaaay before Goog.
First, MSN has NOTHING to lose by imploying such an algo.
They are last in the SE wars and haven't invested billions refining and marketing a "link graph" algo like Goog has.
Consequently, Goog has EVERYTHING to lose by rolling out a completely different algo that has made them market leader and in reality isn't technically broke (yet...)
And if Goog is wise, they would want to incorporate their link-graph(PR) algo AND the browser-user algo, which again would take MORE time to test, refine, and perfect for everyday use.
MSN, on the other hand, can basically scrape everything, start over from scratch, and test more aggressively with this browser/analytics algo.
edited to add - spamming browser rank will probably be easier than PR over the long run however, so their point about that is moot.
So, did I miss it or did they not mention where they're getting all this information from?
Some researchers in Asia...
PageRank stipulates that a surfer will use random link from those available on a page to move elsewhere, but naturally a handful of links on any given page will have much higher probability of being clicked and some others will have much lower, so what they propose for this BR thing actually makes sense - the other question whether this actually results in good enough quality rank.
but pagerank was always weighted to the first links on a page having more weight than those towards the bottom, it was never linear!
|just one of more than 200 signals |
I heard from someone at G release a statement recently, about "100" signals. I'd rather think it's a lot. 100 or 200 is a throw away indicative line.
My thoughts are that PR is an important component of the page quality assessment at Google, which is why it is specified being against the guidelines to "transfer PR".
The moment someone's PR increases greatly, i can almost guarantee a manual check , preceded by an algorithm flag . Too easy to achieve. What's more , you don't know the strength of the link as it's hidden for up to 3-4 months.
[edited by: Whitey at 9:58 pm (utc) on July 25, 2008]
|I heard from some official release recently, about "100" signals. I'd rather think it's a lot. |
If i can extend my analogy, i don't think 100 or 200 is a lot.
I'd say up to 50% of those are the "insecticide formulas"
ie. the spam-fighting signals to PREVENT ranking as opposed to influencing specific "positive growth" rankings
|doesn't mention TrustRank or all the other Google assessments. |
I think when most people talk about PageRank nowadays (in this type of discussion) they aren't talking specifically about the original PR formula, but a link-graph biased algo.
That naturally assumes TrustRank, certain links being weighted differently, etc.
"PR algo" as opposed to keyword/content-biased algos, user-analytics-biased, etc.
I wonder how BrowseRank will operate?
I suspect it will rely on MSIE to feed back information to MS (shades of Phorm and Nebuad?).
If so, with the increase of other browsers (Firefox, Opera) this could backfire anyway.
And, of course, there is the arbitrary definition of "length of time on a page" - some of my browser tabs have been on reference pages for weeks!
It's a better concept than PageRank, I think, in that it (apparently) reflects real-world users, whereas PageRank merely reflects the number/quality of links to a site, which SERPS has repeatedly shown is irrelevant in search terms. Not convinced it's going to work, though.
|I suspect it will rely on MSIE to feed back information to MS. |
If so, with the increase of other browsers (Firefox, Opera) this could backfire anyway.
Actually, i think a good advertising campaign would make this a HUGE selling point. (ignoring privacy concerns for now)
"True input from real-live users like you! Unlike Google"
80% of users still use IE primarily (firefox/opera trending to the more computer savvy/geeky of the population)
I remember something about 3-6 months ago about how last summer or last year google made something like 250 changes to their algorithm, which i think would lend the idea that there are alot of factors (read above 200) that havent been glanced at...
But unfortunately i am at the atlanta airport on my blackberry so i cant post a link to that reference right now, but look for it in the next couple of days unless someone else remembers what im talking about and posts the link or summary...
Here's the reference: [webmasterworld.com...]
The quote from Udi Manber was "Last year we made over 450 improvements to the algorithm."
Heeh, yeah, i just found it.. Sitting at the airport for 5 hours, i got the time...so i wont post it again.
But that 450 could include changes back to the original, so im not sure how accurate my statement is...
But i guarantee that there are those already planning on a way to browser-rank to their benefit just link link building was manipulated..
Id go so far as to say that because google has so much data from google analytics, that they could incorporate that into their algo if microsoft ends up getting some traction on the new wave...not to mention all the various major sites google already has (youtube,blogger e.t.c) and all the information that comes from it
The CNET article referenced above makes a similar claim:
|...the company updated its search formula more than 100 times in the second quarter. |
I found the last paragraphs in the story particularly interesting:
|...And researchers have huge infrastructure at their disposal to try new ideas. |
"My group at Google has at its disposal many thousands of machines, with storage measured in petabytes," Udi Manber, head of Google's search quality, said of Google's search research infrastructure in a June talk. And, he added, engineers are empowered to try their results, with meetings once or twice a week to see how well they worked: "There is no separation of research and development. Everyone does both."
Makes me wonder how often the occasional odd anomalies discussed here are actually a result of "everyone doing both"? Or, are all these research tests in-house only and never seen outside the plex?
In one of the related posts about that, they do an A/B test to see what works or how results are affected.. It goes back to Udi's post on blogspot...
Focusing back on the "article" that started this thread, I must admit, I'm not impressed. I think it's just Microsoft trying to build a buzz and generate some spin for their search engine. In other words, it sounds like a technical piece, but it's really just a publicity piece.
|I'm not impressed. I think it's just Microsoft trying to build a buzz and generate some spin for their search engine. In other words, it sounds like a technical piece, but it's really just a publicity piece. |
you crack me up! Have you seen the TRUTH?
Here's my comment about Udi's "450 changes/diversity" interview that you referenced
( [webmasterworld.com...] )
|Like EVERY company, politician, public figure.... |
National recorded interview = lots of bs spin that may or probably won't resemble truth.
|Makes me wonder how often the occasional odd anomalies discussed here are actually a result of "everyone doing both"? Or, are all these research tests in-house only and never seen outside the plex? |
Well we know of ONE for sure. The "Position #6 Gaffe of Cero-Ocho"
Always good to read any piece with a view to tweezing out:
1. Why talk about this NOW?
2. Is there real value here - anything I can use to inform my actions?
I'm always looking for actionable information, and in general, Google spokespeople have shared a lot more of "real value" than MS has.
But I always ask question #1 as well - and sometimes asking question #1 gives more interesting information than any of the explicit content in an article or interview.
|1. Why talk about this NOW? |
I agree. This piece has more to do with Yahoo talks falling thru, failed Search CEO leaving, keeping stock prices up, but that's a topic for another thread. ;)