homepage Welcome to WebmasterWorld Guest from 54.167.173.250
register, free tools, login, search, pro membership, help, library, announcements, recent posts, open posts,
Become a Pro Member
Visit PubCon.com
Home / Forums Index / Microsoft / Bing Search Engine News
Forum Library, Charter, Moderators: mack

Bing Search Engine News Forum

This 32 message thread spans 2 pages: 32 ( [1] 2 > >     
MSN Search - Relevance. Where is it?
BigUns




msg:1540292
 6:04 pm on Feb 1, 2005 (gmt 0)

I have read, with some fascination, many of the postings in recent weeks in regard to the new MSN search. These postings seem dominated by those who really like the MSN SERPS. I disagree. I am reminded of the universal hoopla and fanfare that greeted Netscape 6 when it debuted - oh what a wonderful standards compliant browser Netscape 6 is, and Internet Explorer is terrible blah blah. Actually, Netscape 6 su*ked. And, in a similar fashion, MSN search su*ks. Many of the searches I do in Google or Yahoo with good results are gibberish when done in MSN.
Among my experiments are 'geographical' type searches(example): Keyword1 Keyword2 western PA

In one search, only one of the top ten results were in western Pennsylvania - 9 of the 10 were in: western Virginia, western Michigan, and several in western Australia.(!) Google is very good at this type of search; Yahoo is excellent; MSN falls flat. In order to get meaningful results in MSN, you actually have to put quotes around "western pa" - then the results are much better. I have also noticed that in other searches MSN gets much better with portions of the search enclosed in quotes - something I virtually never do in Google or Yahoo - and the typical searcher won't either. I have concluded, on the basis of a variety of searches (mostly 4 word), that MSN often serves up random gibberish because apparently:

(1) Word Proximity means nothing to MSN
(2) Titles mean almost nothing to MSN

It would be interesting to know if others have their own observations in regard to (1)Word Proximity and (2)Titles - and the types of searches that seem to support or refute these assertions. Some of you who are pleased with MSN may be using less keywords or 2 keywords with a strong Inherent proximity bias - (example: Boston MA)- which perhaps explains why those results may be better / more relevant.

I have no vested financial interest in how good or how bad MSN search is or isn't. I don't make money from any website and I have no $$ bias relating to SERPs. I use what works, and for the last 5 years, what works is Google - although Yahoo is becoming my favorite SE for many searches - something I would have regarded as heresy six months ago - but that's a different subject. I don't believe that MSN Search is bad - I believe it is Breathtakingly Bad. I do hope it gets better. Feb 1 2005

(On a related matter, has anyone else seen MSN indexing both www.MySite.com AND www.MySite.com/index.html. I have - and this will cause SERP problems ...)

 

zeus




msg:1540293
 7:01 pm on Feb 1, 2005 (gmt 0)

www.MySite.com AND www.MySite.com/index.html yep got it here to.

About the results I agree fully the results are not the best, I do rank goog for my main keyword, but I miss a more theme related search result, there are many results where a site is ranked goog because they only have a single page about the topic, but the rest is whole other topic, so you dont get a perfect result - A search of a site about NASA, then you get a page from lego where they have NASA Bricks, just a example, not found in the search.

Also many pages are missing from the big sites and internal page dont rank well eather.

HughMungus




msg:1540294
 7:28 pm on Feb 1, 2005 (gmt 0)

Interesting that you say you have no vested interest yet you joined the same day that Microsoft released its search and your first thread is about Microsoft's new search.

I did the same search you did. The first 7 results were for Pennsylvania.

BigUns




msg:1540295
 7:50 pm on Feb 1, 2005 (gmt 0)

HughMungus:

You couldn't possibly do the same searches I did because I didn't tell you. But here's one I just tried (and so should you) used cars eastern MA. The first result in MSN is a Kerry-Edwards Blog. To appreciate how bad the MSN results really are try Google and Yahoo.

GameMasterM




msg:1540296
 8:25 pm on Feb 1, 2005 (gmt 0)

Wow, what did I do wrong? My sites are so far down I stopped looking. What is relevance anyway, I have adequate optimization and direct hit keywords. Can't give up so I'll hope they just show up.

Crush




msg:1540297
 8:35 pm on Feb 1, 2005 (gmt 0)

"I have read, with some fascination, many of the postings in recent weeks in regard to the new MSN search."

WELL SAID! I can't beleive why some of my sites simply do not rank. I have put more time into them, got more links to them, they are older than most sites on the page and they are beaten by sites of a few pages that have: no content, no links, not been updated and are just there because of this DUMB algo.

I just do not get it :O/ and until I work out way I am not going to be a happy man.

From what I see all good seo works for some sites on msn and then they just chuck some losers in there to piss you off! <scratches head>

Goog and Yahoo are much better than this new found peice of junk!

Matt Probert




msg:1540298
 8:42 pm on Feb 1, 2005 (gmt 0)

With total disregard for being thought of as an AOL user, I shall agree and say I am not impressed by the MSN search facility.

Though I was amused to find I have a namesake at York University, doesn't look anything like me, but the name's the same <g>

Matt

flashfan




msg:1540299
 8:43 pm on Feb 1, 2005 (gmt 0)

We have some phrase with top 10 rankings in Google and Yahoo. However it even doesn't show up in the first 5 pages in MSN. Do you guys know how MSN indexes page?

houseofsecrets




msg:1540300
 8:47 pm on Feb 1, 2005 (gmt 0)

But here's one I just tried (and so should you) used cars eastern MA.

That search doesn't deliver appreciably better results for the Other Big Two. If you substitute a specific word (maybe a city) for 'eastern' you get much better results on all three. Heck, leave it out, and the results get even better.

Again, I will say in the keyword sectors I work with, MSN Search is mucho better than the other two. And it's really barely a contest.

BigUns




msg:1540301
 9:04 pm on Feb 1, 2005 (gmt 0)

houseofsecrets:

With respect to used cars eastern ma , none of the top 10 MSN results are about used cars in eastern MA with the exception of one forum on SUV repair. I actually looked at the sites. Mostly garbage or not relevant. (Try it)

dickbaker




msg:1540302
 9:08 pm on Feb 1, 2005 (gmt 0)

Interesting. If I do a search for "widget stores Illinois" or "widget stores Peoria Illinois" or other states, my site comes up either #1 or #2.

I wonder why the discrepancy between my site and others.

houseofsecrets




msg:1540303
 9:48 pm on Feb 1, 2005 (gmt 0)

With respect to used cars eastern ma , none of the top 10 MSN results are about used cars in eastern MA with the exception of one forum on SUV repair.

I don't disagree with you about that. I'm just saying that following the links for the same search on G and Y isn't really any better especially on the front page. Google pulls up some superpages listings but the rest of it is pretty similarly bad (i.e. it doesn't lead you to pages of used cars in eastern MA)

And it's a fine idea, IMO, to flog MSN for bad results but MSN is generally being held up as bad relative to Y and G and I don't usually find that to be the case. I can site specific searches where it absolutely blows them out of the water.

houseofsecrets




msg:1540304
 9:49 pm on Feb 1, 2005 (gmt 0)

dickbaker, are those the terms you optimize for?

adamas




msg:1540305
 10:30 pm on Feb 1, 2005 (gmt 0)

Some of my 'not in the top ten's are a little odd but they are blown out of the water in strangeness by some of the number one positions I do have!

I don't think I'll be switching.

HughMungus




msg:1540306
 11:53 pm on Feb 1, 2005 (gmt 0)

I think it depends on the search you're doing. I did used cars eastern Pennsylvania and I got car dealers. And since we can't keep posting search terms here I doubt we'll ever settle it.

BigUns




msg:1540307
 1:47 am on Feb 2, 2005 (gmt 0)

My original assertions were:

(1) Word Proximity means nothing to MSN
(2) Titles mean almost nothing to MSN

I have read nothing here and have seen nothing in the SERPS to indicate otherwise. The top 10 results for used cars eastern ma show NONE of the keywords in the Titles*(!) and little if any evidence of keyword proximity in the actual results snippets or pages other than that one would expect 'randomly' as contrasted with Google and Yahoo where Titles and keyword proximity are an integral and obvious part of the algorithm and the resulting SERPS (abeit unknown quantitatively to those of us on the outside). I have a suspicion that the MSN search was designed from the ground up not to rely on TITLES or keyword proximity, perhaps as a way to combat web page spam. I happen to believe the results for many searches are bad as a result of that design decision, if in fact it is part of the indexing algorithm. I am open to being convinced otherwise but I have seen nothing to indicate that MSN places much emphasis on those two factors. In a nutshell, MSN seems to pick out pages based on everthing else, keyword density etc, and if the keywords are in the Title fine, and if there is proximity fine, but these factors do not seem to be used in the algorithm itself - which if true is a remarkable departure from other engines.
*except for one incidental MA which did not stand for Massachusetts.

houseofsecrets




msg:1540308
 2:06 am on Feb 2, 2005 (gmt 0)

That's a very interesting observation. I can see how that would be a drawback in situations like a search for MA, where proximity identifies it so darn well, that you would want to focus on that. But, the bulk of search queries don't benefit overly from proximity, IMO. I would like to construct better reading and sounding titles than best practices for title construction for Google tend to encourage. Looser proximity requirements would tend to reduce the need for active SEO on natural text, also IMO, rather than as the product of research into this issue.

dickbaker




msg:1540309
 5:29 am on Feb 2, 2005 (gmt 0)

houseofsecrets, there are so many keywords and phrases I have to optimize for that getting backlinks to my interior pages will take a lifetime.

I have "widget stores" on my site in just about all of fifty states, soon to be all fifty. That means trying to optimize for "widget stores Peoria Illinois" as well as "widget stores Fairbanks Alaska."

Then there's the keywords and phrases that really bring the visitors in: the brand names and models. Probably 80% of my pages feature photos and descriptions of individual widgets from the various manufacturers.

This was my plan all along, as I'd seen it drive traffic to other sites I've created.

The problem--and the particular problem with Google--is that I'm now competing with other sites I've created for clients who deal in the same brands and models. If I do a search for "Acme widgets," both my site and my clients' sites are up on top on MSN. On Google, my site isn't even in the top twenty to top 250 for those same terms.

The problem becomes even worse when searching for "Acme model 150 Widget." My older clients' sites rank well for this on all SE's. My site does not do well, even on MSN.

MSN seems to have put some sort of priority on first and second level pages.

Here's another oddity. My page titles by necessity are along the lines of "Illinois widget shops--Illinois widget stores." The individual brand pages are titled "Acme hammers--Acme mallots." I hope you get the idea.

Anyway, if I do a search for "Illinois widget shops", I'm #1. If I do a search for "Acme hammers," I'm generally #1 to #5. But if I do a search for "Illinois widget stores," I'm nowhere. If I do a search for "Acme mallots," I'm in the nowhere zone, too.

By contrast, a widget site I created over two years ago ranks at the top for just about any variations of "hammers" or "mallots" or "stores" or "shops."

The new MSN has brought my new site to the front page for most of the search terms I want. Google is still slowly moving those pages up in the SERP's, after eight months from submission time.

What I don't understand is why MSN Search hasn't been able to drill down from the second-level pages to the third-level that link off of those pages.

And, as long as I'm on a minor rant, has Bill Gates looked at the new MSN homepage on a Mac? I've seen "Little Bobby's Britney Spears" sites that had better cross-browser/platform design.

zeus




msg:1540310
 1:09 pm on Feb 2, 2005 (gmt 0)

I just did a search for bla bla eagles pi... at no.4 was google.com - what does that tell you about the results.

patoruzu




msg:1540311
 5:13 pm on Feb 2, 2005 (gmt 0)

I’m a bit suspicious. I think that the beta version worked better than this “new MSN”. Maybe all of this is part of the strategy: to relax the enemies and put them off guard.

houseofsecrets




msg:1540312
 5:16 pm on Feb 2, 2005 (gmt 0)


I just did a search for bla bla eagles pi... at no.4 was google.com - what does that tell you about the results.

um, nothing? Like most of the "this search is bad" posts, all it reveals is that SEO's don't know much about constructing search experiments that are meaningful. You people that do this should observe the TOS and confine your specious analysis to red widgets.

The person who started this thread presents a specific example search that shows convincingly to me that all the search engines do a sucky job of sending results for that term. But, of course he or she does it in a thread entitled "Msn Search is really bad" when the thread should be titled "Why don't any of the major SE's understand what I mean when I type 'Eastern MA'?

The answer to that is simple: because eastern MA doesn't have an objective description and there are no canonical results for queries that include the terms.
If you put a city or just eliminate the vague geographical locator that throws the SE's off--you get superb results from MSN. If you throw garbage in however, I think most of us know what's going to come back as results.

ddogg




msg:1540313
 6:52 pm on Feb 2, 2005 (gmt 0)

The title tag is the most important thing to MSN search from what I can see based on my own sites' rankings.

BigUns




msg:1540314
 8:10 pm on Feb 2, 2005 (gmt 0)

UPDATE:

Today, for the first time, I am seeing clear evidence that MSN is now using word proximity. I just finished some tests at about 2:30 PM CST Feb 2 on a wide variety of searches. These results are clearly different than those I've done periodically for about a month. There are clear indications that position on SERPs depend on word proximity. What I don't know is if this is temporary, whether it impacts large segments of the web, etc. For now, for searches I am doing, the SERPs are clearly better.

I have no good evidence (yet) that search words in Titles have any impact on SERPs. That's a tougher one to isolate.

Another bit of good news for MSN search: For the first time today I am able to enter a certain 4 word search and essentially isolate some Microsoft technical documentation that I often seek out. Prior to today, a Google or Yahoo search would put what I wanted at the top - typically #1. Now, MSN search does the same, and rightly so. A month or even two weeks ago, I literally had to use Google or Yahoo to find Microsoft documentation - now MSN search finds it too - and that's a big step forward.

The name of this thread is now changed to:

"MSN Searchis bad, but definitely getting better..."

BigUns




msg:1540315
 8:19 pm on Feb 2, 2005 (gmt 0)

UPDATE:

Today, for the first time, I am seeing clear evidence that MSN is now using word proximity. I just finished some tests at about 1:30 PM CST Feb 2 on a wide variety of searches. These results are clearly different than those I've done periodically for about a month. There are clear indications that position on SERPs depend on word proximity. What I don't know is if this is temporary, whether it impacts large segments of the web, etc. For now, for searches I am doing, the SERPs are clearly 'better'.

I have no good evidence (yet) that search words in Titles have any impact on SERPs. That's a tougher one to isolate.

Another bit of good news for MSN search: For the first time today I am able to enter a certain 4 word search and essentially isolate some Microsoft technical documentation that I often seek out. Prior to today, a Google or Yahoo search would put what I wanted at the top - typically #1. Now, MSN search does the same, and rightly so. Previously, I literally had to use Google or Yahoo to find Microsoft documentation - now MSN search finds it too - and that's a big step forward.

The name of this thread is now changed to:

"MSN Search is bad, but definitely getting better..."

HenryUK




msg:1540316
 11:16 pm on Feb 2, 2005 (gmt 0)

Hi, without pushing a strong opinion as to quality overall I'd like to share some experiences that I hope may be of some use to other users. My site is doing both very well and very badly on MSN search and I thought it might be of interest. If this is not the best thread for this, someone point me to it...

I run a site in the UK that is a market leader backed by a large publisher. It has >30,000 different pages all with unique (user-generated) content in a fairly specialist sector. The site ranks #2 in Google for its main 2-word keyphrase. It ranks #1 in MSN's UK search for that phrase, out of c. 59m reported results, and in the top 10 on MSN.com.

The home page mentions the key phrase two or three times in the text (within natural language) and it's obviously in the title and description tags. The main optimisation *for this phrase* is that it is linked to on this phrase by a large number of pages, some with PR as high as 8. The home page itself has PR7.

Most of the site's referrals however come from unique phrases (c 50,000 different phrases per month) - along the lines of [product type] [UK town], which for our specialised "products" we usually rank #1. These pages are accessible via a site map placed there to allow spiders in - although users are signposted to use a search form instead of browsing them (for the sake of usability). The pages are db-driven, but the content is genuine. It's all white-hat SEO, putting key phrases from the data into the titles, meta tags etc.

These pages, which do incredibly well for relevant searches on Google, and which (I would say) represent accurate results for searchers, are nowhere at all on MSN search.

The first problem that I have is one of being indexed at all - about 2% of pages so far are indexed. This I hope is just a question of patience.

More puzzling is the fact that most of the indexed pages are not coming up in the SERPS except on a site search. EG a search for [obscure-ish product] [small UK town] does not return an indexed page with the exact title of that search which also has those keywords repeated (but not "stuffed", a couple of times only) in the text of the page. And I don't mean it's not #1, I mean it's not there at all.

The pages would typically have PR2,3,4. The only theories that I have are (a) pages are newly indexed and appearing on site search but not on main search yet (seems a bit hopeful) or (b) that it is a problem with how FAR they are from the home page - having so much data and sticking to the 50 links per page rule, these pages are typically 4-5 clicks from the home page.

MSN talk about pages ideally being 1-3 clicks from the home page in their help files - has anyone else had experience of deeply nested pages being rated lower by MSN than by Google. I'm seeing less relevant intermediate pages indexed way above the much more relevant pages they link to, which also would tend to support the hypothesis of "anti-depth" bias.

Incidentally, not whinging, just sharing and hoping for enlightenment!

thanks, apologies for length.

H

lgn1




msg:1540317
 2:11 am on Feb 3, 2005 (gmt 0)

I get good results if i search for widgets {mtch=0}
(site builder variable, for exact keyphrase), on a two keyword phrase

If I use {mtch=50} which appears to be the default,
I get a lot of garbage.

BigUns




msg:1540318
 5:36 am on Feb 3, 2005 (gmt 0)

HenryUK:

I wish I could help or offer advise but I can't; I simply don't know enough about your particular situation or about MSN Search. One possibility, however remote: have you done a site:www.MySite.com to see if MSN is indexing multiple root pages, eg www.Mysite.com and www.MySite.com/default where default is whatever your default page is?

Ign1:

I get inconsistent results using that slider. I am not convinced it is labeled properly, nor am I convinced that the % "match" shown is what MSN is using, nor do I really know what "match" means quantitatively, and most of all, those sliders are a confusion factor for what I am trying to see in the SERPS; so I typically leave them in their default positions which is where 99.9% of people will leave them anyway. ( I can't even be sure that the SERPs I see without moving the sliders are what other people see - for all I know MSN is using reverse ISP to dish out different results or has multiple data sets or ... But thanks - I do think those sliders have a lot of useful potential but that's an issue for me for another day.)

houseofsecrets:

It is abundantly clear you don't like the title of this thread. You've made your point. You clearly don't understand and/or don't care about what I am trying to analyze - word proximity and keywords in titles. As indicated in my very first post, I can't analyze the SERPS if I use city state search words like Boston MA or Peoria Illinois. It obscures the issues I am investigating. If I use Boston and MA as search words, even a dumb as a rock search engine that knows nothing about word proximity will have lots of SERPs with Boston MA in close proximity - even if it took pages at random - simply by virtue of the way humans write web pages. It is abundantly clear from postings here and in other threads that you don't give a D about word proximity - well I do and so does every modern search engine I am aware of.

*********************************************
To the individual who sticky/emailed me seeking $ to get me good MSN rankings: You are a fool who has made unwarranted assumptions. Do not confuse my "new" status in this forum with helpless amatuer. Be advised - I already have excellent rankings in Google and Yahoo and pretty darn good rankings in MSN, despite what I believe to be MSN Search shortcomings. I need to pay you as much as I need BiggerUns. If you solicit me again you will be reported to the forum 'authorities'.

HenryUK




msg:1540319
 10:28 am on Feb 3, 2005 (gmt 0)

thanks BigUns but I hope my post didn't sound too much like a call for help, that wasn't the intention.

There's no issue with multiple indexing of the home/default page, and this is where the site is doing really well.

My point in posting was really to speculate about the way MSN search treats deeply nested or "buried" pages, and to seek other users' experiences of this.

H

BigUns




msg:1540320
 4:39 pm on Feb 3, 2005 (gmt 0)

HenryUK:

Here's another remote possibility: I was recently informed of a site where all the pages have been indexed by the msn bot in the sense that a site:www.MySite.com shows the pages. However a subset of these pages never show up in the SERPs. One thing those pages have in common is they all have the same META description tag - sort of a typical bland generic tag. Although purists will demand that all pages have there own unique description tag, it is very common for people to use the same one on multiple pages. Is this your situation by chance?
I have no idea if MSN cares about this kind of thing; Google and Yahoo don't (as far as I know); but who knows. I would immediately dismiss this, but the MSN search is a new beast with some very strange characteristics; even the MSN page that talks about this META tag is odd. They actually give an example that uses both the HTTP-EQUIV and NAME attributes; something I don't recall ever seeing. Indeed it is interesting that MSN devotes a good part of a page called "Site Owner Help" to a tag that some people don't even use.
I wonder if other people out there have noticed any peculiar things with MSN and META description.

[beta.search.msn.com...]

houseofsecrets




msg:1540321
 6:27 pm on Feb 3, 2005 (gmt 0)

BigUns, please don't misunderstand me. The simple fact is the query you gave as an example does not deliver appreciably better results in Y or G. I think that obscures the point of your discussion for those of us trying to follow it actively as opposed to just reading it and saying "yep, MSN search is bad..." I rather appreciate the rest of your post...

My experiments with keywords and title, show me that yes, proximity in title is not so important as other factors. It is noticeable on 3 word queries that MSN seems to be using a pretty large lexicon of synonyms (like a ~keyword search in Google) and ranking those relatively higher than G does on the same search

If I do something like
handmade red widgets

I see pages ranking highly with
homemade red widgets

or crimson handmade widgets or handmade red wozzles (where wozzle is a widget synonym) in the title

and I notice when you shuffle kw1, kw2, kw3 around in my kw range, that MSN delivers much more consistent results for the 6 possible queries, than G does. This is a very welcome change, at least to me.

I fail to understand why proximity in keywords is such an appealing feature. My instincts tell me that any example of a query where proximity helps you get the right answer can be countered by another well-formed query where proximity factors would decrease the quality of the results. When you look at the big shuffle that G does when you toy with kw proximity for search queries that, no matter the word order, are looking for the same thing, it seems to me to be undesirable.

This 32 message thread spans 2 pages: 32 ( [1] 2 > >
Global Options:
 top home search open messages active posts  
 

Home / Forums Index / Microsoft / Bing Search Engine News
rss feed

All trademarks and copyrights held by respective owners. Member comments are owned by the poster.
Home ¦ Free Tools ¦ Terms of Service ¦ Privacy Policy ¦ Report Problem ¦ About ¦ Library ¦ Newsletter
WebmasterWorld is a Developer Shed Community owned by Jim Boykin.
© Webmaster World 1996-2014 all rights reserved