MSN Ranking Factors List

I've been scrutinizing MSN beta to determine why irrelevent PR1 pages are so easily beating much much more relevent PR5 pages (with great inbound anchor text). I am not impressed with their results yet, especially for more complex searches.

Nevertheless, I've distilled some key recurring ranking factors listed in order of importance (I didn't include cloaking which is extremely powerful):

1) Title (obviously) and keywords in url- this hurts relevence, IMO.

2) Keywords in bold. This alone can make the irrelevent PR1 pages beat more relevent PR5 pages. I can't stress enough the importance of bold font in MSN Beta.

3) Important keywords in center table layout. This relates to MSN's algorithm of Block Level Analysis. Basically keywords (and links) in the center block are weighed more than keywords on the side or bottom. Read about it here: www.technologyreview.com/articles/04/10/rnb_100604.asp

4) H1, H2 and H3 headers. These have a fairly large weight.

5) Keywords in italics (even better when bold as well)

6) Links. I've found links to have surprisingly little weight in the MSN beta. Much of the algo is on page.

Of course these factors are 1999 SEO 101, but they are sooo effective in MSN to the detriment of relevence.

What isn't important in MSN Beta:

1) Being listed in DMOZ (I know because it hasn't done cr*p for me)

2) Being listed in the Yahoo directory (I know because it hasn't done cr*p for me)

3) Inbound anchor text is weighted very little.

4) Meta tags don't seem to count (good thing)

***Caveat***

Make SURE you do not have H1 or H2 headers and bold them at the same time. I did this, and my page fell 10 serp pages. I removed this and I recovered.

This is some sort of overoptimization filter.

So, give me some feedback as I am trying to crack this algo some more!!!!

i've found the opposite in regards to inbound anchor text. the key though in the linking is internal linking. the use of doorways is very successful. cloaking as you've pointed out is also very successful.

i've found very spammy sites be very successful by taking a list of keywords, making the doorway pages with keyword-phrase.html, keyword.html, keyword-phrase-2.html..etc. having menu's with links to all the doorway pages with the keyword phrases as the anchor tags. i've also found a nifty old school trick involving frames that works quite well.

I would put Keywords in URL at 4.5. Any weight given to URL can be easily overcome with onsight optimization.

3.5) On-site anchor text. A trailer link on a large site with good anchor will pop a homepage to top ten from oblivion.
5.5) Alt image tags also carry good weight.

To much alt text can hurt your serps, and keywords in h1, h2, in combination with the same keyword in bold
can get you banned for over optimalisation.

My site has a bold H1, at the very top of the site. I am at #3 or #4 in the existing MSN. In beta I got shoved down to 11 to 12 and stayed there for a while. Some 2 weeks ago started moving up a position every 5 to 7 days or so and am 6th now... The bold H1 is still there. I have been adding content at the rate of about 200 to 300 words a week.
raz

Thanks for your analysis, this seems to be the first really decent job at uncovering the msn mystery. i agree with most, but i havent yet seen the effects of bold. h tags seem to make a bit of a difference on my pages.

It seems to produce basic, logical results for some queries, but for some keywords it really seems to be producing very unrelevant results. It almost seems as if they are relying too much on AI to figure out which site is the best.

Now, I have to take everything I said back! This is exactly what happened 6 weeks or so ago. As soon as I posted here my SERP in MSNBetta went down... Last night as I was posting I was at 6, This morning at 15th.... ouch...
raz

thats pretty interesting, i guess he could have been doing just that, but ill give him the benefit of the doubt.

GChaney,

You can remove your tin foil hat- this is not a conspiracy!

Forgive me for trying to share my observed results gained from spending my Saturday analysing MSN beta serps. I'm just hoping to get a discussion going. Oh and aren't we all interested in learning how the MSN algo works?

Next time I just won't post, just so that no one is offended and thinks I have ulterior motives. I'll just keep my observations to myself.

Good job on the article. I think you have a very valid observation in the state of MSN. I have too also agree that keywords in the URL is around #4 in importance.

Well done gchaney.
I agree with your analysis.

BTW, your first post did appear totally out of character.

I agree with most of what you have said, I am testing out the Italics now..
just to see..

Thanks for the info

i am also testing out the bold and italics and havent yet seen improvements, but will post with updates as soon as i see any large movement