Why Microsoft Wants to Buy Google

xihr writes "Harry Fletcher writes in The Inquirer about an obvious discrepancy between searches for "linux windows" on Google and MSN; the former comes up with almost 9 million hits, but the latter only comes up with -- wait for it -- 16. The author then speculates on Microsoft's ulterior motives for their attempted (and failed) purchase of Google."

Re:Not so fast

[BabyDave]

The author of the article has no clue how MSN search actually works. MSN search returns only sponsored results first. There are 16 sponsored results and then if you go to the next page you see:

True, but it doesn't actually say that anywhere - it just says "Results 1-15 of about 16 containing linux windows". Also, there is a 'sponsored links' section on the right of the page, separate to these 15. You need to click on the 'next' button to see "Results 16-30 of about 8898833". Which is an abysmal design decision, if nothing else.

Re:Not so fast

[Ubergrendle]

No where, *anywhere* on that first search results page does it indicate these are "sponsored results". Maybe "featured sites" is a euphemism for "paid for adverstising", but you have to ask yourself...why does Microsoft not come clean? Oh yeah, right, the credibility factor...

Having run a search engine for a major banking internet portal, i know from experience that less than 50% of your users are going to move to a second page of results...more often than not, if they don't find what they're looking for in the first 20 results, users are just going to abort, or chose the closest looking link.

And...

[Spackler]

...And, it returns a pop-under for your trouble. So, I get sponsered links, and then a pop under, and then another click, and I can get the results I was looking for.

Wow, the mind boggles.

Google company meeting

[Tancred]

I hear that during a Google company meeting a couple weeks ago, the crowd was asked if they'd heard the rumor about Microsoft buying Google. Then they were asked to raise their hands if they believed the rumor. Only one person raised his hand - founder Larry Page. I like his sense of humor. ;)

Re:General Public-Active Desktop.

[Verteiron]

Actually, most people asked a tech to disable it for them the first time the computer went into the shop for anything. I can't tell you how many times I've turned that thing off for people... along with changing their homepage (my internet is provided by Yahoo!), removing some bookmarks (but I've never BEEN to hotteenlesbians.com!), and "installing Google.com" (I can't use Google, I told you my internet is from Yahoo!).

I guarantee you these people and others like them have racked up plenty of hits for MSN's advertisers because they don't even know they're "allowed" to use another search engine!

Re:Not so fast

[gnu-generation-one]

"I got 9.3M in 0.11 secs at Google, did someone add 1.2M links this morning? ;)"

It depends on the country you're viewing from [harvard.edu]

The google results can vary depending on per-country censorship rules (French and German laws are the ones discussed here, but other examples include China [clearharmony.net]and the US [google.com] where the results may be less than expected)

WRONG! MSG has plenty of hits

[fleener]

Click the "Next" button a few times on that "Linux Windows" search.

You'll see:
Results 31-45 of about 8897853

Re:Not so fast

[Pivot]

Well, but if you search for windows, it displays the correct hit count on both pages. For linux, it hides the hit count on the first page. Try it.

Re:General Public

[mcpkaaos]

They wont understand either...

Sure *we* can see past this and get facts, but *we* dont line their pockets either...

Did you ever stop to think that one of the reasons people don't support or take the time to learn about the various OSS movements is because of high-brow, elitist comments such as yours?

Instead of alienating the "average consumer", perhaps we should be working as hard as possible to present a viable alternative without all the attitude. If we perpetuate an image that Linux users are different from the "average consumer", then guess what, we will be the only ones using it.

Linux, as many OSS projects, is not too dissimilar from a business in many ways, except that instead of a highly trained marketing department, it has us. That is a greatly simplified statement, but I think it stands to reason.

Makes me wonder...

[c4ffeine]

I'm pretty sure that I've heard of a few virii or worms going around that, among other things, blocked google. What an amazing coincidence... Anyways, here goes M$ with more anticompetitive behavior, maybe if we all get really pissed and make a big deal of it they'll get fined $500k. Nothing we can do about it. Oh well, shit happens, at least they're losing customers to apache. That's a start, all we need is to get linux more popular as an os for all those aol users out there

Re:WRONG! MSG has plenty of hits

[saddino]

Yes, but the question is: why would you have to hit Next a few times to see the actual number of results?

Bah Humbug!

[jefu]

A search for "Charles Dickens" shows us that it finds :
Results 1-15 of about 50 containing ""charles dickens""
Going to the next page gives 16-30 and so on.

It is only the patient searcher who will discover that after all 45 of the 50 hits are seen that there are 151383 results left to go.

So I agree, its not that MSN Search is biased against Linux. It is that MSN Search is biased toward those willing to spend money to be listed. And (as a direct effect of that) MSN Search is biased against those trying to find information.

They're not being dishonest for ideological reasons - they're just not interested in honesty or usefulness so much as they're interested in selling themselves. (I'll resist direct comparisons with the kind of prostitutes who are honest about what they're doing.) (OK, so my resistance crumbled.)

And of course, they'll never go broke underestimating the american public.

controlling access to information

[jasonbowen]

I'd wager a bet that most of us replying to this thread use the internet for research purposes daily. The fact that I just searched for linux on http://search.msn.com and came up with roughly 350+ hits shows that the owner of that search engine wants to limit my access to content on that topic. I think it's time for legislation involving search engines providing information on how they provide their results. Microsoft is obviously either blocking results that pertain to Linux or they need to state exactly how their search engine works if this is not the case.

Linus.

[Jacek Poplawski]

MSN results (yes, I read about sponsored links):

office - 5204
windows - 2373
server - 2121
money - 982
terrorism - 249
osama - 103
bill gates - 63
linus torvalds - 323738

Re:MSN search doesn't return huge numbers of hits

[kalidasa]

Try George W. Bush. I just did. It'll say something on the order of 301 hits. Scroll through the pages to the last hit. Suddenly, the number jumps through the roof. I find it hard to believe that the first 300 hits are all sponsored links. I think something else is going on here: MSN has not only sponsored links, but some kind of edited directory scheme going here, and it doesn't care to let you know that the first number it quoted is of those links which are sponsored or added editorial, and the second number is a raw search result.

Re:Microsoft Biased? Never!

[gamma male]

search for:

linux switch

and you get more than a hundred thousand. And some in the first 10 are about switching to linux. I suspect that some "searches" on MSN don't actually return a search, but rather a pre-selected set of results.

Re:Why corporations must be stopped.

[sacrilicious]

I'm a big fan of holding people responsible for their own stupidity.

I am too; stylistically I wouldn't choose to state it as above, but I am a fan of accountability for individuals and corporations alike. If your position were a machievellian one that rationalizes doing something because it's possible to do then I wouldn't agree, but if I understand correctly then perhaps you and I simply have a different threshhold of what we'd consider "stupid" in this particular case. It seems your position that when people don't have an awareness of how search results may be skewed in favor of a given business, then they deserve to be taken advantage of and critics should be silent.

I think there's a component of untruthful advertising to this issue. Microsoft may not actually commit themselves to objectivity by saying "our search results are unbiased with regard to any commercial interests" but it's not an unreasonable expectation of a search engine; after all, google has set the precedent of making money via advertisements, not by covert manipulation of results.

If I drive past a gas station and ask to have my engine checked, the station may well try to bilk me out of money by claiming I need more work done than is the case. There are ways for the station to claim this without technically breaking the law, but I don't regard myself as stupid for not knowing enough about cars to certify the station's results; at least not stupid in a perjorative way. And, I wouldn't regard the station as on uncriticizable moral ground for doing so. This may all come down to my world vision, where I'd like to steer the world towards honest exchanges without people stabbing each other in the back at every opportunity. I realize that not everybody wants to live in such a world.

There's No Such Thing As 9 Million Hits

[Effugas]

The maximum number of hits a particular search can sustain is between 5 and 100.

This was a great surprise to the engineers building search engines -- the original problem was, how do we find particular keywords across hundreds of millions of documents? This was solved relatively trivially -- index by keyword, and distribute the search space across the memory of many, many machines. All of the sudden, it became apparent that search had much less to do with how much you could search and much more to do with which results came up first.

That's a much harder problem -- fundamentally, without the user telling you what he wants, how can you figure out what he's most likely to desire? This actually uses artificial intelligence techniques, much to the consternation of the eternally discredited AI folk who point out that "the moment AI becomes useful, it ceases to be called AI" (which is true). It's AI because you effectively need to programmatically derive what an intelligent surfer is most interested in, as an abstract subject instead of a concrete phrase. Google gets alot of credit for their Pagerank algorithm, which uses links from other sites to weigh which links are more "authoritative" than others, but interestingly enough their system is noticably robust even without outside links. Corporate websites all tend to run 1998-era search engines -- all quantity, no quality (and in this case, quantity has no quality all its own). Some time ago, I worked at a massive company that was testing Google for internal searches. Corporate web pages are far less cross-linked than the web itself. But Google-internal worked just like...well, Google :) So there's some really smart AI there.

Anyway, as I've said before [slashdot.org], MS can't buy Google; they'd just create the market segment of "what Google used to be". Speaking as someone who has a healthy respect for MS as a company, they've simply burned through too much goodwill for people to trust their search results as authoritative. Yes, mysql only gets a few hits on MSN search. So does pancakes, and I doubt MS is part of the great Waffle Iron Conspiracy.

Yours Truly,

Dan Kaminsky
DoxPara Research
http://www.doxpara.com

Maybe you are giving MS too much credit...

[Maltese Duck]

it is possible that Microsoft wants to buy Google not because it wants to reduce Linux hits, rather the MSN search engine just isn't very good.

Here are the results for different OSes:

OS: Google vs. MSN

aix: 3,550,000 vs. 40

solaris: 8,340,000 vs. 87

tru64: 676,000 vs. 93,433

hp-ux: 1,890,000 vs. 287,294

ms windows: 6,650,000 vs. 291

vms: 1,770,000 vs. 22

os2: 702,000 vs. 118,273

linux: 96,800,000 vs. 365

freebsd: 8,810,000 vs. 1,136,552

openbsd: 3,500,000 vs. 341,343

netbsd: 4,750,000 vs. 223,075

unix: 23,700,000 vs. 164

symbian: 1,400,000 vs. 420,701

"windows 2000": 6,750,000 vs. 315

"windows xp": 970,000 vs. 297

"windows server 2003": 4,170,000 vs. 30

Re:Microsoft Biased? Never!

[fishbot]

Yep, page 2 of the search for 'linux windows' should theoretically bring up a single hit (the remaining 1, out of 16). However, it brings up 15 more pages, out of about 8879023!

Looks like if there is biasing going on, it's just to try and prevent people hitting 'next'. Don't forget MSN don't provide a nice goooooooooooogle style page index, just next/prev links. Ick.

No microsoft bugs! Gates explained!

[Anonymous Coward]

about 2202403 containing "linux features"
about 829257 containing "linux bugs"
about 3119922 containing "microsoft features"
about 20 containing "microsoft bugs"

No *wonder* Gates thinks M$ software is bug free!
He uses MSN!!!

Linux sucks?

[madaxe42]

Here's an interesting one... try searching for 'Linux' on msn. 300 odd hits, right? Now try searching for 'Linux sucks'... Whoa....!

The Enquirer missed the real story...

[kale77in]

The Enquirer completely missed the real story. Try searching on more specific terms:

google.com.au:

'Microsoft Windows' ~= 7,560,000
'Debian Linux' ~= 3,900,000

search.msn.com:

'Debian Linux' ~= 793,659
'Microsoft Windows' ~= 713 (!)

This demonstrates the opposite of the kind of bias The Enquirer was alleging...

Is Microsoft limiting Windows search results to only 'approved pages'? How can there be only 700 pages dealing with Microsoft Windows? Unless they're truly *ridiculously* incompetent, which I don't buy. Or unless they are losing some serious market share to Debian!

Now there's a happy thought. :)