Fog Creek Software
Discussion Board




Why Google will not win in the search war

http://www.google.com/press/zeitgeist.html

Notice the Top 10 Gaining Queries( Week Ending May 17, 2004) and Top 10 Declining Queries (Week Ending May 17, 2004)

Around 30% of them can be called "intelligent". Questions about Iraqi prisoner abuse (lynddie england),  the American who was unfortunately killed (Berg)

Majority of them are just moronic. Like "Orlando Bloom", "Alexandra Kerry (who showed her tits), "Troy" ( a movie).

This is more or less the pattern in the zeitgist for any month.

Inference? The people who use google - or the **people who matter** are the ones who dont look for real relevance. If they type Alexandra Kerry, they just want to see Kerry's tits. and not Sharon Stones.

The barrier of entry to satisfy such people is very very less.
It proves beyond doubt that people are not looking for brilliance or even **relevance** here.  Google will tell you which Alexandra Kerry tit site is the **most** popular. But the jokers who are searching are not looking for brilliance.

Till now, all other search engines sucked abysmally. Even if they improve a bit and provide some relevant results, the days of google being undisputed No 1 is over. Even with their search results sucking, Yahoo is just some percentage points away from google.

This is not to say Google would lose the search war. Hardly that. But it cannot win decisively either and become a monopoly.

K
Sunday, May 23, 2004

I think the biggest way Google can win is to create two version of the site.

Google, and google moderated.

One would be militant about filtering out crap, one would be as it is now.

I think there are people in Google who probably feels the same way but aren't really convinced it's possible for such a search engine to exist (where will all the moderation monkeys come from) and still be as cool as the current google.

If Google can seriously fix 90% of the problems they are having now, and still keep the secret to itself, other search engines will have to work much harder to proof their worth.

Li-fan Chen
Sunday, May 23, 2004

Best way to go about it is probably in fact using things like Gmail and snooping for PGDs... or Paul Graham Dollars... or words that are relevant to your life. If you use this and map like minded people together, you (as google) will be able to tell relevent clicks from irrevent ones.


I click on 5 links.. 3 are deceiving...
Someone clicks on 5 links.. 3 are deceiving...
Someone+1 ... same thing
Someone+2... same thing..

Statistically the one that gets clicked on the most are the most relevant.. the ones everyone is ignoring is crap.

You keep at this for a while and the order of things will be just beautiful.

Li-fan Chen
Sunday, May 23, 2004

Okay so you need a spy like Toolbar that listens to clicks too, but then a url redirector will do that for you too (see google cached pages).

Li-fan Chen
Sunday, May 23, 2004

Hell, just add a survey to every google cached page at the top.

[Click here if this is page is irrelevent] [Click here if this is really relevant]

Again, you need PDGs using something like GMail to put people into groups. Unfortunately to link the two together people will have to learn to give a little. They need to log into Google Passport so that GMail PDGs talks to Toolbar PDGs.

Li-fan Chen
Sunday, May 23, 2004

Once you have this and thousands of CS BSes will be writing a paper about this.

Li-fan Chen
Sunday, May 23, 2004

K >

I think you are missreading the data. You are assuming that people only ever search for one type of things.

I would think its more like, everyone has special interestes, but those are very diverse.  But the car people, the computer people, the sports people and whatever, all go googling for brittny spears every once in a while.

If they has statistics for "Frivolus" vs. "Serious" searches, I bet the serious once would win.

Eric Debois
Sunday, May 23, 2004

What the hell are you some of you talking about? Google has this thing called "Google Safe Search". You can choose to filter results with strong or moderate filtering. There's no need for "two separate Google sites". Now, I don't think that Google will become a "monopoly"*, but it has strong advantages over every other search engine. The closest one compared to Google is AllTheWeb.

Google has some major advantages over other search engines. The thing that I appreciate the most is that it is a spartan website (so is AltaVista and AllTheWeb), without lots of crap all over the place. It's just about the business of providing me with search-results. The text-ads that come up are even useful to me at times (especially when I search for something like "Health Savings Account"). Google appears to have got the message: Users don't like pictures (unless they ask for them, see porn). When searching for something, I just want plain-text and very fast results.

One should also note that there is a psychological advantage to speed. In terms of significance, would it really matter if Google took 2 or 3 seconds to produce results, instead of it's current instantaneous results? No. But there is something deeply satisfying about getting a result in a time-frame that appears almost instantaneous.

* People talking about Google becoming a monopoly have been fooled by Statist definitions of monopoly. On the free market, with no State-instituted barriers to entry or State-priviledges, there can be no monopolies or monopoly-prices, even if one company has 100% market-share, because if that company earns enormous profits, it encourages other entrepreneurs to enter the field, until the profit-rate approaches the natural rate of interest. The reality is that all real monopolies (e.g., patents) come because of State-granted priviledges.

David Heinrich
Sunday, May 23, 2004

Google wasn't the first search engine, you know -- there were others before it that Google sent to the grave.

Having said this, and just repeating what Eric said, the "most popular" are intersection sets between millions of people, but each of those millions of people also look for highly specialized things as well: I don't think many people will proclaim "Well I'm just looking up a boy band, so I'll use X, but when I want to look for something serious I'll use Y". You find one that satisfies both.

Dennis Forbes
Sunday, May 23, 2004

If there are one billion people making searches (which is a not unreasonable estimate), why on earth would the top ten, or least ten be in any way relevant?

continued in statistical/odious_idiocies_I_have_known

Simon Lucy
Sunday, May 23, 2004

The days of "winner take all" are over.  And that applies to Google as well.  The difference is that Google seeks (from what I can tell) to merely do their job well.  They employ just under 1000 people, and from what I've seen are quite profitable at that.  They seem to be rolling out new services aggressively. They will succeed.

Microsoft on the other hand does not consider a project a success unless its a Windows/Office grand-slam homer.  By this standard, Microsoft will perpetually fail.  The days of winner take all are over.  Success belongs to companies that can execute efficiently at many things.  Ones that can adapt quickly through true added value.

hoser
Sunday, May 23, 2004

I expect Google will be the search site of choice as long as it treats it's users with respect.

Many web sites seem like they're trying to annoy or deceive people with unmarked advertising, spam, intrusive or popup ads, etc. Google is one of the few who understand that people come back because the site is nice to visit.

I'm not saying the other heavyweights are bad, just that they'll have a tough time beating the leader unless the leader drives it's own customers away.

Tom H
Sunday, May 23, 2004

Hmm. Not sure if I buy that. Just because the majority of a persons searches may be moronic, doesn't mean they won't expect good relevant results in say the 20% of searches in which they're looking for something more obscure/detailed.

If the world split neatly into 'moronic search engine users' and 'non-moronic search engine users' then it might work better, but everyone does some amount of non-moronic searching, just some more than others.

Matt
Sunday, May 23, 2004

"Google, and google moderated."

You mean Google and Yahoo?

Philo

Philo
Sunday, May 23, 2004

Lol, how true Philo :D

Li-fan Chen
Sunday, May 23, 2004

>> all go googling for brittny spears every once in a while.

Today I googled through MSDN for half an hour.

And then googled for "Kelly Rowland wallpapers".

(Good thing Google returns good results for both.)

Alex
Monday, May 24, 2004

"If they type Alexandra Kerry, they just want to see Kerry's tits. and not Sharon Stones."


Ummm.... I'd settle for seeing the tits of either of them

Horny Teenager
Monday, May 24, 2004

Tom H says it well.  There is a leftover notion from the dotcom, turn of the century days, that the goal is to "make a portal" that will keep people on your site.  Thus the Brittany Spears smiling pics and pop-ups of J. Jackson's anatomy.  All busy and irrelevant.

Google gets it right.  The objective is to find and leave.

hoser
Monday, May 24, 2004

" * People talking about Google becoming a monopoly have been fooled by Statist definitions of monopoly. On the free market, with no State-instituted barriers to entry or State-priviledges, there can be no monopolies or monopoly-prices, even if one company has 100% market-share, because if that company earns enormous profits, it encourages other entrepreneurs to enter the field, until the profit-rate approaches the natural rate of interest. The reality is that all real monopolies (e.g., patents) come because of State-granted priviledges. "

That's assuming two big things which shouldn't be taken for granted.  1) The barrier to entry for the industry is low (which is true for search engines, but not for many commercial products) and 2) The companies play fair.  Except the whole point of monopoly restrictions is that monopolies DON'T play fair.  They can force their suppliers to not supply competitors.  They can force retailers to not carry competitors.  They can make their product incompatible with competitors.  All of these raise the barrier to entry incredibly high.  Sure other people might try, but it is highly likely that they will fail.  The presense of failing competitors doesn't mean there is no monopoly.

MikeMcNertney
Monday, May 24, 2004

Also, the "free market system" is not totally unregulated.  And, once you introduce regulations, those regulations inevitably have bias.

For example, an S Corporation limits its number of shareholders to 75.  This allows companies with less than 75 shareholders to have an option that companies with more than 75 shareholders don't.  Whether or not it is fair is debatable but there is definitely a bias.

Since there is bias, the structure of regulations influences how competition works and whether and how permanent monopolies can be.  It's a range.

But monopolies exist everywhere.  Every company has a monopoly in its own product.  Microsoft has a monopoly in Microsoft Windows.  The debate is whether that Microsoft Windows monopoly is effectively the same as a monopoly in the much broader category of operating systems.  But that's "effectively" a monopoly, not an absolute monopoly.  Clearly, other operating system exist.  So, it's a range, too, how "effective" is effective enough?  It's a matter of opinion.

Daniel Howard
Monday, May 24, 2004

On topic response: Philo's riposte ("you mean Google and Yahoo?") is the succinct answer to why search engines that censor and mickey with the input set by hand in order to satisfy what human reviewers *think* should be the most relevant result are doomed to fail.

In short, Google's algorithim is set up so that web content creators democratically vote which sites are the most popular and relevant.  Such a system will always be more correct than the opinions of a handful of reviewers.

Off topic response: the claim that truly "free market economies" (should such a hypothetical beast exist) would be immune from monopolies is incorrect.  In addition to the points made by previous posters, I also want to point out that government is not the only barrier to entry into a market.  There are financial barriers to entry as well.

Suppose that telephone company A took a risk and invested billions of dollars into building a nationwide telecomunications infrastructure which they now own.  Company B sees that company A gets outrageous margins for its service, and considers getting into the game.  But there's a catch.  First, company B must invest an equal amount into infrastructure merely to be on equal footing with company A.  Secondly, the entrance of company B into the market triggers a price war which lowers margins and makes it so company B won't recoup its investment for decades.  As a result the management of company B decides that it's better if they focus their energies elsewhere.

This thought-experiment demonstrates that even in a perfectly fair "free market environment", there's a limit to the number of enterprises a market can sustain.  For low-barrier enterprises like coffee shops and grocery stores, that could be hundreds.  In the case of utilities and telecomunications, where there's a high barrier to entry, not to mention pressure to consolidate, the limit could be only one or two enterprises.

Alyosha`
Monday, May 24, 2004

Another reason non serious searches come in first over serious ones, is the simple fact that people tend to find a set of sites they like and stick with them.  How often do you go look for new sites in your favorite hobby?  Personally I have maybe a dozen programming sites I frequent in a week, they are all bookmarked, so I don't go to google to search for them or other programming sites.  Everyonce in a while someone on one of them will mention x,y,or z about programming then I turn to google for relevant info on that topic, and that is if whoever brought it up didn't give a site with good info on the topic.

Brian
Tuesday, May 25, 2004

*  Recent Topics

*  Fog Creek Home