Why Google Will Always Dominate Its Peers

Anybody can build a search engine. And anybody can take a pen and draw a crude map of the world in a few minutes. But guess what? The most precise free world map/ Internet search engine gets most of the queries.

And the search company with the highest number of queries makes more profits and can afford a higher level of search engine improvement. Google (GOOG) has the leading search map now, and can profitably spend the most money on map improvement.

But this virtuous cycle of search engine development and financing only points to a high barrier to entry for the paid search business -- It does not explain why Google’s search network is probably invincible. Google’s invincibility stems from its market share of humans doing searches.

You see, the biggest network of internet crawling software robots only finds the biggest set of possibly valid websites for your search. And improved indexing algorithms only give a preliminary machine-based preference for these search returns. It is ultimately the gleaned preferences of the humans that have gone before you that give human meaning to the robot searches. Without the human network, the robot index is still only as smart as a bunch of dumb computers. That is to say, it is easily tricked by cleaver humans that want you to visit their spam website. Can you recall drilling down a list of spammy search returns before before Google existed?

So the human user network lets Google's indexing computer understand among other things, which search results are the last ones that people click on (ie. The human found what it was looking for). Other humans that type in the same thing will like this webpage, it goes to the top of the list.

The search network with the most humans searching gets the most accurate search returns in human terms -- And the search engine with the most humanly accurate search returns maintains and extends its dominance.

Some people argue that a breakthrough method of obtaining greater statistical accuracy with half of the human sample size would unseat Google. They argue that Yahoo (YHOO) may find some way to obtain as much search accuracy from a particular search by 50,000 searching customers as Google generates from 100,000.

With large sample sizes, the idea is at least theoretically possible. If we compare this situation to a public opinion survey, the difference in survey methodology could easily matter far more than increasing the sample size from 50,000 to 100,000. However, this idea falls apart when only say 2 people undertake highly specific searches in that particular day. With small samples, the methodology (within reason) is less critical than the sample size. And perhaps half of all searches could be considered small sample size searches -- especially when we introduce timeliness as a variable.

If Google has seen two people searching for something on a particular day, Yahoo, with its second place market share, might not have had any customers at all doing that search. And Yahoo might not be able to draw any up to the minute customer inference, no matter how many thousands of pages of websites Yahoo’s robots turn up.

In order to make any serious headway in search, Yahoo and MSN (MSFT) must stop losing human market share and start gaining it. And in order to do that, Yahoo and MSN must prove to the public that they are not merely equal to Google, but superior. And they must now do so while overcoming the handicap of a smaller human user network to sample and base their search returns on. Absent that miracle, Google will simply continue to return more accurate search returns, and its reputation will continue to spread.

Eventually, even the folks who barely know the difference between an email address and URL will know of Google’s superiority. That will be the end of the other search engines.

Perhaps there is only room for one dominant global search engine. Maybe it is similar to how there tends to be only one consciousness in a person, and one language among a people, and one operating system for communicating among PC's: All are communication systems.

Google already has more than twice the market share of Yahoo and nearly four times the market share of MSN. According to Metcalf’s law (the value of a network increases roughly in proportion to the square of its size) this implies that Google’s search network would be something like four or more times more valuable than Yahoo’s, and something like 16 times more valuable than MSN’s. This is why search is so much more wildly profitable for Google's than the other companies.

So according to Metcalf's law, when Google's market share rises from 50% to 55%, while Yahoo's market share falls from 25% to 20%, Google's market value does not merely grow by 10%, while yahoo falls by merely 20%.

What really happens is that Google's value goes from 50 squared to 55 squared, while Yahoo goes from 25 squared to 20 squared. Google's value thus goes from 2500 to 3025, while Yahoo's value goes from 625 to 400. So Google starts with 2500 relative to yahoo's 625 or 4 times Yahoo's value. Google ends with 3025/400 or 7.5 times Yahoo's value. In other words, that small gain of 5% of marketshare makes Google's value increase by 87% and not 10% relative to Yahoo's.

Disclosure: Author has a long position in GOOG

GOOG 1-yr chart

Comments

[JeffMolander]

Andrew:
What you describe in paragraph 4 is important -- in particular the delicate balance that GOOG is continually striking between ad saturated parked domains (arbitrage sites as an example) and relevant search results pages. Everyone involved -- human searchers, advertisers, distribution partners -- must be closely monitored to extract maximum value (advertising revenue).

GOOG and YHOO both have much to lose if users search experience reaches a certain pain threshold -- one which Jason Calacanis (mahalo.com) would have us believe is soon to be reached (he, and many like him, believe that search is fundamentally flawed... broken... and a 'human guide' model is the future).

As an example, you gave an example of how "The human found what it was looking for" -- and how this enables GOOG to dominate. Others will suggest to you that, as you point out, this is not necessarily true given the trickery that Web affiliates and advertising arbitragers use. There's a lot of user ignorance that gets converted into cash -- at the expense of user experience but not to the point of users leaving. Why? Google still maintains the reputation of offering the most relevant search results -- regardless if this is actually true or not it's the defacto "portal to the Web" based on <strong>human habit <strong>(IMO).

Habits change... but rarely quickly.

Your comment "In order to make any serious headway in search, Yahoo and MSN (MSFT) must stop losing human market share and start gaining it" is on target but isn't it really a question of when humans will stop using traditional search engines as a Web navigation tool (and start using something else -- i.e. social media)? I suggest this is the REAL question and why YHOO may, in the long term, be better positioned to monetize the social Web.

GOOG's success has been based largely on its ability to offer advertisers a black box ad solution based on the 1998 premise of its natural search prowess (best search results -- hands down). It's a black box business -- advertisers can't peer inside. Publishers can't peer inside. Nothing is really controllable but guess what? Removing control (i.e. Panama) SCALES much better. It "dumbs down" search for all parties. As well users are easily tricked into clicking on ads offered up by arbitragers (i.e. the multi-million dollar business of domaining which Marchex, a very important YHOO partner, plays so successfully in).

Combine the "folks who barely know the difference between an email address and URL" with advertisers and publishers who **today** enjoy the black box and you have a complete "ignorance economy" that is responsible for a majority share of all online ad dollars being spent in the U.S.

My point: Is this advertising revenue model the future? Doubtful. Hence, we have everything from Mahalo to Snap.com being touted as the future -- with a dash of "the pageview is dead" for good measure.

How long before a more serious shift in models??

Keep up the good work. Your insights are valued and appreciated.

[akindman]

</strong>

[AndrewMelcher]

Recently I was looking for tips on doing a road trip in mexico (driving Baja California actually) and no internet guides were needed. Google worked excellently. The first page taught me that I probably wanted to buy a cheap Mexico registered "beater" car and sell it/ give it away of it at the end of my trip. For this micro-subject, I had no need a guide site like Mahalo. People like to help out and they post all sorts of useful information without being official guides... modesty aside ;-)

I think that the future of search involves a more robust function like the "this is spam" button in Gmail. Maybe a new Google browser will have some sort of rating feedback mechanism. Maybe there should be five buttons that rate all URL's whenevery you want. 1) Total spam garbage, 2) spammy, 3) neutral, 4) informative, 5)fantastic. The URL's that get a statistically significant number of IP's selecting 1's or 5's get driven to the top or bottom. This would go a long way towards combating search engine spam for popular subjects.

There probably is a place for guides somewhere, but I believe in democracy and "the wisdom of crowds" (a terrific book) . I believe (thanks to this book) that well aggregated large group opinion is nearly always as smart and usually smarter than the smartest individuals in a group... but the group opinion has to always be aggregated properly.

It may take some tweaking, but eventually, Google will figure out how to get you good information on subjects that are now are profitablely jammed with spam.

Also, perhaps the social networking sites are doing so well now because they are sort of hitching a ride on Google's coat tails as mini networks of "artificial" popularity. When Google starts to distinguish this sort of social network popularity from general popularity, there may be a harsh reassessment of the value of these sites. To clarify, user generated content will always be great, but the value of search popularity per user/contributer may be near a zenith.

[c1ue]

There are several things which Google can do to screw up their lead:

1) Quality of searches decreases perceptibly.

Already for certain keywords, the ability to find quality content not cluttered up by paid advertising is difficult. Google has much goodwill to work on, but it is not infinite.

Don't forget that the majority of searches do not yield income for Google; the paid ones pay for all of the other free riders.

This leads to the second weakness

2) Sector specific searches

As Mahalo is trying to do via the 'human touch', what is more likely to be successful is the sector specific search engine.

The Internet is precisely a medium where groups can form irregardless of location and time zone. For many areas of interest, I posit that expertise in the area will long term be more valuable than a great generic search engine.

The question is whether a true aggregator can take on Google in the areas where Google makes money.

The answer to that will be likely something like Craigslist's business model but with higher quality output in the Mahalo vein.

Honestly speaking a lot of Google's benefit at present is just because it is cool and new. I know of a person who recently went into appliance repair and bought adwords on Google.

He is getting a significant amount of traffic from this, but perusal of his adword clearly shows he is the only one in my area doing so - and I live in the techno-center of Silicon Valley!

I would argue that it is not very likely that multiples of this type of existing revenue will be forthcoming - I just don't see many appliance repair types starting a bidding war over adwords.

[JeffMolander]

c1ue:
Indeed but I'd like to just point out that Google just voted against sector specific search with Universal. Some think that's a mistake (you will be interested to read this www.searchengineguide....).