Home | How To Build Next Big Search Engine

The Next Big Thing Is "Be Small"

September 8, 2010

If you ask people "Is the Google a good search engine?", a typical answer will be "Sure it is!". Really, why not? Google provides us with millions results. And almost immediately. However, if you dig deeper and start asking related questions, the situation may suddenly turn to the opposite:

- Do you think Google gives you the best answers?
- Yes, of course!
- How do you know they are the best?
- Everyone says that.
- Did you ask them how they know that?
- OK, my own personal experience says it.
- Always?
- Well, I do skip some results, but most of them are good.

And now it's time to ask the main question:
- DO YOU KNOW HOW MANY GOOD RESULTS FOR YOUR ENTRY EXIST? Do you know Google has BILLIONS links in its database? Have you seen them all? So HOW YOU CAN BE SURE Google shows you the best?!!

The unlimited choice is a key factor for Google's success. Considering high level of dublicates on the Web, it seems the biggest Google's secret is that almost ANY result from millions similar ones may be called the best one!

It's practically impossible to find out a precise rank for any given page in Google (its public PageRank services may show outdated values; besides, the PR is just a part of Google's formula, and seems, - not the main one). It's difficult even to say how many billions pages we are talking about: Google says it has discovered 1 trillion (1,000,000,000,000) unique Web pages on the Internet (adding to that "Strictly speaking, the number of pages out there is infinite"), but does not say how many of them it indexed.

Americans love gigantic scales. The Google started as an attempt to digitize the entirety of the universe! "Let me tell you what the challenges are of a search engine", Sergey Brin said. "You have to index the entire Web." But it's useful to remember there always exist a more poor world. Where Less Is More.

So, the Google conquered the world by indexing big amounts of pages: "Google's leaders believe that it can facilitate breakthroughs in numerous realms simply by making unprecedented amounts of information available". It's very good. But do you really think people will read all the data?

To compete with Google you don't need to be as big as Google
I realized that during testing my Mini-WWW. The size of database is not really important (Quil claims it's bigger than Google, but hasn't beaten Google yet). What's the point in indexing the infinity, if people never go beyond first 1-3 pages (SERP) anyway?!!

Why would not call the Google a "Search Engine With Just Three Result Pages" then? Well, it may be fair. However, it's not about Google only. Practically any search service is used for viewing just few top results most of the time. Consider it a psychological human factor. Anyway, it definitely should be applied for building your own search engine :)

By the way, do you know that in the most cases Google never shows you ALL its search results? Say, if the number of found documents for the term you search is a million, you will be able to view a first few hundred of SERP only. For example, Google for "mini" finds 458,000,000 results; nevertheless, stops showing them after 747 (page 75). At this point there is simply no "next" link.

But what about the rest? Nope. Access not allowed. It's probably a strong suggestion that you should always trust the choices Google makes. Well, we probably would. Provided we know and are able to discuss its algorithm. Can we? Is Google's ranking process transparent or clearly open? As the famous saying goes "Trust, but always check". Are we able to check the Google?

What Google actually does is creating its own sub-Web, - a privately ranked (with a secret algorithm) collection of links to high-ranked destinations of the Web, as Google sees it ...
Why not to apply the above tricks during creating your own search niche?

So, if I ask you now whether it's important to use a search engine that is as big as Google, you may say "Yes". But do you really need to open a billion documents? Instead of indexing all available websites, the Mini-WWW includes only minimal pages!

Google accesses the Web.
Mini-www - the mini Web.

How to outperform Google
I'm not trying to compete with Google! Because we are so different. (When Steve Jobs was asked why he entered into competition with Google, he said he simply ... didn't. He just wanted to create a best product. Then suddenly Google decided to compete with Apple). But if you ask my advice on that, the answer would be simple:

The Next Big Thing Is "Be Small"


COMMENT

About Blog Submit New!
Copyright(C) Beloy ::: Mini-News