The CashRank Algorithm

CashRank is a new way to reduce search engine spam by limiting indexing to pages that cost at least $1 per year to maintain. As money is a finite resource this automatically removes 99% of spam.

When making a search engine you realize just how much junk there is on the Internet.

It's made for AdSense sites, link farms, blog spam networks, dubious search results, scraped content, sites still under construction or classified ads from 1995, ecommerce sites made from Amazon feeds. There's a lot more but you probably get the picture.

This is useless or duplicate pages that you usually won't find by following a link from a reputable site but that often pollute the search results of your favorite search engine, especially for more specific long tail search terms.

As a search engine we want to have a way to make sure we don't index this kind of low value content. Not only is our users not interested in it, but in the case auto generated content like search results and web applications the number of URLs is infinite so we simply cannot index all URLs.

How We Find Useful Webpages

To solve this problem we made the CashRank algorithm, a way to limit the number of URLs we index.

CashRank is based on the fact the money is a finite resource in the world. The CashRank algorithm basically requires that for a web page to be indexed someone has to pay one dollar per year to someone else. Primarily the owner of the page and secondarily another web page owner that donates his CashRank to your page.

This sounds a bit weird, I know that, but listen:

Anyone buying a domain has to pay a domain renewal fee every year, lets say it's $10/year. That domains index page then has a CashRank of $10.

The first 9 pages that this domain index page links to receives a donaition of $1 and gets indexed. The domain index keeps $1 for itself to earn it's index right.

The Details

This means that from the start every domain will have 10 and only 10 pages indexed, but as some websites are popular they will have inbound links, all donating a CashRank of at least $1. This givs the popular domain a total of $10 in it's own CashRank plus lets say $50 in donated CashRank from inbound links. This would let us index 60 pages from this popular domain.

If the $10 domain link only to 2 other pages, each page will get a CashRank donation of $4.5, that's ($10 - $1)/2

Additionally if a link donates $2 or more of it's CashRank to another page, like in the example above an additional damping factor of 0.5 is applied to prevent the CashRank from expanding to infinity by link loops.

Secret Search Engine Labs