As a web site owner myself I have an ongoing page rank battle to fight. My problem is that there is a famous artist named John Chamberlain so my site is always getting bumped by art-related sites. On the whole, however, I have won this battle (currently I am #2 on the "John Chamberlain" search). I already wrote about this in my article "I'm Feeling Lucky" from last January. Waging this little war has made me aware of the basics of boosting page rank. Ignore this advice and be condemned to anonymity.

Site Content

Make sure your site is readable - People will ignore your site if they can't read it. Check your site in all browsers (IE, Mozilla/Firefox, Netscape, Opera, AOL) and make sure it looks alright. Don't do wierd stuff like use yellow text on a magenta background. Use moderate contrast and do not use any image that flashes or moves.

Simplicity and links - People are attracted to clarity, simplicity and links. For example, look at this web site for Doug Lea. This guy has the highest page rank out of all java-related personal pages for one main reason: his site has good style/content. Don't pooh pooh it. His page rank is much higher than a lot of famous programmers essentially because of the way it is formatted and its breadth of content.

Title is crucial - When your site is indexed the single-most important part of the search will be your page titles. For example, the title of my site is "John Chamberlain, Software Developer and Author". These words are what I want to be indexed. Do not waste words by including fluff like "Welcome to" or "Home Page", etc.

The first paragraph is important - After the title the next most important factor is your first 250 words or so. You should write your pages as though only the first 250 words were being indexed. If you have a highly graphical site this may be a problem because the first 250 words might be useless graphical gibberish. To solve this problem you can setup a bots-only page by configuring your web server to react to the user agent. The server returns a different, text-only page to a robot.

Proof-read - Use a spell checker and proofread your site on an on-going basis. Typos and other such errors look sloppy and will turn people off your site.

Change your content daily - The search engine keeps track of which sites have dynamic content (change daily) and which are static. Dynamic sites receive a higher page rank. The actual indexed content has to change. Just changing a date tag or something will not have an effect.

Dealing with the Search Engines

Submit your site - When you first start out you will need to make sure your site is submitted to Google and the other major search engines. Each one has a form that allows you to do this. Once you are submitted you are on their their list and you do not need to resubmit unless your URL changes (very bad) or you have some special content you want to make sure gets indexed.

Supply robot guidance - Most search engine spiders recognize a robots.txt file which gives instructions about how to traverse a web site. You can use such a file to control the way your site gets indexed. For example, look at http://www.cnn.com/robots.txt to see how CNN does it.

Submit to DMOZ - DMOZ is the Open Directory Project. This directory is extremely important because it feeds Google and other search engines. Not only that it is widely copied so a link there will be rebroadcast by heavily used sites all over the world. A human has to review and accept your dmoz application so it takes time to get and is difficult to change. Carefully prepare your dmoz application because it is the single-most important factor in your page rank. As a rule sites with a dmoz entry will blow away any non-dmoz site in page rank.

Do not try to trick the search engine - Do not do things like include hundreds of keywords in a meta tag or use other tricks to try to fool the search engine into linking you. The search engines have special rules that heavily penalize such behaviour if it is detected.

Cross-Linking and Popularity Building

The core factor in your page rank is how it is cross-linked by other sites. The more other sites link to you the higher your page rank becomes.

Content is King - The main reason another site will link to you is your content. For this reason you want to provide useful content relevant to whatever it is that you do. For example, if you sell sprockets you might have a "Sprocket Design Guide" on your site that allows users to read web pages that describe in detail how to design sprockets. Having software downloads is another key content item. People will link you if you have useful software downloads. By the way this does not mean posting demos and trial ware. You've got to have running, useful software. It is much better to post a small utility program that actually does something than a big useless 10-megabyte trial. For example, on my site I post a lot of Java source code and a bunch of small programs.

Getting linked one site at a time - You can boost your linkage by laboriously contacting web masters of resource sites. You find logical sites where you could be linked and then email the owner asking them to link you. Don't even bother trying this unless you have serious content on your site (see "Content is King" above).

PR - Generating fake news stories is a great way to boost recognition. For example, Arthur D. Little, a well-known consultancy, used to do stunts like having their engineering section make a lead balloon and a silk purse out of a sow's ear. Once their engineers had the item they would call up a bunch of reporters. Very effective.

Getting Slashdotted - A key site for computer geeks is slashdot.com. It is a rolling news blog for techs. Depending on your field/audience a different news blog may be more logical, but for someone like me in technology a key way to build awareness is to get cross-linked in a slashdot news story. I succeeded in doing this on January 11, 2004. A slashdot cross link will generate 50,000 to 100,000 hits in one day so make sure your web server can handle it before hand (make sure there are no big images on your site, for example).

Publish or perish - Another key way to build cross links is to get published. If your site is mentioned in books or magazine articles you will get cross linked. For example, I have written a bunch of articles for programming trade journals. These journals often include an online link to my web site. They also generate cross links from resource sites.

Beware of blogs - In the old days blogs used to get the highest page rank because they would all cross link to each other and thus generated an artificial pool of high link intensity. To compensate for this Google has taken measures which tend to diminish blog cross link importance. Therefore, unless you are a blogger you should shy away from soliciting cross links from bloggers.

Keeping Up With It

Up time - If your URL changes or your site goes down that is very bad. Other sites use link checkers which check if you are there, and if you are not there you will get deleted. It's hard to get cross-linked, don't spoil your work by having down time. It is important to have maximum stability over long periods of time.

Use a link checker - Use a link checker to validate that the links on your site are good. Visitors do not like bad links. For example, if another site is thinking of cross-linking to you, but they go to your site and try a link and it is bad they may change their mind.

Monitor your stats - You can use a program to generate statistics from your web server logs. I use a free program called Analog. It puts out a web-accessible report which will tell you a lot of key stats for your site. By monitoring these stats you can gauge the effectiveness of your page rank strategy. This is what my report looks like as of today.

The aging factor - The older your site is the more popular it will become. For example, look at my report as shown in the previous topic. I have not really done any major work on my site since around January 2004. As you can see from the report my page hits per day increased exponentially during the fall of 2003 when I was actively working to popularize my site and then went into slow growth mode. This slow growth has a snowballing effect. As people cross link to you, your page rank organically grows, but it takes a long time. For example, even though I did a bunch of major things in January of 2004 (like get slashdotted) my page hits now (when I am doing nothing) are more than double now what they were then. This is purely because of the aging factor.

The Economics of the Web

The plight of the small company - A lot of small companies create web sites in hopes of selling their products but experience little result. Nevertheless having a good web site is vital because any potential customer will visit your site and base part of their perception of you on it. The key to making your company site get noticed on its own is to add interesting content other than your own product materials.

Paying for links - You can boost your page rank by shelling out bucks for links. Most big portals like MSN do not allow alien linkouts on their home page, but most minors do. For example, drudge.com, a major news portal has ads with direct links to their advertisers. If the ad stays there for a long time (like a month) it will generate page rank. Since a site like drudge has massive page rank, its cross links are powerful. Be prepared to pay big time.

Advertising effect - It takes a lot of visibility to have a genuine advertising effect. For example, I get about 250 requests per day, which is very good for the kind of site I have. In comparison, web advertisers generally are not interested in banner advertising on your site unless you get more than 10,000 hits per day. My 250 requests per day gives me about 7500 requests per month. I get about one email from a viewer a month. I get about one job/interview offer per three months. That gives you a sense for the value: basically you need to get about 10,000 hits for every one time somebody actually wants to talk to you.

---------------------------------------------------

In the end what does all this get you? Page rank. When someone does a Google, Hotbot, or whatever search your site comes right up at the top!