I have a basic non-U.S. visitor redirect system in place.

First, you need to get GeoIP set up on your server, which I detail in my post MaxMind GeoLite Country and GeoLite City Made Easy

Once that is set up and running properly, follow these 2 steps:

1) Download my GeoIPFunctions.php file and put it in your GeoIP folder on your server.

Be sure to change the parameter on line 6 to match your server setup.

2) At the top of your PHP landing page, put this code:

Again, make sure the path in the require_once matches the path to your GeoIP files.

Now when non U.S. based visitors come to your landing page, you can send them to another offer that accepts non U.S. traffic.

This system is very simplistic, and a setup that forwards any individual country's visitors on to a specific URL would be a better solution, but I'll leave that for later.

Update: The #83 Alexa rated site has a search box that is vulnerable to link injection.

Update 2: I used a social bookmarking service to bookmark the vulnerable pages, pointing to one of our affiliate sites. I'll collect ranking data now, and 6-8 weeks from now to see if there is any measurable lift from having links to the page that "links" to our site.

Even before these splogs popped up, I had been interested in GeoIP lookup techniques for Google Maps mashup purposes, but the splog trick finally made me figure out how it was done.

DISCLAIMER: I don't condone the use of these techniques for deceptive advertising techniques. At best, they are disingenuous, and at worst, may be illegal, according to the FTC. Like The Force, this technique can be used for good or evil. Which path you choose is up to you.

MaxMind has a paid GeoIP City database service that costs $370 up front and $90 a month for database updates, which they advertise as 99.8% accurate in determining location from IP address. However, they also make a free database available, that is 99.5% accurate. We'll cover the free version since the accuracy difference is negligible for our purposes.

The crucial information is spread out over four pages on the MaxMind site:

APIs for GeoIP products
GeoLite Country information
GeoLite City information
and our API of choice, the GeoIP PHP API

It should be noted that the GeoIP PHP API has three methods of being implemented, each being more efficient with webserver resources than the previous: a pure PHP module, a PECL module that allows you to embed the GeoIP C Library inside PHP, and mod_geoip, an Apache module that you can access via PHP. Since our needs are fairly minimal, we'll go with the easiest to implement, the pure PHP module, which requires no PHP or Apache changes.

1) Download the following PHP files from MaxMind: geoipcity.inc, geoip.inc, and geoipregionvars.php .

2) Download the latest GeoLite City Binary Format file, which is updated on the 1st of each month, and gunzip it.

3) Put all of those files in their own directory somewhere on your webserver, and for this example, we'll name the directory GeoIP.

4) Put the following code in a web-accessible PHP page, changing the PATH/TO parts to match your setup, and view the page in a web browser:

The database says I'm in Los Angeles when in fact we're in Mission Viejo (Orange County), which isn't a huge deal, but shows the possible slight inaccuracy in IP->location lookup.

One thing this could benefit from is a quick state abbreviation to full state name conversion function, so the text would be "Los Angeles, California" instead of "Los Angeles, CA". I think the full spelling looks more natural in some instances. I've uploaded a PHP state abbreviation to state name conversion function for your use.

Update: Danger Brown steals the show for GeoIP made REALLY easy. He's got a two line javascript snippet that does the same thing, but is reliant on the MaxMind servers each time the script is called. My experience with using javascript that is resident on outside servers is that it works great until it doesn't. One day your site will be slow or completely down because the outside javascript is failing to load. Some well crafted code can mitigate the risk, but I prefer to have all executing code on my server. Except for Google Analytics, because I tend to be no match for their servers.

Update 2: I've uploaded my modified geoipcity.inc that includes the abbreviation to state/province function, with ON->Ontario included, and the call to that function in GeoIP_record_by_addr. Note that I could have also put the $record->region_name = Abbreviation_GetStateName($record->region); code within the _get_record function, but I chose to put it in the higher-level GeoIP_record_by_addr function.

Update 3: SEOcracy.com has a post about a very simple way to use the Google javascript API to do geolocation

In a nutshell, Mundu wants to charge $11 lifetime to use their web-enabled iPhone chat software/service, and Nick thinks they are out of their minds to be charging money. He makes two statements that are absolutely ridiculous in defense of his argument:

1) "There are way better ways to monetize software. Offer a free version and drop an advertisement ..."
2) "The marginal production cost of software is zero. That’s what the price should be."

If there are "way better ways to monetize software", then why are two of the richest people on the planet sellers of software? Being a fan of all the new concepts, services and software falling under the Web 2.0 umbrella, this is the one part of this movement that I can't always agree with; software should be free, supported by advertising. I can't find the site now, but several months ago I went to a new Web 2.0 service that was launching, and on its front page it said something like "It has become accepted that this type of service should be free, and be supported by advertising". Wow, way to be proud of the service you've built and the expectations of your users. You've just said your software isn't worth much to your users, and its best feature is that it can be turned into a billboard.

Next, we have the almost-always true statement "The marginal production cost of software is zero." followed by the utterly ridiculous "That’s what the price should be.". Yes, that is one of the beautiful things about software compared to hard goods; the marginal cost of production of a web service/software is zero. Marketing guys love to take this fact and extrapolate it into all kinds of falsehoods, like "therefore the cost should be zero", or "so giving away 100 copies for free costs us nothing.". Wrong. What really matters is the marginal cost of distribution, the cost to have another person use the software. There are support costs, upgrade costs, and for web services, additional server and bandwidth resources, which cost more money. Having the marginal cost of production be zero means you can put a zero in your spreadsheet for that category, where you could not for physical goods, and that's all, nothing more.

If I can take serious liberties with what Nick has written, I think under there is a good argument. If we can say "There are way better ways to market this type of software at the point in time they find themselves", then there is value here. As he notes, Mundu's main challenge is to get a critical mass of users before Apple releases their own chat application for free. To do so, he asserts that Mundu needs a freemium pricing model; free with advertisements, or paid with no advertisements. I'm a big fan of this model, as it allows users to try the service, and make some money from the users who would never pay for it in the first place. It also allows users who find it valuable to pay for it, and clear up some of the valuable screen space (incredibly valuable on a handheld device) that would have been taken up by ads.

However, getting a critical mass of users to beat the big guy to the punch is not the only way to make money, nor is it the best. The best way is to make a superior product, and people will have no problem paying for it. I know this for a fact, because I am in the middle of a 14-day trial period for some contact software for my Windows Mobile phone that is far superior to the built in stuff, and I'll definitely pay the $30 after the trial is up.

So the takeaways from this are: 1) If you want to make some money, and the 800 pound gorilla is breathing down your neck with his free version, make a superior product, and serious users will pay for it. For the rest of the users, monetize them with ads. 2) Having the marginal cost of production of software be zero only means that it is a cheaper thing to manufacture than a physical good. Nothing else.