Affiliate Tracking Technology Comes Full-Circle
Saturday, December 29th, 2007It’s been about 10 years now since I wrote my first affiliate “script”. And it really was a “script” - not a full-blown application.
I was an electrical engineer at Motorola at the time, trying to sell direct marketing books (hardcopy, not ebooks) over the internet for some extra income.
In order to add some interactivity to my own site, I learned the predominant web-server script language of the day, “Perl”, and I was on my way. Soon I started getting requests from other webmasters to write scripts for them
One of the first people to actually PAY me for a script was Allen Says, of “Internet Marketing Warriors” fame. As I recall, he wanted a script that would take an identifying code (”affiliate code” in today’s vernacular) and propagate it from one page to the next so that it would eventually be inserted into a hidden form field on the order form.
If I remember correctly, the script was 280 characters in size, and Allen paid me $250 for it - that was almost a dollar per character, which at the time was heady stuff indeed! (Actually, it still is - I don’t think I’ve ever been paid that much per character since!)
This was shortly after Amazon started their affiliate program, and it seemed like a great idea. So I decided to create a full-blown affiliate program system, which I started selling in 1998.
Back then, browser cookies weren’t used for affiliate tracking. Either the affiliate code was propagated from one page to the next, or static copies were made of the site pages for each affiliate - so their affiliate ID was hardcoded in these static pages. This was done automatically when the affiliate signed up.
The problem with both of these approaches is that affiliate site “templates” were required. They were just standard HTML pages with some “merge codes” added. For instance, if I wanted the affiliate ID to show up in a certain place on the page, I’d use the merge code of AFF_NUM in that location in the template page. Or for the affiliate’s name, I’d use “AFF_NAME”
It wasn’t difficult, but it was more than most people wanted to deal with, especially if they had a lot of pages on their site.
In order to prevent “affiliate abandonment”, the entire site would need to be templated, OR a portion of the site would need to be templated and designed such that no links led to pages outside the templated pages.
With the widespread adoption of cookies, these things were no longer a concern. You merely “tagged” the visitor with an affiliate cookie when they entered the site, and then read it back again at the time of the order to determine which affiliate got credit for the sale.
At least that’s how it worked in theory . . . as long as someone had their cookies enabled (which most did).
But in the years since then, the browser environment has become more and more hostile to cookies. The default security settings are more discriminatory and various applications designed to “protect” the consumer delete cookies on a regular basis.
In this day of dynamic IP addresses (I’ve seen my own DSL IP address change while in the middle of a shopping session!) using the IP address as a cookie-backup system is hit-and-miss at best.
It seems that as soon as new tracking methods come out (such as the “flash cookie”), new methods are devised to defeat them for the consumer’s privacy and security.
Natural consumer behavior is another factor to consider. For instance, in a recent high-profile promotion, I tracked the orders that didn’t get attributed to the correct affiliate. I actually picked up the phone and called a couple of the customers, asking them about the circumstances surrounding their order.
One person had received the promotional email on their work computer, but had waited until getting home to place the order. The 2nd person had used a different computer for their order as well. What kind of affiliate tracking technology is going to overcome that? Ironically, it’s the “old” technology.
If you want to make absolutely sure the correct affiliate gets credit for a sale, provide them with a static page hard-coded with their affiliate ID. And have that page link to an order page that is also hard-coded with the affiliate’s ID. Make sure their affiliate ID is coded visibly on the order page for cases where they FAX or call in the order. And don’t have any entry or order pages OTHER than the affiliates’ - that way it’s not possible for someone to make an “end-run” around your affiliate.
In this case your promotional home page (e.g. “www.example.com/widgets”) might have something like this:
| Oops, you didn’t type the complete web address! Your link should be something like this: http://www.example.com/widgets/abcde
|
Alternatively, you could create static pages for all of your affiliates for a particular promotion using one of the available “site replicator” scripts. You’ll have to do a bit of work to import your affiliate list and “replicate” the promotional pages for each one, but using a replicator script should at least make it feasible to do so (doing it by hand is NOT feasible if you have more than a couple dozen affiliates).
In fact, I just remembered I created a free “feature limited” version of my affiliate software some time ago - it will replicate pages just fine, and in fact is very similar to what I used for a couple of past high-profile client launches to create static pages for the VIP JV partners. You can download it here:
http://www.paulgalloway.com/utilities/yoapfree.zip
This will probably be part of a paid package at some point in the future, so download it now while you have the opportunity.
Note: Please don’t ask for support for this free software unless you’re willing to pay for my time ($150/hour, 1 hour minimum).
Keep in mind that this software was designed to replicate pages in real-time as people filled out and submitted a “signup” form. If you already have a bunch of affiliates that you want to replicate pages for, you will need to populate the affiliate data file first, then run the “Update” utility to create the replicated pages.
If you’re handy with a text editor, you can sign up using the included signup form, and then examine the “adm-dealers.txt” file to see how the data is stored. You can then add your existing affiliates to that file using the same record formatting.
The above “home page” example can also be used if you’re using standard cookie-based affiliate tracking - at least you’d make certain the customer comes in through an affiliate link.
But there will still be “breakage” in cases where people have cookies disabled or have their browser security set to reject cookies - so if you can capture and propagate their affiliate ID from one page to the next, that would overcome the cookie deficiencies.
The free software provided above will also do this kind of propagation for you, even without importing your affiliate list - just read the section on “dynamic pages” on pages 22-24 of the user guide.
I don’t know about you, but I’m very curious about where affiliate tracking will be in 10 years. I really will not be surprised if using old, boring “static pages” is still the most reliable way to track, even then!
Best,
Paul
