Coder’s Eye

Sometimes I have to remind myself why I am writing a brand-new shopping cart system. It seems on the face of it to be an incredibly foolish thing to do.

I know.

The thing is, I am determined to use modern, systematic approaches to selling, as described in “Call to Action: Secret Formulas to Improve Online Results” (Bryan Eisenberg, Jeffery Eisenberg). I need the flexibility to experiment with the systems, I need the explicit knowledge about how my system works and why.

I cannot be held-up, waiting for the “next new release”, available “real soon now” of this critical component of my store. Also, during this experiment, I can’t sit there and grit my teeth poking at horrible, ugly PHP code, guessing and hoping that I am changing the right thing.

I was going to use Zen Cart, but its latest release has been stuck for far too long, and it can’t even do the very first experiment I want to try, which is to try not requiring registration before purchase.

I actually don’t think it will be so hard.

I’m using Django, so I’m not writing any admin pages at all. What I do write is working very quickly, and I trust it. That’s key for me. I can sit down in an evening and hammer out a new functional app.

I’ve broken down the functionality I want into manageable chunks, small apps that work together. With focus, I am usually able to put a working app together in a remarkably short time, with non-trivial functionality.

For example, I wrote my email auto-responder subscription system in just two evenings. It’s full-featured, too. It has multi-level fallback templates, and generic templates as well. With this system, I can manage an unlimited number of simultaneous email campaigns, with customized signup templates, across an equally unlimited number of domain names. It figures out everything it needs to keep the campaigns separate, and tracks all the click sources as well as abandonment rate.

Now I’m working on the catalog. I got all the model work done in just a couple hours. The templates will be built using my fallback mechanism, so any product will have a default template, but can be customized independently by product line, by product, or by hosting site. See, the apps are additive, and the system is so clean that they work together nearly seamlessly.

Next will be the cart itself, then the authorize.net gateway. Add some shell templates and some design work and I’ve got a cart system!

Yes, I am insane, but I am an insane guy who knows exactly what he wants and who cannot find it anywhere else.

Technorati Tags: django, shopping cart