When I was at Internet Retailer in Chicago a few weeks ago, I heard at least three speakers give as story that, abstracted, went something like this:
We started off building our own ecommerce platform, then we switched to a vendor supported product. After we almost went broke, we went back to building our own ecommerce platform.
Your reaction to that might be like mine was: "why would a retailer want to spend money building their own platform?" After all, shouldn't they concentrate on their core competence--retailing--and leave software development to the experts?
Here's what it comes down to: most online retailers aren't selling unique products. They're sourcing product from a supply chain that their competitors have access to as well. So, they're all selling the same thing with roughly the same margins. What do they compete on? Shopper experience.
The one thing that can make a huge difference in their top-line revenue is the overall experience that a shopper has when they visit the online store. If it's slow, ugly, full of friction with too many clicks, breaks, doesn't offer features shoppers expect, and so on, shoppers will go somewhere else.
All of these depends on the platform and if you're using the same platform as your competitor, you're reduced your degrees of freedom substantially.
Amazon, of course, is the biggest example of a company that uses a custom ecommerce platform. They're a premiere technology company because that's what it takes to be the Net's biggest retailer. Amazon wouldn't be Amazon if they were running on ATG (ignoring issues of scale). Amazon is the biggest retailer because they run their own platform--not the other way around.
Every business has to know how they compete and who they compete with. In retail you might compete on a unique product, but usually you're competing on price and experience--and only the latter is sustainable.