You might want to look into the editors of Optimizely and VWO. Even though they are A/B testing tools, I've used them just for thinking through quick design changes. From looking at the code they generate, it looked like they are just altering the CSS using jquery (often just assigning absolute positions to elements). So maybe you are already doing something better than that.
Have you looked into working on a small subset of web design? For example, make your editor work really well for Woothemes or sites built on Twitter Bootstrap. It will give you constraints to make at least one kind of popular design format really easy to edit and also possibly an easier way to get users because people recognise those brands.
I don't want to pivot your pivot, but positioning your product primarily for editing existing web pages, rather than creating news ones from scratch might be a good idea.
I don't do everything in it, and the code that comes out isn't quite production ready, but it's damn close.
It's hard to get folks to switch over to a less-than-complete web design tool (and building an unconstrained editor that generates quality HTML is a challenge), but I think it will get there eventually. I'd love to hear your feedback on the current product or advice on how to proceed long term.