Build it from Scratch

I’ve never worked anyplace where I haven’t completely revamped my company’s website. This is for a lot of reasons – not the least of which is that I’d rather have to deal with my own screw-ups rather than those of the person before me. On some days, I even feel like my work might even be better than the crud I inherited.

There is only a very small number of really good sites out there. Even where a site was good once upon a time, it may be time now completely revamp it to bring it up to date or accommodate new requirements. This is not so say that good work has not been done – just that a revolution is in order.

Every company in either situation faces the question of whether to incrementally change or rebuild from scratch. While there are cases where the architecture is fluid and the problems minor enough that the incremental approach can work, I’m going to argue in favor of the full redesign.

First, let me start with a theoretical reason (big surprise, I’m sure). When you try to be build a new strategy off a flawed system, you’re pretty much guaranteed to perpetuate at least some of the underlying problems nknowingly. It’s just a product of how we think. So trying to evolve a current site pretty much ensures that some of its issues will be retained.

Now to the practical point: folks, it’s just much harder to try to overlay serious changes over an existing architecture unless it’s the hallmark of flexibility. Even if this is true, there are serious brainpower issues to consider. From this perspective, developing a new vision is so much easier than layering a solution over a flawed foundation.

I tried this once, at a major dotcom site with page views in the +100 million per month. There was so much political investment in the “channel” structure that we tried to evolve the site’s navigation by overlaying an entirely new “superhighway” over this existing structure. This probably cost 3-6 months in extra time to reconcile the opposite approaches in one space. And this is probably a positive outcome compared to how it could go. At least we resolved the experiential issues.

So today, when I find a site with serious problems or growth constraints, I have one statement: burn it to the ground. Out of the ashes, something much better may arise and probably with many fewer brain cells burnt in the effort.