The main one is how customers are going to react to your product. Spending a few hours on XML sitemaps and a build process for minification might seem like a good idea if you are a perfectionist, but if you end up spending time on these things it means less time for other features.
It's not so much about solving issues to make the perfect product up front, but more about making progress quickly. Number of iterations is one of the best predictors of product quality, and by spending time on things other than getting out the next iteration, you're slowing down your rate of iteration. It might be tempting to think of a website as "done", but a launch is just the beginning.
For example, the biggest issue with our product at launch (http://sonalight.com, a mobile app that allows you to text through voice) was the pricing structure. We realized it was better to give the product away for free instead of limiting usage. We would never have realized this if we hadn't launched and heard complaints from our customer base.
It's not so much about solving issues to make the perfect product up front, but more about making progress quickly. Number of iterations is one of the best predictors of product quality, and by spending time on things other than getting out the next iteration, you're slowing down your rate of iteration. It might be tempting to think of a website as "done", but a launch is just the beginning.
For example, the biggest issue with our product at launch (http://sonalight.com, a mobile app that allows you to text through voice) was the pricing structure. We realized it was better to give the product away for free instead of limiting usage. We would never have realized this if we hadn't launched and heard complaints from our customer base.