This question does not exactly keep me up at night, but if it did, my answer might sound like "Our most important product isn't the one we're selling, it is the processes we use to make and market the product(s) we're selling and will sell in the future, and those processes are non-trivial and so ridiculously superior to what the competition will be using that they will
never catch up."
IMVU or Zygna are good examples of this. Neither has a patent wall around them, either is about as well defended as any business based on computers can be. The reason is that, while you can make your own Big-Chested-Anime-Chicks-Farming-Together-Chat-Client in not a whole lot of time, you won't have their creation process... and they're learning so fast about their markets that they will bury you. The advantages they get from the whole lean startup thing -- fast learning and fast production -- are compounding, because it buys them more traditional advantages like network effects, kickstarts their virality, gives them money to buy insane amounts of FB advertising, etc etc.
This is totally not limited to these market segments. (If either of them were in search-focused verticals I'd tell you about how this helps build the self-reinforcing authority cycle for SEO, too.)
