Depending on pending...
If someone was to put together the "Founder's Manual" for pitching a VC, the section on barriers to entry would say something like: "When asked about your barrier to entry, tell the VC that you have several patents pending for proprietary technology." Heck, I'm sure I used the same line when I was raising money for Half.com -- after all, we filed for our first patent in October 1999.
Well, I'm excited to announce that the original Half.com patent (#7,373,317) was finally approved by the USPTO last week - eight and a half years after we filed it (and seven years after eBay acquired Half.com).
Most people don't realize that the average software patent will take many years from initial application to issuance. (According to the USPTO, software patents have the longest backlog of any type of patent. In 2005 the backlog was almost four years...and it has grown larger since then).
Given that a patent needs to be issued before it can be enforced, and given that it takes 4+ years for a patent to issue, you can't really call it an effective "barrier" to entry. Before you can even consider enforcing the patent, your company has already succeeded or failed in the marketplace. And by the time you try to close the "patent door" to the barn, the horse has not only left the barn, it has probably died of old age.
I believe there is little-to-no value in depending on pending patents for a "barrier" to entry.
So should startups invest in patents? I agree with much of what Brad Feld writes in his AskTheVC post on this topic - especially his conclusion that pending "patents are probably the most overvalued 'asset' that startups promote when raising funds". [UPDATE: Brad has a great post today on IP protection]
There is value in owning issued patents -- but that typically comes after you've experienced success in the marketplace. (If a company has failed in the marketplace, their patent filing is typically abandoned/withdrawn). I think Fred Wilson said it best when he said "Patents are like nuclear bombs, you just got to have some". But if you're fighting a war today, it's better to count on weapons you have at hand today -- don't rely on a nuclear program that could take five years to come to fruition.
You can't keep competitors out of your market just by depending on something that's still pending...