What kind of initial patenting costs should we calculate?

So I have 2 published patents and 4 pending patent applications and I have been through more patent litigation/enforcement/brokering then I care to remember. At this point in time, I really think you need to consider all the ramifications of "patent protection" prior to spending money therein.

First, I have yet to see a patent published for less than $25K. Sure an ideal patent with limited office actions and a blessed specification might, somehow, go out the door for $10K (good luck). Instead, especially if you are in software land, your patent is going to be relentlessly challenged, responses are going to cost more than you expect and 3 years later you will be in the hole $25K.

Now you are protected during this time...right?

Unfortunately your protection is only as good as your lawyer and your ability to pay said lawyer. A typical IP case can take 5-7 years and literally cost millions - especially if you go up against a large corporation. And that was before the whole reexamination nightmare that the patent reform fiasco threw upon us. There is no such thing as an airtight IP case. I have literally seen memos from defendants blatantly saying that they stole IP - and still the prosecutors lost the case.

Now having a patent does provide some sense of security, albeit a false sense at times, and most people do pay attention to said protection. In the end, you need to figure out how deep your moat is, how long it will take somebody to copy what you created and where you will be financially at that point in time.

If your moat is deep and your work impossible to easily copy then perhaps a patent is not required. If you are creating something that is downloaded to a client and can be reverse-engineered overnight then perhaps you do want that patent after all. Sure investors still clamor for patents but a very detailed prior art review combined with a good legal opinion can suffice in many cases - especially is your secret sauce is hard to obtain without laying it out in a patent.

On the other hand, if you know your company is going big fast then file away as you will have at least a year until you need to start paying money into your prosecution. If you know you will have the money to fight people later then get your patent(s) filed now so you have something with which to fight.

Finally, determining patentability comes down to a thorough prior art and current state of the art analysis. Can somebody as smart as you easily combine current options together to create your solution? Are there prior patents out there that are one simple, fairly obvious, step from your invention? Are your competitors already selling highly similar products with a small change here or there? If you answer yes then you have nothing to patent. If you answer no, well then the question is how much do you believe in your startup's ability to succeed?