Developing a patent strategy for startups


There are six questions to ask when considering patent strategy for startups: (1) Why should we build a patent portfolio? (2) Which inventions should we file as patent applications? (3) How often should we file? (4) When should we file? (5) Where should we file? and (6) Who should we engage for our patent work?

Why do startups build a patent portfolio?

Fortune 500 companies build a patent portfolio to enforce against a competitor,  to generate licensing revenue, and to market technical and creative ability. None of these reasons, however, applies to startups. Enforcement against a competitor is too expensive, the return on investment in generating licensing revenue is too low relative to the time and effort required, and there are better alternatives to marketing technical and creative ability. In the short term, a more valuable approach is to file a few pending patent applications that can help a startup reinforce a technology narrative to an investor, create a hurdle for smaller competitors, and establish “background IP” for technology partners. The significant impact of building a patent portfolio, however, comes in the long term. A portfolio of issued patents can deter patent infringement lawsuits from larger competitors and can increase valuation during an acquisition or an IPO. Our patent strategy has deterred patent infringement lawsuits for Twilio (which competes against AT&T) and Farmlogs (which competes against Monsanto) and has created significant value for Cruise (acquired by GM for $1 billion) and Accuri Cytometers (acquired by Becton Dickinson for $200 million). When a startup first stops to truly understand the reasons to build a patent portfolio, it can then focus on the appropriate goals and budgets for its patent work.

What inventions should startups file as patent applications?

In an ideal situation, patent applications are pursued if they are both highly patentable and highly valuable. Identifying and prioritizing inventions that are highly valuable can be a daunting task for large companies, especially if they have multiple divisions and product lines. For this reason, larger companies often optimize for identifying highly patentable inventions and then filing hundreds to thousands of patent applications with the hope that some are also highly valuable.

The system that larger companies use to accomplish this can be considered a bottom-up approach, which includes incentivizing engineers and scientists to identify their own inventions and submit an invention disclosure form to a patent committee that evaluates and selects the inventions that are highly patentable. The fact that the low-level engineers (and sometimes even the patent committee!) do not understand the bigger picture and do not know what inventions are the most valuable to the company does not matter, because the larger company will simply file patent applications with a “quantity over quality” approach. Filing hundreds to thousands of patent applications on an annual basis is really expensive in terms of both dollars (in the range of $10 million) and time (in range of decades of people-power).

Startups do not have the luxury of either resource, but are often misguided to institute a version of the bottom-up approach when building their own patent portfolio. This is not only wasteful but also dangerous.

Failing to file patent applications on the inventions that are both highly patentable and highly valuable is clearly problematic in relation to a startup’s ability to deter patent infringement lawsuits and increase valuation. But, filing patent applications on the wrong inventions is potentially worse, because this additionally wastes precious resources of time and money and, because patent applications are automatically published by the patent office 18 months after the filing date, teaches competitors how to make and use their (unprotected) inventions.

Why does the bottom-up approach fail for startups? Startups, which are often not large enough to harness the wisdom of a crowd, struggle to correctly identify inventions that are highly patentable.

Because patent applications are not published for 18 months, there is no way to conduct an accurate patentability search. There is another problem with the bottom-up approach: I have encountered an inverse proportionality between the brilliance of inventors and their ability to identify their inventions and evaluate the patentability of their inventions. Average people think that all of their ideas are brilliant, while brilliant people think all of their ideas are average. In other words, a brilliant startup chief technology officer (CTO) cannot accurately identify the inventions that are highly patentable.

So, what approach should a startup use to identify inventions? We recommend a top-down approach. In a top-down approach, we first start with the two to three core differentiators that articulate the reasons the startup will succeed in the marketplace. This could be as simple as the REST application programming interface (API) and multitenancy attributes of Twilio or the low cost and compact features of Accuri cytometers.

Next, we help our clients identify the technologies that enable the core differentiators. And finally, we interview the engineers and scientists to capture and select the inventions that support these technologies. Instead of a “quantity over quality” approach to the patent application, we recommend the opposite. Patent applications with better claims and with more embodiments and variations in the specification will overcome the rejections from the patent office.

While Fortune 500 companies identify patentable inventions and then use quantity to get great patents, startups should identify valuable inventions and then use quality to get great patents.

How many patents should we file?

In my experience, the ability of a patent portfolio to deter a patent infringement lawsuit from a competitor has a value that can be graphed as an “S” curve based on the number of issued patents. The reason is based on the power of exponential numbers. A defendant in an infringement lawsuit of one patent can expect to invalidate the patent or avoid infringement roughly 50 percent of the time. But to escape the lawsuit, the defendant must try to invalidate or avoid every patent in the portfolio. If every issued patent in the portfolio offers a 50 percent chance, then, as the number of patents in the portfolio increases, the chance for the defendants to escape the lawsuit decreases at an exponential rate. While escaping one issued patent might be easy, escaping a hundred issued patents is close to impossible.

In my experience, the sweet spot is around 15 issued patents as the “S-curve” rapidly climbs from five issued patents to 25 issued patents and tends to increase only marginally thereafter.

On the other hand, the value of the patent portfolio is more linear. Large companies that sell hundreds or thousands of issued patents in bulk tend to fetch $500,000 to $1 million per issued patent. For instance, when Facebook bought 500 issued patents from IBM, it paid $1 million each.

There is probably, however, a limit to how many valuable and patentable inventions that a startup can produce in a given year.

With an understanding that we have the S-curve for the ability to have a patent portfolio to deter an infringement lawsuit and a linear relationship for the value of the patent portfolio, we can now consider the timing for future events for the startup. A startup is not likely to be sued by a competitor until the startup has reached $100 million per year in annual revenue. Depending on the startup, this revenue milestone takes many years to reach, but can often be predicted with enough accuracy within a two-to three-year range. Similarly, most startups can predict an acquisition or an IPO with enough accuracy within a two- to three-year range.

In years past, the expected four- to five-year life cycle of a patent application would make it impossible to hit a moving target that is two to three years away. But now we now have the ability to “fast-track” patent applications (for only $2,000 in government fees) and quickly move from filing to issuance in less than one year. Thus we can set a goal to have 12 issued patents in five years and 25 issued patents in seven years, and work backwards to determine how many patent applications should be filed on an annual basis between now and then.

The actual pace of patent application filings often mimics the valuation and engineering headcount of the startup. It is typical for our clients to file 2 to 3 patent applications in the first year to cover the core differentiators, 4 to 6 applications in the next couple years to cover the improvements, and then 6 to 12 patent applications on an annual basis to pursue the features enabled by the core technologies. These patent filings, however, are always dictated by the goals of the patent portfolio.

When should we file patent applications?

To maximize the success rate of a patent application, one should attempt to get an early filing date to beat the competitors in our first-to- file patent system, and one should include more details in the patent application to distinguish from prior inventions. These twin goals (file earlier to beat the competition and file later to discover more details) appear to be in conflict. An appropriately timed provisional application followed by a full patent application, however, solves this.

I have a motto that good ideas are simply not patentable. In my experience with over  2,000 patent applications, the inventors that have  built and tested their inventions have discovered the important details that help distinguish their invention from prior inventions. This level of inventing typically does not happen during a morning jog, a shower, or any other eureka moment but rather happens with a great team that has significant funding and focused direction. Thus, the optimal time for a full patent application to be filed is after the invention has been built and tested. The optimal time for a provisional patent application to be filed, however, is exactly one year before this date.

In the software space, technology development is more predictable. And thus, when an invention has been conceived and it is believed that the invention will be built and tested within a year, we encourage our clients to file as soon as possible. In the hardware space, however, technology development is often less predictable. And with our clients in the hardware space, we often encourage them to delay the filing of the patent application until they are confident that they will build and test the invention within the next 12 months.

The twin goals to file early to beat a competitor and to file later to distinguish from prior inventions do not apply equally across different technology spaces. For instance, we often encourage our startup clients in the clean tech space to delay the filing of their provisional patent applications because the space in which they are inventing is often very crowded, and the goal of distinguishing from prior inventions is more important than the goal of beating competitors. In contrast, we often encourage our startup clients in the software space to speed up the filing of their provisional patent applications because the goal of beating competitors is more important than the goal of distinguishing from prior inventions.

Where should we file patent applications?

The question of whether or not to pursue foreign patent protection is, by far, the one area of patent strategy that produces the most anxiety and, unfortunately, the most regret.

Our most successful startups often regret not filing more foreign patent applications, while our less successful startup clients often regret spending so much money in the pursuit of patent protection in faraway lands. For this reason, we spent a significant amount of time analyzing the historical data of our more than 250 startup clients, and we found that spending roughly 30 percent of the patent budget in the pursuit of foreign protection was ideal.

Pursuing patent protection in the United States is expensive, and pursuing patent protection in foreign countries is no different. One can expect to spend approximately $30,000 in the pursuit of issued patents per foreign country. When considering the 70:30 ideal split within the patent budget, and knowing that the foreign patent applications (approximately $30,000 each) are almost as expensive as the domestic patent application (approximately $40,000), one can quickly calculate that the ideal pattern is to file two foreign patent applications for every five U.S. patent applications.

There are three factors to consider when choosing which countries to select for your foreign patent applications: (1) where might you make and sell your product in the next three to five years, (2) where might you have competitors that make and sell an infringement product in the next three to five years, and (3) where might a potential acquirer of your startup be located? Keep in mind that while there are over 200 countries with patent systems, one can cover a very large portion of the global market by filing patent applications in the United States, the European region, and China. When our clients choose to pursue foreign patent protection, we often recommend filing in Europe and There are, of course, many exceptions. Our medical device startups, which often have smaller but more valuable portfolios, often pursue patent applications in Canada, Australia, and Japan as well as Europe and China. And, our manufacturing startups often pursue patent applications in Japan and Mexico as well as Europe and China.

Who should we engage for our patent work?

There are several different roles in building a highly functioning patent portfolio: (1) the patent strategist who determines the why, what, how often, when, and where of the patent portfolio, (2) the patent agent or attorney (“patent associate”) who interviews the inventor and crafts the patent application, and (3) the technologist who reviews the patent. The CTO is almost always the person who assumes the role of the technologist, but there are a few different ways to structure the patent strategist (who often has seven or more years of patent experience) and the patent associate (who often has two to six years of patent experience). One model is to hire an in- house patent counsel as the strategist and use outside patent firms for the patent associate. The challenge with this structure is that a good in-house patent counsel can often command a salary in the $250,000 range. Another model is to hire a consultant as a part-time in-house patent counsel and use an outside patent firm for the patent associate. While this solves the financial challenge, it often fails because the consultant and the associate (who are both external to the startup) rarely communicate, and the patent strategy is not properly implemented. Twilio, like all of our other clients, used a third model: engaging a patent law firm for both the strategist and the associate. I was the person who designed the strategy and one of my associates was the person who crafted the patent applications (while the Twilio CTO was the technologist who reviewed each of the patent applications along with the inventors).

If a decision to use a patent law firm is made, the next question is to determine the best fit for the startup. I recommend optimizing for four factors: (1) experience, (2) technical background, (3) startup focus, and (4) proximity. The Supreme Court has stated that patent applications are the most challenging of all legal documents. It pays to work with someone who has traversed the steep learning curve of developing patent portfolios and writing patent applications. It is also important to work with someone who is fluent in your technology.

Patent applications stand and fall based on the words that are chosen in the claim section of   the patent application. The patent associate  must be fluent in your technology to be able to choose the right words. For instance, while I am fluent in mechanical, electrical, and software technologies, I could not write a high quality patent application on a pharmaceutical invention.

I simply do not know the right  words.  As  I hope it is abundantly clear, patent strategy for startups is wildly different than patent strategy for Fortune 500 companies and, for this reason,  I strongly recommend that startups work with someone that has extensive startup experience.

Finally, I recommend that startups choose someone that they can meet with and brainstorm in a face-to-face manner on a regular basis.


By answering why should we build a patent portfolio, which inventions should we file as patent applications, how often should we file, when should we file, where should we file, and who should we engage for our patent work, startups can build a patent portfolio that deters patent infringement lawsuits from their competitors and increases the value of their startup.

Jeffrey Schox, Founding Member and Patent Attorney, Schox Patent Group