Any views or opinions expressed by me in this article are solely my own and do not necessarily represent those of Oracle Corporation, its subsidiaries or affiliates.
Please stop filing extremely short, overly broad patent claims. This may seem like a simple request, one for which it is easy to say: “He’s not talking about me—my claims are reasonably short and broad, but not extremely short and overly broad.” If you are a patent attorney who prosecutes applications in the software-related arts, it’s safe to assume I am talking about you.
Over the past couple of years in my roles as Senior Patent Counsel at Oracle and as Adjunct Professor of Software Patents at Chicago-Kent College of Law, IIT, I have often found myself discussing extremely short and/or overly broad patent claims with patent attorneys and patent agents who draft patent applications. Whenever I suggested the claims were extremely short and/or overly broad, I received responses that were, to my surprise, mostly defensive of the strategy: “this claim would be a really valuable claim if we got it,” and “we might get it—I’m not aware of any prior art on point.” My initial response was “you’re wrong, fix it” with reference to several examples of prior art off the top of my head. That said, at that time, the discussion had become one of competing dogmas (checked by my knowledge of the prior art) rather than one of science.
I am not the first person to make the observation that extremely short independent claims can serve as one good indication . For example, the U.S. Patent Office (USPTO) has recognized that overly broad patents are an issue and has taken steps to address it. The USPTO created a Sensitive Application Warning System (“SAWS”) to provide additional review for improving patent quality. As described on the USPTO web site and in response to a FOIA request, SAWS was designed to provide additional pre-issuance review of, for example, “applications with claims of broad or domineering scope” and “applications with claims of pioneering scope.” The SAWS program received some criticism for being kept secret, and because there were no apparent rules to guide its potentially broad but practically narrow application. The program was retired in 2015.
The USPTO examiners and patent practitioners have also loosely and less formally used a “One-Hand Rule” to gauge extremely short and overly broad independent claims. The one-hand rule is used by some to determine that a claim is unlikely to contain allowable subject matter if it is not longer than the reader’s hand. As will be explained in more detail below, it turns out that the One-Hand Rule (or a more precise variation of it) is a good way to identify low quality patent applications (i.e., applications that are not expected to efficiently proceed towards allowance).
The clear need to improve patent quality has been well documented in both mainstream media and technical and legal sources. The discussion for improving patent quality needs to proceed scientifically, driven by an empirical analysis of data, to avoid the pitfalls of competing dogmas. I recently conducted a study to measure the effectiveness of various prosecution strategies. The study covered over a hundred thousand patent assets pursued by software companies, and for this sample, I found that filing extremely short, overly broad patent claims is a bad strategy in just about every way imaginable.
My study addressed the following questions:
- How short is extremely short?
- How prevalent is the strategy of pursuing extremely short, overly broad claims?
- Are people really getting these claims?
- If so, are some firms/companies better than others at getting these claims?
- If not, how much time and money is being wasted on this strategy?
- What are the other measurable costs of this strategy?
I found that:
- Cases with independent claims of less than 300 characters (or approximately 60-75 words, 4 full lines of text, or less than half the length of an abstract, which is capped at 150 words) had dramatically different prosecution lives than cases with independent claims that were all longer than 300 characters.
- At one extreme, a major software company used this strategy in as many as 13% of their cases as filed. Other software companies used the strategy significantly less frequently.
- Consistently, roughly 0% of the issued patents had extremely short independent claims.
- Nobody seems to be better than anyone else at getting these claims.
- On average, the strategy costs 2-3 more office actions and 1 extra RCE (roughly $10,000).
- This strategy leads to longer (presumably narrower) claims as issued for claims near the threshold of 300 characters, and the strategy also leads to more prosecution history estoppel.
The study and its findings are explained in more detail below and were also detailed in a recent Topic Submission for the USPTO Quality Case Studies.
Through their outside patent counsel, companies pursue extremely short claims to a variable extent, yet to their consistent detriment. The study investigated characteristics of the patent portfolios of four large companies with a wide range of software-related products and services. The identities of these four companies (“Companies A-D”) are being kept confidential. For each of companies A, B, and C, all published U.S. applications and patents were included covering the lifetime of each respective company, as of the beginning of 2016. For Company D, published U.S. applications and patents were included with filing dates ranging over a fixed period of 6 years during an exploratory phase of the study. This article focuses on results from two of the companies, Company A and Company B, that have pursued extremely short claims to varying degrees. The results for Company C fell between the results for Companies A and B. Company D rarely pursued extremely short independent claims.
Companies pursued extremely short independent claims in 3-13% of their cases, but these claims were present in roughly 0% of their issued patents. In other words, nobody is getting these extremely short independent claims through the patent office. For highly developed technologies such as computer software and hardware, it is extremely unlikely for the USPTO to issue a patent with extremely short independent claims. The English language might just require more space to adequately define the scope of these highly developed technologies.
The study focused on finding cases that statistically have almost no chance of allowance in their condition as filed. I found that extremely short independent claims were present in 1% or fewer of issued patents even though originally-filed applications had these extremely short claims much more frequently. I also found that companies differed in how aggressively they pursued these extremely short claims. For example, Company A pursued the extremely short claims in 13% of their applications as filed, and Company B pursued the extremely short claims in 9% of their applications as filed. Despite the variable inputs (13% and 9%, respectively) to the patent office, the output was roughly the same: these extremely short independent claims were present in 1% or fewer of issued patents for both of these companies. In fact, Company B obtained a slightly higher percentage of patents with extremely short independent claims despite the lower frequency with which they were initially pursued. In other words, pursuing these extremely short claims almost never results in an issued patent, but some companies still aggressively pursue them.
Chart 1 below shows the percentage of applications filed and percentage of patents granted for Company A and Company B, respectively, based on the length of the shortest independent claim (rounded down in increments of 100). About 9% of cases filed by Company A had an extremely short independent claim. About 13% of cases filed by Company B had an extremely short independent claim. For both companies, 1% or fewer patents had an extremely short independent claim.
Avoiding the strategy of pursuing extremely short claims could result in significant savings for an applicant (your client) in terms of direct prosecution costs and prosecution history estoppel. The study concluded that, on average, the strategy of pursuing extremely short overly broad claims costs 2-3 more office actions and 1 extra RCE, as well as increased prosecution history estoppel and possibly even narrower claims. This is particularly clear when looking at the independent but similar-looking datasets from Companies A and B. The estimated 2-3 additional office actions cost applicants approximately $6000-$9000 in attorneys’ fees alone. Given the cost of a second RCE at $1700, the total cost of this ineffective approach is roughly $10,000 per case. The study also concluded that the expense was not well spent. For applications with independent claims near the threshold length of 300 characters, those with fewer than 300 characters actually ended up with longer claims as issued than those with more than 300 characters.
Charts 2-3 below show the number of Office Actions received by Companies A and B, respectively, based on varying starting claim lengths. Applications starting with longer, but not extremely long, independent claims (1600-2000 characters, or about 2.5 to 3 times the length of an abstract) on average have 2-3 fewer office actions than applications with extremely short claims. The trend lines are shown in black with a polynomial fit order n=6.
As would be expected, the number of RCEs filed in each case roughly tracks the number of Office Actions, as shown in Charts 2-3, with a delta of about one fewer RCE between the strategy of pursuing extremely short claims and the strategy of pursuing longer (but not extremely long) claims.
Charts 4-5 below show claim lengths issued to Companies A and B, respectively, based on varying starting claim lengths. The local minimum at about 300 characters indicates that claims starting with less than 300 characters are likely to end up longer as issued than claims starting at or slightly above 300 characters. If taken as a percentage of prosecution history estoppel added to the claim as filed, claims below 300 characters are changing in length by more than 200%.
I am not suggesting that longer claims are “better.” If the applicant’s goal is breadth, and if claim length is a proxy for breadth, then shorter claims are better to the extent that they define the metes and bounds of new and non-obvious technology. For independent claims less than 300 characters in length, there is generally not enough content to define the metes and bounds of new and non-obvious technology, as evidenced by the roughly 0% of issued patents having these extremely short independent claims. What I am suggesting is that that a strategy with a 0% likelihood of success is a bad strategy.
I would supplement these empirical findings with my own experience (as a result of attending roughly 100 disclosure meetings a year) that extremely short claims are a good indication of laziness at the disclosure meeting and application drafting stage. A disclosure meeting can last as long as 2-3 hours where the attorney(s) and inventor(s) ultimately understand and reach an agreement on what is new and non-obvious about the technology in light of a wide range of prior art; or, a disclosure meeting can last as little as 15 minutes where the attorney(s) and inventor(s) may understand the field of the invention but end the meeting before the real discussion begins. As shown by this study, the field of the invention turns out to be a bad starting point for patent claims—in other words, laziness is bad, and I am calling you out on it.
For the sake of applicants, I hope all patent attorneys carefully consider the implications of this empirical evidence before pursuing an extremely short, overly broad independent patent claim. In-house counsel and non-attorney clients should carefully consider the impact of this strategy on their patent budget. For a company that files 1,000 cases per year and pursues this strategy in 10% of their cases, the strategy could be costing the company $1,000,000 per year. Outside counsel also need to recognize that, with respect to impressing clients, actions speak louder than words. Any patent attorney can draft a broad claim, but it takes hard work and skill to carefully consider and craft a claim that is expected to:
- provide the needed scope of protection,
- efficiently survive patent examination by the USPTO, and
- survive additional review by the Patent Trial & Appeal Board (PTAB) and courts.
Eric L. Sutton
 Chicago-Kent College of Law, Illinois Institute of Technology, has the highest-ranked intellectual property program in the Midwest according to U.S. News & World Report (March 2016).
 Gaudry, Kate S. “Re: Freedom of Information Act Request No. F-15-00004.” 9 Oct. 2014. Made available by AlyssaBereznak. “SAWS FOIA Response.” Scribd. Web. 10 May 2016.
 Gaudry, Kate S. “Secret PTO Program Subjects Apps To Heightened Scrutiny.” Law360. 3 December 2014. Web. 10 May 2016.
 USPTO. “Sensitive Application Warning System.” 2 Mar. 2015. Web. 10 May 2016.
 just_n_examiner. “The One-Hand Rule.” LiveJournal. 20 Jul. 2005 Web. 10 May 2016. Stein, David. “The New Practitioner/Examiner Relationship: Collaboration For Valid Patents.” USPTO Talk. 8 Feb. 2015. Web. 14 May 2016. Philip G. Emma, “Five strategies for overcoming obviousness”, IEEE Micro, vol.26, no. 6, pp. 72, 70-71, Nov./Dec. 2006, doi:10.1109/MM.2006.110.
 “When Patents Attack!” This American Life: Episode 441. 22 July 2011. Web. 9 May 2016. Goble, Gord. “10 tech patents that should have been rejected.” PCWorld. 21 May 2013. Web. 14 May 2016. See El-Gamal, Yasser M., LaPorte, Lawrence R., Mikulka, Yuri. “Down the Rabbit Hole: Trends in Software Patent Court Decisions Post-Alice.” Lexology. 22 Dec. 2015. Web. 9 May 2016.
 Extremely short independent claims appeared in less than 3% of Company D’s applications and in about 0.2% of Company D’s patents.
 Defined as less than 300 characters, or approximately 60-75 words, 4 full lines of text, or less than half the length of an abstract, as summarized in the findings of the study.
 The data for my study was gathered using claims exported from Innography® for applications filed and for applications issued within each company’s corporate tree, using data from applications as published to measure applications as filed. I then automatically and uniformly determined which claims were independent claims and automatically and uniformly determined the length of the independent claims for each case. I combined this information about the length of independent claims with information from LexisNexis PatentAdvisor® about the number of office actions and RCEs for each case. I would like to thank Addam Kaufman and Tonya Yan for their hard work and contributions to my various patent quality studies over the past two years.
 See Charts 2-3.
 Gaudry, Kate S. “How USPTO Teleworking Program Impacts Patent Applicants.” Law360. 12 May 2014. Web. 9 May 2016. (noting on page 3 the average attorney fees of $3000 for an office action response); Report of the Economic Survey 2015. AIPLA. Prepared under direction of the AIPLA Law Practice Management Committee by Association Research, Inc. June 2015.
 See Charts 4-5.
 Charts 2-5 were capped at initial claim lengths of 2000 characters and, for charts 4-5, issued claim lengths of 4000 characters. The caps were introduced to increase the visibility of the 200-299 character marker.
 The RCE charts were excluded as redundant.
 See El-Gamal, Yasser M., LaPorte, Lawrence R., Mikulka, Yuri. “Down the Rabbit Hole: Trends in Software Patent Court Decisions Post-Alice.” Lexology. 22 Dec. 2015. Web. 9 May 2016.