Startup IP due diligence

Maat – goddess of truth

Startup investors know that they are mostly investing in just your IP.  So get your IP ownership situation under control before due diligence starts.

What does a startup investor get for his money? Startup companies are really just a bundle of information (usually ideas), people (founders, contractors, consultants), and things (tangible property, cash). A startup’s “things” will usually have minimal value. The people can walk out the door at any minute. So all that an investor may actually get for his money is some sort of ownership option in the information – the startup’s intellectual property (IP).  Prudent startup investors thus often conduct a fair amount of due diligence on this “IP”.

A startup’s IP can be viewed as being a mixture of alleged inventions (e.g. actual or potential patents), alleged copyrights, alleged trademarks, and alleged trade secrets.  At least some of this information may still only exist in the minds of the founders, contractors, and consultants. What is potentially real here, and what is not?

Sophisticated investors also know that according to our IP laws, unless actions are taken to protect the IP, the IP may either still belong to the individual that created it; or it may have become public domain.  Investors also know that startups often don’t fully understand IP law.  Has that intriguing startup already accidentally lost its key IP?  Was it missing essential IP to begin with?  This is why investors want IP due diligence studies.

Some ways that IP can be missing or lost include failure to obtain written IP assignment contracts, inadvertently placing IP into the public domain, and reliance on IP owned by someone else.

Failure to obtain written IP assignment contracts: IP assignment agreements should be obtained from anyone who has materially contributed to the company’s IP. Don’t assume that just because you paid for something, your company owns that IP — you often don’t. Try to do these IP assignments up front, because the longer you wait, the more leverage the other party has. For example, they could make unreasonable demands, or deny assignment altogether.

Inadvertently placing IP in the public domain: For patents, assume that anything you publish before a patent application is filed can be used against you, and be careful to update your patent filings as your technology advances. For trade secrets (i.e. important, not-generally known information which is often not patentable), take active steps to restrict access, require confidentiality agreements, and of course never publish this at all.

Reliance on IP owned by someone else: For patents, consider prior art searches to reduce the chances that you are infringing on other patents. For copyrights, check the status of your software licenses, and that you and your contractors are not copying other work. Avoid using names that might confuse customers from related companies, since this may cause trademark problems. For trade secrets, stay away from inside knowledge (not generally known in your field) that you or your contractors may have obtained from other companies.

Remember that from an IP due diligence perspective, if you can’t prove you did things right, the assumption will probably be that you didn’t.  So keep your documentation on file.

Open source software licenses for startups

BSD open source license
BSD open source license

Open source software is a wonderful thing.  The open source community has given us a gift pack animal that everyone from individual hackers to the largest companies in the world can ride.  However not all open source software licenses are alike. In particular, some of the most famous open source software comes with legal obligations that can be hazardous for startups.  So if you are doing a startup, resist the temptation to immediately grab your favorite open source software and start hacking.  Instead first take a bit of time to look that gift pack animal in the mouth.

What’s Gnu? Back in the 1980’s, Richard Stallman, who wrote the original and very influential Gnu OS open software license, had a deeply held opinion that in order to create an open software sharing community, it was necessary to “poison the well” for many commercial uses. This underlying hostility towards commercial applications is very evident in “The GNU manifesto”.

These “poison the well for commercial use” Gnu concepts in turn influenced the open source GPL (General Public License), which Linux and MySQL use; along with a number of other popular open source licenses.  Some of these open source license terms can negatively impact your ability to patent your software, as well as your other attempts to make your business profitable.

This problem scares investors.  Sophisticated investors now frequently include open source software questions as part of their routine, pre-funding, due diligence process. So yes, to avoid starving, this stuff matters.

Enter BSD: Not all open source software licenses have these problems. In the 1990’s other software experts decided that an open software sharing community could develop without also poisoning the well for commercial uses.  They developed the Berkeley Software Distribution (BSD) license (initially for their Unix-like operating system). The BSD license encourages sharing, but does not usually limit commercial use or patents.

Although BSD type open source licenses and software are not quite as famous as GPL open source software, a large number of well-respected and highly reliable BSD distributions are available.  Like Linux, there are different flavors of BSD for different applications. Look into FreeBSD for large scale servers, OpenBSD for secure applications, and NetBSD for smaller scale devices.  There are also BSD web servers (e.g. Nginx, httpd, Lighttpd), BSD databases, and many other BSD licensed open source applications.

Use your business model to pick your open source software (license), and not the other way around.  Google’s main revenue is from advertising.  They could afford to base Android on GPL licensed Linux. They have to make Android available for a free download, but given their business model, this isn’t a problem for them.  By contrast Apple has a business model based on selling closed devices.  Apple can’t survive with non-proprietary software. As a result, Apple chose to build iOS and OS X on a BSD foundation.  So pick the license that is best for your business (hint, think permissive or public domain).

Medical device patents

Medical device patents
Medical device patents

Why are medical device patents important? New medical devices are both expensive and risky to produce. Without patent protection, few new medical devices would ever reach the market.

The number of filed and granted medical device patents has substantially increased over the last few years, with certain US states (California, Minnesota and Massachusetts) leading the way.  Certain countries, in particular Japan and Germany, are also heavily active in this area.

Write carefully: In contrast to some high tech areas, such as computers and software, which often rely on a larger number of lower impact patents, the medical device industry tends to rely on a smaller number of higher impact patents.  What this means is that any given medical device patent is more likely than average to be subjected to a high level of competitor scrutiny.  Remember also that in the medical device field, both the patent office and the courts may view the “person having ordinary skill in the art” (PHOSITA) as being highly skilled, such as a PhD or MD. Thus care should be given to write accordingly. It is helpful to include a lot of detail and discuss alternative approaches. Research and disclose prior art, and of course try to write claims in a manner that distinguishes over prior art.

Remember Europe: Medical devices are an international market, and often due to slower FDA review times, US firms first introduce new products in Europe.  Many US medical device patent applications end up being filed in Europe.  Thus European patent rules are often important.

Although there are a few glaring exceptions, US medical device patent eligibility rules generally tend to be expansive. The USPTO typically reviews these patents in its 3700’s art unit.  In recent years, this art unit has continued to grant medical device patents according to its historical percentages.

European patent eligibility rules are also generally quite permissive with regard to “gadgets” and in-vitro devices, as well as technical aspects of in-vivo devices.  However Europe (See G-II 4.2.1) doesn’t allow “methods of treatment of the [living] human or animal body by surgery or therapy and diagnostic methods practiced on [usually interpreted as “in”] the human or animal body “.  Fortunately the Europeans tend to read this exclusion narrowly, so this isn’t actually quite as restrictive as it sounds. However from a claim writing perspective, be careful of method claims with limitations that read on the patient’s body.  Try to focus more on the details of how the gadget (device) itself works.  Watch out for methods that remove a tissue, treat it, and then return it to the body, as these could run into trouble.

International patents

World map by Japinderum (CC BY-SA 3.0) license
PCT international patent treaty countries

Applying for international patents: Most countries maintain their own separate patent and trademark offices, and typically (Europe being a partial exception), patents operate within national borders.  Thus a US patent, for example, covers just the US.

The Paris Convention: The process of extending a patent application to cover multiple countries was worked out by several international treaties. One important treaty, which is still in effect, was the 1880’s era Paris Convention. This established that, for example, a foreign patent office will recognize the priority of a US utility patent application (or provisional patent application) for up to 12 months after its initial filing date.

This was a good start, but it often takes more than 12 months to determine if a new invention is commercially important. It is very expensive to file patent applications in many different countries. So for many applicants, 12 months just wasn’t long enough.

The Patent Cooperation Treaty: To extend this time, in the 1970’s the Patent Cooperation Treaty (PCT) was enacted. This treaty is a bit complicated. Think of it as a temporary “holding tank” for patent applications. The key idea is that if you file a PCT application within 12 months of the filing date of the original patent application, you can extend your ability to then file international patents to 30 months from the filing date of the original patent application.

After 30 months from the filing date of the original patent application, however, you then need to make some tough decisions. It typically costs thousands of dollars to file (i.e. enter the national phase) in each separate country. So unless the invention is really very promising, often applicants either continue on only in the US, or else only file nationally in a few select countries.

The Patent Prosecution Highway: Typically, each local (national) patent office will then want to examine the patent application itself. This is both expensive and duplicative of resources. To help streamline the process, in the late 2000’s, various Patent Prosecution Highway (PPH) agreements were initiated, typically on a bilateral (country-to country) basis. Here, if a patent application has claims allowed by one (PPH) country’s patent office, the applicant can petition to have these claims accepted by other participating PPH offices. The US participates in the PPH, as do a number of other countries. The PPH is great when it works, but it is presently a patchwork of agreements rather than a comprehensive international treaty. So view the PPH as being somewhat in the “pilot” or “beta” stage.

Patent non-publication requests

Keeping confidential

Non-publication requests: Patent non-publication requests can be an important part of your IP strategy. Under US law, patent applications can be filed with patent non-publication requests. When this request is made, the USPTO will hold a patent application secret (i.e. will not publish it) until when and if the patent finally issues. Otherwise, absent this request, the patent office will automatically publish the patent application 18 months after the initial filing date.

When to consider filing a non-publication request: Non-publication requests are particularly appropriate for certain types of software patent applications, such as “business methods” and related software patents. This is because the US patent law for this type of software is presently in an unsettled state, and the international acceptance of this type of patent is also more limited. By filing with a non-publication request, your disclosure remains a trade secret. If the USPTO grants you a patent that is broad enough to be worth disclosing your work; great. If not, then you can continue to elect to keep your work hidden from the public.

When filing a non-publication request may not be appropriate: If you have plans to file outside the US, then by international treaty, you must file for international patents within 12 months of your initial filing date, and also allow your initial application to be published within 18 months. Here you can either not include a non-publication request on initial filing, or alternatively send in another form rescinding your non-publication request.

Other considerations: Unless a specific reason for non-publication can be identified, I generally recommend filing using the default, “publication” mode. This is because published patent applications can be useful. They help provide published prior art to help establish priority over patent filings from competitors. Additionally, published patent applications look impressive to investors, and can help give you more credibility.

Software patents in China, Europe, and the US

Software patents in China, Europe, and the US

Summary: In general, software patents that deal with technical effects, physical parameters, and improving computer function tend to be favored. By contrast, software patents focused on business methods tend to be disfavored.

China: Before April 2017, China restricted software patents to methods that solve technical problems, control internal or external processes according to the laws of nature, and produce technical effects in accordance with the laws of nature.  Machine readable medium type claims, and game patents, were not allowed. However these guidelines are now somewhat more expansive. For example, business methods that also have a technical aspect may now also be eligible for patent protection.

Europe: Software patents should have a “technical effect” (e.g. controlling industrial aspects, improvements in computer technology, data pertains to physical properties,). Pure methods directed “only” to doing business and computer programs without “technical effect” are not allowed.

United States: The rules are presently incoherent. Traditionally (before 2014-2015) US policy was permissive. Indeed the present written (i.e. statutory) US legal standard is that all types of software patents are permitted. However the latest judicial rulings and USPTO policies are that at least some software patents may represent “abstract ideas” run on “generic computers” that are not patentable unless there is “something more”. Unfortunately the terms “abstract”, “generic computer” and “something more” are effectively undefined (and thus can vary according to the whims of the judge or patent examiner at hand).

How will US software patent rules evolve in the future? This is a bit like reading tea leaves, but the latest US software patent decisions appear to be trending more towards a Chinese or European-like approach. On average, “Business method” software is increasingly being rejected as being “abstract”. However software that exhibits technical effects, operates on data pertaining to physical properties, or improves computer function is often accepted on the basis that it is not “abstract”, not run on “generic computers” or contains “something more”.  Arguably the US, presently lacking coherent rules, may be at least temporarily filling the void by borrowing rules from other countries.  The constitutionality of this approach remains to be determined.

About Silicon Valley and the San Francisco Bay Area

Golden Gate
Golden Gate Bridge, San Francisco

Silicon Valley and the San Francisco Bay Area have a worldwide reputation for innovation. But is there any subjective evidence for this? What do the patent statistics say?

According to the USPTO statistics, at, as of 2013, Silicon Valley “classic”, characterized as San Jose, Sunnyvale, and Santa Clara (Metropolitan Statistical Area 141940), led the field in patents with 12,899 patents granted. By contrast, the remainder of the San Francisco Bay Area minus Silicon Valley, characterized as San Francisco, Oakland, and Fremont (Metropolitan Statistical Area 141860) came in a respectable second at 8,721 patents.

Here the San Jose, Sunnyvale, and Santa Clara area includes other cities such as Campbell, Cupertino, Los Altos, Los Gatos, Milpitas, Monte Sereno, Morgan Hill, Mountain View, Palo Alto, San Jose, Santa Clara, Saratoga, and Sunnyvale.

By contrast, the San Francisco, Oakland, and Fremont area includes other cities such as Belmont, Burlingame, Emeryville, Foster City, Fremont, Menlo Park, Millbrae, Newark, Oakland, Redwood City, San Bruno, San Carlos, San Francisco, San Mateo, South San Francisco, and Union City.

Combining the two, the San Francisco Bay Area as a whole dominates the rest of the country, at an impressive 21,620 patents granted in 2013. By contrast, the next runner-up, the New York-New Jersey area, comes in at 7,886 patents. The Los Angeles area is close behind at 6,271 patents, followed by the Boston area at 5,610 patents. So from a patent perspective, yes the San Francisco Bay area is, in fact, pretty unique.

Federal Circuit rules against PTAB “Chewbacca defenses”


One thing that I will never forget about my experience with European patent oppositions is that to American eyes, the European process appears to be rather “due process of law” impaired.

For example, although in theory, issues should be argued in advance by written briefs, the European opposition process also allows parties to introduce new issues during the last minute oral arguments.

This allows for litigation by “unfair surprise”, rather than by reasoned arguments.  In this sort of setting, I have seen that illogical but last minute “Chewbacca defenses” can work quite well.  The clock is ticking, proceedings are going to finish in an hour, and suddenly you have to discuss entirely new and logically irrelevant issues. It is as if you suddenly have to shift gears and focus on if the Star Wars character Chewbacca lives on the planet Endor or not.

In this regard, it is refreshing to see that the Federal Circuit in Dell Inc., v. Acceleron, LLC (March 15, 2016) has confirmed that this sort of litigation by unfair surprise is unacceptable for US PTAB patent reviews. Due process wins — arguments made during oral argument must be restricted to only those arguments previously discussed in writing beforehand.

The hidden complexity of trademarks

Registered Trademark
Registered Trademark

Trademarks help protect providers of products and services from unfair competition. An example of unfair competition is when company “A” sells a product or service, and business competitor “B” sells a competing product or service in such a manner as to confuse company “A” customers into accidentally buying from company “B”.

Three general principles of trademark law can be summed up as “don’t try to trademark the normal names of a product or service”, “a trademark only has meaning in the context of a particular class of product or service”, and “use it or lose it”.

Trademarks that too closely resemble the normal name of a product or service will be rejected as being “generic“. By contrast, a trademark that is “fanciful”, — that has no obvious correlation to a product or service, will often be accepted. Trademarks falling in between these two extremes, which may be “descriptive” of the underlying product or service, often will be put on probation in a “secondary trademark register” for 5 years. After 5 years of use in trade, absent objection by outsiders, the descriptive name can then be granted full trademark status on the “primary trademark register“.

Note that the word “Trademark” has the word “trade”. This begs the question, trade in what? Under trademark law, products and services are shoehorned into a limited number of different trademark classes. Although recycling the same names within a class is generally not tolerated, the same names can often be reused for alternate products or services between classes, so long as there is not confusion. The USPTO generally examines trademarks on a per class basis.

Trademarks are granted on a “use it or lose it” basis. Although it is possible to reserve trademarks (at least fanciful trademarks) for up to a few years on an “intent to use” basis, ultimately to get the trademark, you need to provide proof that the mark is actually being used in commerce by submitting a “specimen”. This is often a photograph of sales receipts, website purchase pages, and the like, showing that the trademark name is actually being used in real trade.

There are many other complexities of trademark law as well. Searches can be tricky, because the underlying legal standard for trademark infringement is “possibility of confusion”, and not “the same spelling“. Thus merely avoiding other trademarks with identical spelling is not enough. Alternative spellings, word prefixes, word suffixes, and other name variations can often cause problems if potential customers might be confused.

Why patent searches and prior art searches are a good idea

Thinking like a detective

Prior art searches: Before investing a lot of time, effort, and money in trying to patent a potential new invention, it is always good to spend at least a few hours doing patent searches to see if there is any prior art (earlier patent or non-patent publications) that might cause problems. As a rule of thumb, if you haven’t found any prior art that concerns you at least a bit, then you probably haven’t looked enough yet!

Why do this? Because eventually, skeptical USPTO examiners will be going through your patent application, looking for ways to reject on the basis that the application is either not novel or obvious in view of prior art.

Patent examination is an adversarial process. Patent examiners have, in effect, quotas which require them to initially reject most patent applications. It is a bit of a game, like tennis, where the examiner initially rejects by citing some combination of prior art.  The examiner then hits the ball back onto the applicant’s side of the court. The examiner is expecting the applicant to hit the ball back onto the examiner’s side of the court by submitting convincing arguments pointing out where the examiner was wrong. If the applicant can convincingly argue that the examiner is wrong, usually (often after a few additional volleys), the examiner allows the patent, and the applicant wins. But if the applicant cannot argue back convincingly, it is like dropping the ball or hitting the ball back out of bounds. The applicant loses, and that patent application is going nowhere.

The patent examiner typically starts by doing computer searches for prior art using search terms that the examiner thinks will work best. This computer search will come up with unexpected citations, which the examiner may in turn combine in unexpected ways to try to show that the invention is “obvious”. Because of the quota to initially reject almost everything, often these initial rejections are unreasonable. Unreasonable rejections are “good news” in disguise, because the applicant can then rebut these by reasonable arguments. It is the reasonable rejections that are the real “bad news”, because these are harder to argue against. So before putting a lot of time and money into an invention, it is a good idea to do at least some initial checking to see what the examiner (or for that matter, a competitor) might try to use against you later. This is the idea behind a “feasibility study” type prior art search.

Sometimes, particularly when the inventor or inventors are experts, the inventors themselves may decide that they are sufficiently familiar with their field so that the search-time expense/benefit ratio is low, and no additional prior art searches are worthwhile. Here, since the USPTO will do an independent search anyway, the only obligation that the inventors have is to inform USPTO of the most important prior art that they are aware of.

Other times, particularly when the inventor or inventors are working in an area that is new to them, prior art searches have a better search-time expense/benefit ratio. Here many good sources are available over the Internet, including Google, Google patents, Google Scholar, the USPTO patent database, the WIPO patent database, and the like.

Some common mistakes that inventors make is to underestimate the amount of prior art that a patent examiner can use against their invention. Although pretty much everyone understands the problem of prior patent applications by others, this is only part of the prior art problem.

For example, for the US, anything published more than a year ago can be used against your invention, including your own publications. This includes that “cool” YouTube video, that “cool” social media posting, or that research paper/thesis that you may have published last year. Foreign patent offices are even worse in this regard. Patent examiners know that often the inventor’s own worst enemy is the inventor’s own earlier publications. As a result, patent examiners often specifically search for the inventor’s own publications and internet postings to use against them.

Another problem that inventors sometimes have is that they are a bit too emotionally attached to their cool new idea.  This problem can result in sub-optimal prior art searches. The inventor may, perhaps subconsciously, avoid search strategies that may kill that shiny new idea.

Here, there are several advantages to having a patent attorney do a quick feasibility prior art search for you. One is that the patent attorney does prior art searches all the time. Another is that the patent attorney, being both less emotionally invested in the idea, and being expert in finding prior art, may have an easier time thinking of thorough search strategies. The third is that often doing the search helps the patent attorney understand variants of your idea better, which can translate into a better patent application.

For example, if something highly relevant comes up (and often something does come up), then the new idea can be better evaluated in the context of this new-found prior art. The scope of the initial idea may be altered or expanded, and a better patent application can result. Worst case, if it becomes clear that the idea has been done before, then time and money can be saved. Alternative ideas can be explored instead.

In conclusion, “feasibility study” type prior art searches are a good a way to get at least a rough estimate of the legal feasibility of a new idea. Although the prior art search may uncover what looks at a first glance to be a tough obstacle, often this initial obstacle may encourage further thinking and improvements to the initial idea, resulting in a better patent down the road.

Note that in addition to “feasibility study” type prior art searches, there are other types of patent searches called “freedom to operate” searches or opinions as well. These represent a different and more extensive type of search, and this is a different topic.