Trade Secret vs. Patent Protection for Software Startups

In light of recent changes in the law concerning software patents, software startups should more heavily consider trade secret protection to protect their technological and operational advantages.

In light of recent changes in the law concerning software patents, software startups should more heavily consider trade secret protection to protect their technological and operational advantages.

So you’ve launched a software startup. You know that if you are to have any hope of becoming the next Google, Facebook, or Twitter success story, you need to take steps to protect your technology and your brand. Making the choice to invest some resources into obtaining intellectual property protection can be a great decision, but the specifics of how to carry out such a plan might be a little elusive. The question is what kind of protection do you actually need, and how do you go about getting it.

Two types of IP protection that most startups will want to consider are patents and trade secrets. What works best will vary from startup to startup depending on each ones own unique circumstances. However, there are several factors that all startups will want to consider when implanting an intellectual property protection strategy. First, let’s consider the basic types of IP protection that the patents and trade secrets provide, and then we can better understand factors that might encourage you to choose one or the other.


One type of IP protection that most people have heard of but few people fully understand is patent protection. There are several categories of patents, including utility patents, design patents, and plant patents. Plant patents and design patents give inventors rights in specific areas as defined by statute. However, the most common type of patent is the utility patent.

Utility patents grant the holder an exclusive monopoly on inventions of “new and useful process, machine, manufacture, or composition of matter, or a new and useful improvement thereof.” For example, someone could obtain a patent on a new and improved hair trimmer as long as it was different than and not obvious in light of previous hair trimmers. Last year, over 500,000 utility patent applications were filed at the U.S. Patent & Trademark Office.

Patents, utility patents in particular, enjoy popularity for several reasons. First, they grant the rights holders a very strong form of protection. This protection – the right to prevent others from making, using, selling, offering for sale, or importing the patent holder’s invention – is given in exchange for a publication that explains how to make and use the invention. After twenty years anyone is allowed to make and use the invention. When considering whether or not to apply for a patent, a software startup should evaluate whether or not the exclusive protection is worth having in exchange for giving up their invention after twenty years. In many instances, this will be a good tradeoff to make given the rapidly evolving world of software and technology.

Another benefit of utility patents is that the patent holders will not lose their rights due to the actions of third parties, assuming that they justly received the patent in the first place. This may not necessarily be true with trade secret protection. For example, a person who has trade secret protection in a manufacturing process can not prevent a third party from independently discovering and using such a manufacturing process. But if the original inventor of that manufacturing process acquired a patent, the third party could not use the process even if they independently discovered it (assuming that the patent holder had a valid patent that was not acquired through fraud).

While there are some significant upsides to obtaining a patent, there are costs as well. One downside that was already mentioned is the limited duration of patents. Another potential downside is the cost and difficulty of obtaining a patent. As discussed in our recent post, in recent years, it has become more and more difficult for inventors to obtain patents on software, and the scope of granted software patents may be more limited. Getting a patent can also be quite expensive. Depending on how many jurisdictions (i.e., countries) your startup is applying for patents in, the price tag can reach into the hundreds of thousands of dollars. See our prior post outlining the costs of obtaining patent protection.

Trade Secrets

Unlike patents, trade secrets are creatures of state law rather than federal law. This means that specific trade secret regimes will vary some from state to state; however, there are generally accepted principles regarding how trade secret law should work. One attempt to compile some of these generally accepted principles is the Uniform Trade Secrets Act. This document defines trade secrets as follows: “’Trade secret’ means information, including a formula, pattern, compilation, program, device, method, technique, or process, that: (i) derives independent economic value, actual or potential, from not being generally known to, and not being readily ascertainable by proper means by, other persons who can obtain economic value from its disclosure or use, and (ii) is the subject of efforts that are reasonable under the circumstances to maintain its secrecy.” In short, a trade secret is information that is valuable because it is kept secret. State laws prevent the “misappropriation” of trade secrets. This basically means that people aren’t allowed to steal your secret sauce. They can, however, independently create or reverse engineer it. This is one of the main disadvantages to trade secrets.

There are several benefits to weigh this disadvantage against when considering whether or not to rely upon trade secrets as your chosen form of IP protection. One benefit is that trade secrets can be relatively cheap compared to patents. There are no government filing fees for trade secrets, and the main costs result from taking measures to make sure your secret information remains secret. Trade secrets also can last forever assuming the information remains secret. This can be advantageous if you believe your invention will have commercial value for a long time to come.

Choosing Trade Secrets or Patents

One form of protection or the other will not be right for every startup. The following are some factors to consider when choosing between the two: (1) Will the invention be useful beyond 20 years? (2) Is it possible for companies to reverse engineer the invention? (3) Is the invention likely to be discovered independently in the near future? (4) Can you afford more expensive patent protection given your goals? (5) What is the risk that competitors design around a patent? (6) Are you interested in licensing/cross-licensing with competitors? (7) Would you be able to detect if someone was infringing your patent?

Let’s look at a couple of examples in order to understand how these factors can play out.

Example 1:

A software startup has developed a technology that has never been seen before, but as soon as you sell it, everyone will understand how to make it. They believe that this technology will be relevant for 5-10 years, but are not sure beyond that. Finally, the startup thinks there a lot of potential competitors who will likely be interested using this technology.

In this example, the startup would likely want to obtain a patent because all the factors lean in that direction. It will be easy for other competitors to develop similar products once they see the startup’s new technology; this will greatly reduce if not eliminate the value of any trade secret protection. The indefinite duration of trade secrets wouldn’t really be that valuable even if competitors couldn’t reverse engineer the product. The technology will only be relevant for 5-10 years, well less than the time frame protected by patents. Finally, because there are a lot of potential competitors that might be interested in using the technology, the startup could potentially operate on licensing business model to obtain revenue or cross-licenses.

Example 2:

A software startup has made a discovery that they believe is so foundational in nature that they believe it will revolutionize the industry for many years to come. They’re a little low on cash right now, but they expect that to turn around in the next year or two. Finally, the discovery should improve the ease with which current products are made, but the final products themselves won’t necessarily be changed.

In this example, the startup will likely choose to rely on trade secret protection. Given the foundational nature of the discovery (think F=MA2 rather than a bulldozer applying the principal of F=MA2), it is not even clear that the discovery would be patentable.  Further, the startup does not have a lot of money to try and convince the USPTO that the discovery is in fact patentable. Finally, it looks like the manufacturing processes for widgets will change, but this will not be detectable in the final widgets themselves. Thus, it might be very difficult to tell if a competitor was infringing a patent. Relying on trade secret protection will mitigate these concerns and allow the startup to profit for the discovery for many years to come.

It is important to note however, that deciding between a patent and trade secret does not necessarily have to be an either or proposition. It might be possible to go the trade secret route at first, and then obtain a patent further down the road. However, this is a one way street. Once you obtain a patent, you can never go back and rely on trade secret protection. Also, you can split up various aspects of your business. Certain aspects of the business might be appropriately covered by patents while other aspects would be great candidates for trade secret protection.


Startups creating an IP protection strategy should make decisions in light of their business model and technology. If it is not a clear decision to go one way or the other, the startup will need to make a calculated decision as to what factors are more important to them. Finally, when appropriate, startups should consider using patents and trade secrets in a hybrid form of protection.


Biotech Ventures Beware! The Industry May Need a Different Kind of Cure.

Last year's Supreme Court decision in Myriad and the USPTO's subsequent change to its examination guidelines make it more difficult to protect certain biotech innovations

Last year’s Supreme Court decision in Myriad and the USPTO’s subsequent change to its examination guidelines make it more difficult to protect certain biotech innovations


On June 13, 2013 the Supreme Court handed down its decision in Association for Molecular Pathology v. Myriad Genetics, Inc.. In this landmark decision for life science patents, the court held that a “naturally occurring DNA segment is a product of nature and not patent eligible merely because it has been isolated.” In so holding, the court weaved a fine line between unpatentable subject matter – the long-standing rule that “laws of nature, natural phenomenon, and abstract ideas are not patentable,” as they are not any single inventor’s creation, but rather are the “tools of scientific and technological work,” – and patentable subject matter. The court indicated that “the rule against patents on naturally occurring things is not without limits, for all inventions at some level embody, use, reflect, rest upon, or apply laws of nature, natural phenomenon, or abstract ideas, and too broad an interpretation of this exclusionary principle could eviscerate patent law.”  After Myriad, practitioners were uncertain as to whether Myriad specifically made genomic DNA unpatentable or whether other “naturally occurring” entities were also unpatentable.

On March 4, 2014 the United States Patent and Trademark Office (U.S.P.T.O), the regulator of U.S. patent grants, issued guidelines determined to shed clarity on this Myriad ambiguity. In these guidelines, the Deputy Commissioner for Patent Examination Policy identifies 12 factors that must be considered by a patent examiner in determining whether the invention is “naturally occurring.” Some of the factors that weigh in favor of eligibility are:

a)  Claim is a product claim reciting something that initially appears to be a natural product, but after analysis is determined to be non-naturally occurring and markedly different in structure from naturally occurring products.

b)  Claim recites elements/steps in addition to the judicial exception(s) that impose meaningful limits on claim scope, i.e., the elements/steps narrow the scope of the claim so that others are not substantially foreclosed from using the judicial exception(s).

c)  Claim recites elements/steps in addition to the judicial exception(s) that relate to the judicial exception in a significant way, i.e., the elements/steps are more than nominally, insignificantly, or tangentially related to the judicial exception(s).[7]

Some of the factors weighing against patentability are:

g)  Claim is a product claim reciting something that appears to be a natural product that is not markedly different in structure from naturally occurring products.

h)  Claim recites elements/steps in addition to the judicial exception(s) at a high level of generality such that substantially all practical applications of the judicial exception(s) are covered.

i)  Claim recites elements/steps in addition to the judicial exception(s) that must be used/taken by others to apply the judicial exception(s).[8]

The large number of factors and their inherent ambiguity have resulted in unpredictable examinations. Practitioners have been surprised to receive 35 U.S.C. § 101 rejections, when they were expecting patent issuance notices for molecular biology and non-molecular biology based inventions alike. Patent examiners have thus been effectively reading Myriad’s prohibition on “naturally occurring” as not limited solely to DNA-based inventions. As pointed out in a recent article, what’s more is that products that are derived from a naturally occurring product might be considered to be naturally occurring as well, further stifling patent protection for life science inventions.

The result has been a siege of the biotech and pharmaceutical industries. For an industry so reliant on patent protection, that is fixed with high upfront research and development costs, patent ineligibility is a real scare. If these guidelines stand, the industry can no longer be assured that their R&D expenses will pay off down the road.

However, the Myriad-Mayo Guidelines have themselves also been under threat from those companies and practitioners most affected by them. July of this year marked the end of the public commenting period, in which biotech and pharmaceutical companies as well as universities voiced their consternations with the Myriad-Mayo Guidelines. The vast majority of the comments were extremely critical of the guidelines, incentivizing the USPTO to revise the Myriad-Mayo Guidelines. The revision is currently underway and it will be important for life science based startup companies to watch for what occurs at the USPTO in the immediate future. Will the USPTO cure the patent illness it began? Patent protection for your start-up may be at stake




Fallout of the Supreme Court’s Recent Alice Decision: Is Software Still Eligible for Patent Protection? (Part 2 of 2)

In Part 1 of this series, we examined the Supreme Court’s recent decision in Alice Corp. v. CLS Bank and how courts and the Patent Office have started to limit the scope of software patents as a result. In this second and final post, we’ll look at how an entrepreneur in the software space might be able to protect her intellectual property in light of this new guidance.

Software Implemented Inventions Can Still be Patented

Despite what some anti-patent advocates may have hoped, Alice did not invalidate all software patents, and software inventions can still be patented. In fact, the Alice decision, at most, delineated a boundary beyond which a computer-implemented idea becomes too abstract to be patented, without something more. Nor did the court hold that software inventions necessarily constitute abstract ideas and are therefore ineligible for patenting. In fact, nowhere in the court’s opinion is the word “software” even used. Without a doubt, many inventions that embody systems and methods implemented in software will continue to be awarded by the Patent Office and upheld by the court system.

The Scope of Patentable Software is More Limited

With the good news comes the bad for software inventors hoping for a new patent. The scope of patentable software inventions is now decidedly more limited than it was before Alice. Gone are the days where an inventor could program a generic implementation of Bingo and receive a patent for it. While the Supreme Court has not set any hard-and-fast rule on what, exactly, is considered “abstract,” we have an idea of the general shape of the boundaries: any program that simply takes a long-existing idea and generically implements it in code is too abstract to pass the Alice patent eligibility test. If you take your computer Bingo program, and mentally strip it down to its essential functionality, is it really just a Bingo game on a computer, or is it something more? Perhaps there is additional functionality that make it more than just Bingo. Maybe you’ve programmed some unique features within the code that take advantage of the hardware on a specific platform and make the program run more smoothly. If you have these sorts of features built-in to your program, even computer Bingo may be patentable under the second step of the Alice test.

 Carefully Consider Your Options Before Filing a Software Patent Application

Finally, as an entrepreneur, you should think very carefully about identifying the particular and novel aspects of your software before filing a patent application. Your attorney can help you narrow and identify these features, but in general, the following considerations are prudent:

  • Think about what, specifically, makes your software unique. Does it leverage a unique algorithm to create performance improvements? Is there a function that has never been accomplished before? Does the user interface create a new experience? If the novel aspects of your software currently exist, especially if they exist in a non-digital format (like the game of Bingo), is there something “extra” that makes the invention more than just converting it into a digital format? For instance, implementing Bingo on a generic computer is not patentable, but is there something really unique about the user interface that makes the Bingo card easier to interact with? Does the program take advantage of the memory architecture in a unique way to more efficiently store and retrieve data? Try to identify these features in as much particularity as
    possible, even before you sit down to develop the patent application language.
Under the new patent eligibility guidelines for software implemented inventions, patents must claim something more specific than functionality carried out by well-known programming techniques on general purpose computers.

Under the new patent eligibility guidelines for software implemented inventions, patents must claim something more specific than functionality carried out by well-known programming techniques on general purpose computers.

  • When you and your attorney do begin drafting the patent application, consider identifying the novel algorithmic aspects of the software and including flow charts or even representative source code or pseudo-code in the filing. This will help to demonstrate to the examiner that your code is more than just implementing an idea on a generic computer, which can be accomplished in a near-infinite number of ways. Including specifics about the algorithm (or several examples of a possible algorithm) in the application will give you the flexibility during the Patent Office review to narrow the scope of your claims and could potentially save the application from dismissal under an Alice rationale. Since the legal ramifications of Alice have yet to fully play out, this is an attractive way to preserve some (albeit narrow) patentable ground if the courts or the Patent Office further restrict software-enabled inventions.
  • Consider your overall business goals and, perhaps, reconsider the need to pursue a strategy of patenting. You may not need a patent to protect your competitive advantage. Implementing a strategy of keeping software a trade secret may be a viable alternative in some situations, and copyright generally protects unauthorized copying of software. The expense and hassle of obtaining a patent may not be worth the ultimate reward.


Fallout of the Supreme Court’s Recent Alice Decision: Is Software Still Eligible for Patent Protection? (Part 1 of 2)

In June 2014, the Supreme Court struck down a patent on a software implemented invention, perhaps narrowing the scope of software patents we will see in the future.

In June 2014, the Supreme Court struck down a patent on a software implemented invention, perhaps narrowing the scope of software patents we will see in the future.

The Supreme Court, Patent Eligibility, and “Abstract Ideas”

In June 2014, the Supreme Court issued a much-anticipated patent decision in Alice Corp. v. CLS Bank. The question was whether the claims at issue, involving a method for mitigating “settlement risk” between two parties in a financial exchange by using a computer system as an intermediary, are patent-eligible under 35 U.S.C. § 101. Justice Thomas, writing for a unanimous court, answered that question with a definitive “no.” But even though this particular computer-implemented method was struck down, are other software programs still eligible for patent protection? That’s a trickier question.

35 U.S.C. § 101 limits patentable subject matter to “any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof.” Laws & products of nature, physical phenomena, and abstract ideas are generally considered unpatentable subject matter. With the rise of software and software-based inventions, the line between abstract ideas and new and useful processes became somewhat unclear. Note that §101 addresses whether a particle technology is eligible for patent protection. Section 101 does not address whether an invention is new, nonobvious, useful, or sufficiently described. Those requirements are addressed in separate sections of the Patent Act and represent their own unique hurdles to patentability. Therefore, the question presented by §101 (and addressed in the Alice case) is, even assuming a particular invention is new, nonobvious, useful, and sufficiently described, is this the type of creation that should be considered for the powerful protection of a patent.

For a time, courts generally refused to invalidate software patent claims under the statutory language of § 101. In 2010, however, the Supreme Court unanimously held, in Bilski v. Kappos, that a patent claiming a software method for financial hedging against price fluctuations was invalid under § 101. Alice continued this line of decisions for the Court, confirming the Supreme Court’s position that “basic tools of scientific and technological work” should not be monopolized by a patent-holder. Alice communicated to lower courts that § 101 should be used to invalidate software and business methods that are so abstract that the award of a patent would “tend to impede innovation more than it would tend to promote it.”

In order to accomplish this goal, Alice puts forth a two-part framework, based on its previous decisions in Bilski and Mayo v. Prometheus, to determine whether a patent application describes an “abstract idea,” and if so, whether it still contains enough inventive material to meet the § 101 standard.

First, the court considers whether the patent claim is “directed to an abstract idea.” While on some level, every invention embodies at least one abstract idea, if the claimed idea is so broad such that awarding a patent would impede more innovation than promote it, it is likely to meet this first step. Examples of abstract ideas include fundamental economic practices like the ones at issue in Alice and Bilski, certain methods of organizing human activities, an idea of itself, and mathematical relationships/formulas.

Second, if the patent claim is directed to an abstract idea, the court examines whether anything in the patent “transform(s) that abstract idea into a patent-eligible invention. In other words, if there are features of the invention that go above and beyond simply implementing an abstract idea in software, it may still be patent-eligible. Improvements to the technology, improvements to the functioning of the computer elements and other “meaningful limitations” that go beyond simply implementing an idea on a computer are likely to qualify an invention in this way.

 The End of Patents for Generic Ideas Implemented on a Computer

The U.S. Patent and Trademark Office (PTO) took little time in heeding the new guidance from the Supreme Court. Less than a week after the Alice decision, the PTO issued new guidance to its examiners in applying the Alice decision. PTO examiners now employ the two-part test set forth in Alice to evaluate potentially abstract ideas, and if there are no “meaningful limitations” on an otherwise abstract idea, the PTO examiner is very likely to reject the application as non-patentable subject matter under § 101.

Similarly, courts have wasted no time in implementing the Supreme Court’s guidance. The U.S. Court of Appeals for the Federal Circuit, which receives all appeals on patent matters, has invoked Alice in a number of recent cases to invalidate software and business method patents. District Courts hearing patent disputes, too, have seized the opportunity to invalidate claims that are too “abstract” to be patented. Some recent examples of inventions too “abstract” to be protected by a patent under the Alice rationale include:

  • “[A] process of organizing information through mathematical correlations and is not tied to a specific structure or machine . . . taking two data sets and combining them into a single data set” Digitech Image Techs., LLC v. Elecs. for Imaging, Inc., 758 F.3d 1344 (Fed. Cir. 2014)
  • “[C]reating a contractual relationship—a ‘transaction performance guaranty’—that is beyond question of ancient lineage,” implemented by a generic computer. buySAFE, Inc. v. Google, Inc., 2014 U.S. App. LEXIS 16987 (Fed. Cir. Sept. 3, 2014)
  • “[M]ethods and systems for ‘managing a game of Bingo,’ . . . ‘solv[ing a] tampering problem and also minimiz[ing] other security risks’ during bingo ticket purchases” using a computer. Planet Bingo, LLC v. VKGS LLC, 2014 U.S. App. LEXIS 16412 (Fed. Cir. Aug. 26, 2014)
  • “[A] computer program that allows the user to create meals from a database of food objects according to his or her preferences and dietary goals, to change those meals by adding or subtracting food objects, and to view the dietary impact of changes to those meals on a visual display” DietGoal Innovations LLC v. Bravo Media LLC (Div. of NBC Universal Media, LLC), 2014 U.S. Dist. LEXIS 92484 (S.D.N.Y. July 8, 2014)
  • “[A]sking someone whether they want to perform a task, and if they do, waiting for them to complete it, and if they do not, asking someone else. . . . performed ‘in connection with a computer system.’” Eclipse IP LLC v. McKinley Equip. Corp., 2014 U.S. Dist. LEXIS 125529 (C.D. Cal. Sept. 4, 2014).

So what does all of this mean for an entrepreneur with a potential software invention? Check out Part 2 of this series.


FTC Sues Amazon: Lessons for Startups Seeking to Monetize Through In-App Purchases

FTC Complaint

As has been widely reported, on July 10, the Federal Trade Commission filed suit against for allegedly failing to do enough to prevent unintentional in-app purchases by children.

Background on Consumer Protection by FTC

In its complaint, the FTC accuses Amazon of violating Section 5 of the FTC Act, which provides:

Unfair or deceptive acts or practices in or affecting commerce [ ] are [ ] unlawful.

The FTC Act defines “unfair” practices as acts that:

“cause[] or [are] likely to cause substantial injury to consumers which is not reasonably avoidable by consumers themselves and not outweighed by countervailing benefits to consumers or to competition”

The FTC’s website refers to Section 5 as the FTC’s basic consumer protection statute.

The FTC’s recent actions with Apple, Google, and Amazon, show the FTC is primarily concerned with the following aspects of company’s seeking to monetize through in-app purchases when children may be using the app.

(1) making sure account holders understand the nature of in-app purchase and the full scope of what an account holder is agreeing to when they download an app and enable in-app purchases;

(2) technological restrictions imposed at the time of an in-app purchase to reasonably prevent a child from making a purchase the account holder does not intend to make (e.g., a second entry of a password, or having to enter credit card information); and

(3) the notice provided to the account holder when and in-app purchase is made.

FTC’s Complaint Against Amazon

According to the FTC’s Complaint,, billed parents and other account holders for children’s activities in apps that are likely to be used by children without having obtained the account holders’ express informed consent.

Some commentators believe the “express informed consent” standard is too rigid and deprives an app designer of the ability to make its own design choices in deciding how to fairly engage with consumers.

The FTC’s “Express Informed Consent” Standard

Some commentators believe the “express informed consent” standard is too rigid and deprives an app designer of the ability to make its own design choices in deciding how to fairly engage with consumers.

In its January settlement with Apple, the FTC defined “express informed consent” as requiring an affirmative act communicating informed authorization of in-app charges that:

(1) is made promptly in relation to in-app activity;

(2) discloses material information related to the billing, including:

  • for a specific in-app purchase: the activity related to the charges; the specific amount of the charge; and the account to be filled; and
  • for potential future in-app purchases: the scope of the charges (including duration and apps to which consent applies); the account to be billed; methods by which an account holder may revoke or modify the scope of consent

(3) takes reasonable steps to verify that the person providing consent is the account holder;

(4) for consent related to potential future purchases, consent must be obtained at least once for each mobile device.

Current Lessons for Startups Related to In-App Purchases

The following are take-aways from the FTC’s recent activity related to in-app purchase:

1) Stayed tuned and see if Amazon prevails.  Amazon believes the FTC’s requirements are too stringent.  If federal courts agree, then the FTC will have to revise its policies concerning in-app purchases.

2) Use your “Terms of Use” to provide clear notice to account holders about what in-app purchases are available and (if applicable) that such purchases can be made without re-entering credit card information.  A clear warning to parents should also be provided if the nature of the app is such that children are likely to use it.

3) Reasonable restrictions on in-app purchases should be imposed.  The restrictions must be balanced with reasonable design and business considerations.  Possible restrictions include: (i) re-entering a password in order to make an in-app purchase; (ii) entering a different password; (iii) having to re-enter at least a portion of credit card information;  or (iv) asking identifying questions or using recognition technology on phones.

4) Provide prompt notice to an account holder of an in-app purchase and/or provide some ability for account holders to stop or reverse a purchase.

5) If you do not intend for children to us the app, take steps to prevent them from doing so, and make this intent clear in your terms of use.




The Rift Between Companies: Oculus v. Zenimax

ZeniMax Media and Oculus are embroiled in a dispute over the intellectual property rights incorporated in the Rift product.

ZeniMax Media and Oculus are embroiled in a dispute over the intellectual property rights incorporated in the Rift product.

Have you heard of the Oculus Rift (that loveable virtual reality company swept off its feet by Facebook)? How about legendary game designer John Carmack? If you haven’t by now, you absolutely should have! In any case, here’s your major takeaway: Beware the Employer! The situation playing out between Oculus and ZeniMax Media is an interesting one, and it is a useful lesson in what not to do. While we wait for this to unfold (either in the courtroom or behind closed doors), let’s see where things went wrong and learn from Oculus’s potential mistakes.

Our Legal Setup

On May 1, 2014, ZeniMax Media (the parent company of Carmack’s former employer, id Software), sent letters to Oculus and Facebook claiming rights in at least part of the IP in the headset. In other words, ZeniMax believes it owns rights in IP created by Carmack during the course of his prior work for ZeniMax’s subsidiary, and that some of that IP is embedded in the Oculus Rift headset.  On May 21, ZeniMax initiated a lawsuit against Oculus in the Northern District of Texas. The focal point of Oculus’s problems stems from two documents: John Carmack’s employment agreement (and other ZeniMax/id personnel as well) and a Nondisclosure agreement entered into between Oculus cofounder, Luckey, and Zenimax Media.

Employment Agreements: Be Cautious about who helps you

Entrepreneurs, if you seek someone’s help, be sure to have them take a look at their employment agreement! Almost always, an employment agreement will include provisions that immediately make the company the owner of rights in inventions created by the employee in the scope of his employment. Sometimes, there may be an out if the work is unrelated or outside the scope of the Employer’s work or agreement. However, that may be difficult to prove.

Prior to this whole mess, Luckey, a hardware developer had begun the development process on the “Rift” virtual reality (VR) headset. In 2012, John Carmack began corresponding with Luckey and asked for a prototype of the Rift headset to tinker with. This is where their trouble begins. In the complaint, ZeniMax constantly states that it was developing VR technology with Carmack heavily involved in that research. As a game company, ZeniMax researched VR technology specifically for game development. There exists some question as to whether or not they sought to bring a viable consumer hardware project to market.

Carmack’s specific employment agreement reads:

Employee agrees that all Inventions that (i) are developed using equipment supplies, facilities or trade secrets of id Software or the Company… or (iii) relate to the Company’s business or current or anticipated research and development will be the sole and exclusive property of the Company (ZeniMax) or its designee from the moment of their creation and fixation in tangible media…

In other words, even if Carmack did not use his employer’s resources in working on the Rift, his employment contract would still purport to transfer rights to the employer if the Rift relates to the employers “current or anticipated research and development.”  Additionally, Carmack and other ZeniMax employees are said to have used company resources and research to make both hardware and software (including the software development kit or SDK) improvements to the device.

According to ZeniMax, “Carmack made breakthrough modifications to the Rift prototype based upon years of prior research at ZeniMax.” When Carmack and his crew made improvements, the resulting intellectual property had been assigned to ZeniMax via the employment agreements. The complaint constantly tees up that ZeniMax (Carmack with id Software in particular) had been working on VR hardware and software research. However, record exists where Carmack states that this was his pet project.  [Is there something missing in the prior sentence?]

In any case, if the court sides with ZeniMax, these improvements become their property, including the SDK. Carmack had quickly taken to Twitter to protest saying that nothing he has ever worked on was patented. But that might not be the only issue. The code he wrote, which is automatically protected by copyright, became property of ZeniMax (under ZeniMax’s interpretation of the law and facts). Accordingly, it might be the case that ZeniMax owns certain parts of the code used in the Rift.  If ZeniMax was the owner of Carmack’s rights in code he developed, then Occulus would have had to reverse engineer its code, without incorporating any of the copyrighted content from the ZeniMax-owned code.  It might also be the case that Luckey and ZeniMax became co-owners in the software.

Nondisclosure Agreement: Now that it’s not yours, it’s not yours.

Under ZeniMax’s argument, the technology became confidential information and trade secrets of ZeniMax. This is what gives ZeniMax some ammunition (or rather a actionable claim). Their research and proprietary information that was confidential and not readily attainable had been used in the Oculus Rift. Oculus was using it to profit to the tune of a cool $2 billion from Facebook. This could constitute misappropriation, meaning ZeniMax’s proprietary information was stolen. Helping to solidify their claim, ZeniMax points to the NDA entered into between the Company and Luckey.

It was under this agreement that Carmack showed Luckey the improvements made to the Rift. This contract is what governed the relationship between Carmack’s team, ZeniMax, and Luckey. The definition of confidential information includes the work done by the employees of ZeniMax. If these facts are correct, Luckey would likely not be able to use the information without the permission of ZeniMax (for purposes other than those outlined in the NDA). If ZeniMax’s secrets are disclosed, they potentially lose any trade secret protection. Multiple times in ZeniMax’s complaint, the Oculus team is painted as having no software background and no particular VR experience. If these facts are true, this would help ZeniMax demonstrate that they did not reverse engineer the tech (an acceptable method of discovering trade secrets). Multiple email chains affirmed the notion that they needed Carmack. Only days after demonstrating the Rift improvements, Oculus LLC was formed in the state of California.


ZeniMax’s complaint does not paint a favorable picture of how Oculus handled its negotiations with ZeniMax. The only facts are those from the complaint, and we’ll have to wait for more to come to light to know the truth. It is understandable that an entrepreneur will desire to protect and profit from an idea they believe to be there own. But even if there is not a clear owner of intellectual property rights in an idea, an entrepreneur maybe able to secure clear title to important intellectual property by negotiating a business arrangement with others involved.  ZeniMax’s allegations outline what not to do in negotiating over intellectual property rights.  ZeniMax alleges the negotiations broke down in large part due to the Oculus team. According to the complaint, originally, Oculus was to grant ZeniMax some share of equity in the LLC. According to the complaint, Oculus demanded a worldwide exclusive license over the IP disclosed by ZeniMax pursuant to the NDA. Furthermore Oculus demanded marketing support from ZeniMax and 10,000 free copies of Doom 3 BFG edition (a game capable of showing off the Rift) for the Kickstarter campaign. In return, ZeniMax would receive 2% equity interest subject to a 3 year vesting schedule. Vesting would be dependent on Carmack’s ongoing involvement in the project. Oculus also proposed tht ZeniMax could pay an additional 1.2 million for another 3%. However, that term was eventually dropped.

In 2013, negotiations in total broke down. ZeniMax would eventually prohibit Carmack from working with Oculus. As a result, Carmack left the company to become the Oculus CTO. They then began poaching former members of Carmack’s team despite noncompetition provisions in his employment agreement specifically prohibiting this.  Again, this is all based on ZeniMax’s unproven allegations.  However, this provides an example of how aggressive negotiating posture and subsequent aggressive hiring decisions can inflame a situation and lead to litigation.


Clear title to fundamental intellectual property is typically critical to a startup’s success.  While Oculus was able to achieve a favorable exit, their outcome is likely the exception to the rule.  ZeniMax’s complaint in its recently filed litigation provides numerous examples of mistakes startups can avoid making in working with individuals employed by others or subject to NDAs.


LLC Formation: Why Filing Your Articles is Not Enough

While it might be simple to technically establish an LLC, several important documents are required to properly organize as a company with clear membership, governance, and IP ownership.

While merely filing Articles of Organization might be sufficient to establish an LLC, several important documents are required to properly organize as a company with clear membership, governance, and IP ownership.

Entrepreneurs are often attracted to organizing as a limited liability company (“LLC”) due to the perceived ease and cost-savings of formation.  In most states, one can (technically) establish an LLC by filing a single document and paying a small filing fee.  For example, in Michigan, one can establish an LLC by filing the Articles of Organization and paying a $50 fee.  While filing the Articles of Organization may technically establish an LLC, more is needed in order to properly organize an LLC.  This post examines the additional steps needed to properly organize a multi-member LLC.

Information Contained in Articles of Organization

The Articles of Organization are a simple one-page document that provide the following:

  • the name of the company;
  • the duration of the company, if other than perpetual (this is typically left blank for tech startups);
  • the name of the registered agent and address of the registered office;
  • the name of the “organizer” of the LLC.

Most states do not require any documentation other than the Articles of Organization (or that state’s equivalent publicly-filed document) in order to establish an LLC.

Information NOT Contained in Articles of Organization

Importantly, the Articles of Organization do not provide any of the following information:

  • who holds equity in the company;
  • how much equity any single person holds in the company;
  • who makes what decisions on behalf of the company;
  • what happens if a member leaves the company; and
  • who owns the IP created by members of the company.

At the most basic level, the absence of information about who is part of the company is particularly troubling.  If a founder does nothing more than file the Articles of Organization to set up a company, there is no conclusive document indicating who is part of the company.  Without more, individuals could point to vague oral or email statements to claim that they are entitled to some equity interest in the company.

Your State’s Default Governance Provisions May Not Be Appropriate

Most states have LLC statutes that provide default provisions for how LLC’s operate.  In the absence of additional documentation, such as an operating agreement signed by all members, these default provisions will control the operations of the LLC.  While it might be tempting to rely on these default provision rather than taking the time (and perhaps expense) to think through and establish company-specific provisions, founders should be wary of their state’s default provisions.  These default rules are unlikely to reflect exactly the way the founders intend to operate their company.  For example, in Michigan, section 450.4502 of the Michigan LLC Act provides that (unless otherwise specified in an operating agreement) each member is entitled to one vote in making company decisions.  In other words, even if interests in the LLC are divided 80/20 between two founders, they would each be entitled to one vote.

Additional Documents Needed for an LLC

The above deficiencies are why entrepreneurs should view LLC formation as requiring a suite of formation documents rather than the one-page Articles of Organization.  Specifically, a startup organizing as an LLC should use at least the following:

Operating Agreement - An Operating Agreement should be signed by all members of the Company.  Therefore, it confirms whether or not an individual is a member of the LLC, and how much equity that person holds.  The Operating Agreement should also specify at least: how decisions are made (e.g., what % of vote is required, and by whom, for the company to take certain actions); how membership interests can be transferred, if at all; and how profits/losses are allocated and/or distributed between members.  For most tech startups, it is common to create “Units” of membership interest (similar to stock in a corporation), which are established in the Operating Agreement.

Restricted Unit Agreements - If a startup seeks to impose “vesting” it is common to implement the vesting via a Restricted Unit Agreement entered into between the company and each individual member actively working with the company.  This prior post discusses the concept of vesting.  While the Operating Agreement may grant a member a certain percentage interest in the company, and the number of units that correspond to that percentage interest, a Restricted Unit Agreement will grant the company a repurchase option that lapses over time (i.e., the vesting schedule).  The Restricted Unit Agreement should be clear about specifying what action triggers the company’s repurchase option (for example, a termination of service, or a majority vote of the members of managers of the company).  Of course, whenever vesting is imposed (and the company’s repurchase option is less than the fair market value of the equity at the time of repurchase), holders of equity should pay attention to their 83(b) elections as discussed in this prior post.

Proprietary Information and Invention Agreement (“PIIA”) - PIIA’s should be signed by each individual member actively working with the company.  A PIIA assigns to the company rights in any intellectual property created by an individual during the course of their work for the company.  PIIA’s also include confidentiality obligations requiring the individual to maintain as confidential any of the company’s sensitive information.

IP Assignments -  If an individual (such as a founder) has been working on the startup prior to the company formation, then it is likely that individual holds intellectual property rights that need to be assigned to the company.  Because most PIIA’s are designed to cover intellectual property created during the course of an individual’s work for a company, they might not adequately cover pre-existing intellectual property.  Accordingly, IP assignments should also be used to cover any intellectual property created prior to company formation.

Proper Organization is Important Even for Startups Planning to Convert to a C-Corp

As discussed in this prior post on entity conversion, it is common for startups to initially organize as an LLC but later convert to a Delaware C-corp when they plan to raise capital from sophisticated investors.  For a startup contemplating this conversion, it might be tempting to forego the above organization documents, merely filing the Articles of Organization to establish the company as an LLC.  This is not wise, however.  Among other concerns, most state conversion statutes require a specified vote of the members in order to approve a conversion of an LLC into a Delaware C-corporation.  In MIchigan, for example, section 450.4708 of the Michigan LLC Act requires that all members approve a conversion unless the operating agreement provides otherwise.  Accordingly, absent an operating agreement, all members of the LLC would need to approve the conversion into a new corporate form.  However, absent an operating agreement, it is also difficult to conclusively know who is a part of the company (e.g., who the members are).  This lack of information would be ripe for an individual to later claim they were part of the LLC, but not included in the approval of the LLC to a C-corp.  Therefore, even when a startup is planning to convert from an LLC to a C-corp, they should first properly organize as an LLC so that the conversion process is clearly approved by a well-defined set of members under well-defined governance procedures.


With the above documents in place, it is now clear: which individuals are part of the company and how much they own; which members make what decisions on behalf of the company; that the company owns the intellectual property related to its business; and what happens if a member’s service for the company terminates.  None of this information would be clear if a startup merely files the Articles of Organization, and doesn’t take additional steps to properly organize itself.  Using the correct documents to properly organize will lessen the risk of potential disputes down the road, especially if the LLC later converts to a Delaware C-corp and seeks outside investment.


Let Your ELF Help You File in Michigan

By setting up a Mich-ELF account, entrepreneurs can quickly and cheaply make their organizational filings in Michigan via e-mail.

By setting up a Mich-ELF account, entrepreneurs can quickly and cheaply make organizational filings in Michigan via email.

As addressed in this prior post, entrepreneurs in Michigan typically decide between a Delaware C-Corp and a Michigan LLC for their choice of entity.  With either choice, entrepreneurs should be aware of an easy electronic filing option in Michigan available by establishing a Michigan “ELF” account.  With an ELF account, startups seeking to organize as a Michigan LLC can quickly file their organizational documents via email.  Startups organized in Delaware, but transacting business in Michigan, will need to be authorized to transact business in Michigan as a “foreign entity” and can also file their application via Michigan’s ELF service.  This post serves as a guide to setting up, and using, one’s MICH-ELF account.

Establishing a MICH-ELF Account

To set up a Michigan ELF account, one must complete and submit a Michigan ELF application ( CSCL/CD-901).

ELF Application

To complete this form, one needs the following information:

1) Name, address and phone number of the MICH-ELF applicant

2) Fax number for return of document;

3) Type of credit card (VISA, VISA Electron, and MasterCard are accepted);

4) Credit card number, expiration date, name of cardholder, and billing street address and zip;

5) Contact person, phone, and fax number if other than applicant.

The completed ELF application should be faxed to (517) 241-6445.  Note that this ELF application should not be submitted with any other documents.  Your Articles of Organization or Application for Authority to Transact Business will be submitted in a separate filing after you receive your ELF filer number.  Michigan will reject your application if you submit your ELF application with other filings.

Questions Related to the ELF Application

Although the ELF application is relatively straight forward, some questions do arise.  One common question is who to identify as the applicant.  If one is organizing a Michigan LLC, there will not yet exist any entity when one is initially setting up the ELF account.  Nonetheless, Michigan customarily will accept the future company name as the applicant name.   Identifying a founder as the applicant is also acceptable.  For foreign entities, it is likely best to identify the foreign entity name as the applicant.  If you are an attorney filing an ELF application on behalf of the client, it is best not to list yourself as the applicant.  This is because when Michigan faxes your ELF account confirmation (see below), the only identifying information on the confirmation page will be the new ELF filer number and the name of the applicant.  If the applicant name is the attorney name, and that attorney commonly files ELF applications, there will no way to tie the confirmation page to a particular client without contacting the state.

For the contact person, it is likely best to identify someone who can quickly respond to questions from Michigan.  This will likely be the person handling this initial organization — either one’s attorney or a key founder.  Note that the contact person listed on the ELF application can be different than the Resident Agent that one will identify on their Articles of Organization.  So, the contact person is not signing up for any additional responsibilities on behalf of the company, other than questions related to one’s ELF account.

Your ELF Confirmation

Michigan typically processes ELF applications in less than 48 hours.  You will receive a fax from the state, similar to the following:

ELF Confirmation Page 2

It is important to store this document (or at least the ELF Filer Number), because one will need to include the ELF Filer Number in all future filings with the state.

Making Filings Using an ELF Account

An ELF Account permits one to make organizational corporate filings via email.  Corporate filings can be emailed to or faxed to (517) 636-6437.  Note that this is a different fax number from the one used to establish the ELF account.  A MICH-ELF cover sheet (CSCL-CD900) should be used with all ELF corporate flings.  The standard MICH-ELF cover sheet is shown below.

Elf Cover Sheet

Upon making an ELF filing, Michigan will automatically charge the credit card account associated with the ELF account.  Accordingly, it is important to update one’s ELF account when a business credit card is initiated.

Note that the ELF cover sheet is the only document that should include your ELF Filing Number.  Do not include the ELF Filing Number on the Articles of Organization themselves.  A common mistake is to for a filer to place their ELF Filing Number in the unmarked box on the right side of the Articles of Organization.  If you do this, Michigan will reject your filing and require that you remove the number.  Leave this box empty.

ELF Articles of Org


The Michigan ELF account process can save entrepreneurs time and money.  Email filings are simple and lead to more timely approvals from the State of Michigan.  Setting up your ELF account (and keeping your ELF Filer Number) can ease the headaches around organization, conversions, financings, or other occasions when timely filings matter.


Visiting Your Competitor’s Web Page: Should Entrepreneurs be Concerned?

Competitor Website Pic.v3

Entrepreneurs, you can check out your competitors. But it’s likely they’re watching you watch them.

Data Collection and the State of the Internet

Data collection is an integral and expected practice of the internet. In fact, many companies, both established and new alike, rely on it for their continued existence. Data serves a critical role in either their marketing strategies or monetization strategies. Or in some cases, both. This data provides useful information essential to adjusting and tailoring a company’s business strategy. Both websites and third-party advertisers collect data.  This data includes IP addresses, location, and web history. Since the practice is unlikely to go away, the question becomes “as an entrepreneur, should I be worried about visiting my competitor’s website?” The good news is that there are some things that can be done to minimize the information a competitor learns about you from your visit to their website. This blog post will cover: 1) some of the technology involved, 2) examples from Facebook’s data usage and privacy policies, and 3) some suggestions going forward.

Technology involved

Numerous free or low-cost tools exist for a company (such as one of your competitors) to ascertain the affiliation of visitors to the company’s website.


Cookies are small text files created when a user loads a website. Every time the user returns to the site, the browser sends this file to the site’s server. Both websites and ad servers create cookies on a site. Important for our purposes, cookies inform how and when ads are shown.  You can learn more about cookies here.

Pixel tags

Pixel tags are small blocks of codes on a webpages that allow websites to read and place cookies. The resulting connection can include information such as the person’s IP address, the time the person viewed the pixel, and the type of browser used.  You can learn more about pixel tags and how Facebook uses them here.  

Web Analytics

Services such as Quantcast, Google Analytics, KISSmetrics provide extensive data. They include demographic information (age and gender), location information, time on site, and other metrics. IP addresses by themselves will give away a user’s location and network information. For instance, visit IPInfo’s free geolocation databases.  Users are greeted with IP address information on the right-side of the screen (IP address, country, region, city, and time). These services can be free or paid, and therefore, there aren’t significant barrier to entry. A basic (yet still comprehensive) version of Google Analytics is available for free.

Examples from Facebook Data Use Policy and Practices

As it has become a part of many people’s daily habits, Facebook may become a problem for the cautious entrepreneur. First, Facebook may increase your competitors screen time as a user’s information and web traffic affect advertising. Second, it may tip off your competitors as to your presence.

Facebook is rather explicit (but not entirely straightforward) in its online policies. Facebook’s privacy page states that they receive data from and/or about “the computer, mobile phone, or other devices you use to install Facebook apps or to access Facebook.” This information may include your “IP address or mobile phone number, and other information about things like your internet service, operating system, location, the type (including identifiers) of the device or browser you use, or the pages you visit.” Facebook additionally receives data whenever you visit other websites, games, and applications that use the Facebook platform or use a Facebook social plugin.

Certain information on Facebook WILL ALWAYS REMAIN PUBLIC UNLESS DELETED. For our purposes, this includes name, user ID, and networks.

Facebook’s cookie policy

Facebook may read a cookie so they can show you ads that may be interesting to you on Facebook or other websites. They also use a cookie to learn whether someone who saw an ad on Facebook later visited the advertiser’s site. Similarly, Facebook partners may use a cookie or another similar tech to determine whether they’ve shown an ad and how it performed. Information may be shared with partners.  More more information, wee Facebook’s cookie policy here. 

Facebook Page Analytics

For a Facebook page, you can see how many people your post reached, how many people clicked it and how many people clicked it, commented on it, or shared it with their friends. The Facebook Pages feature allows administrators (or rather your competitors) to identify the location of users who have seen “any content associated with [their] page.”  

What does this all mean for the entrepreneur?

The problem for entrepreneurs is two-fold. First, you may be tipping off your competitors about your presence (and your interest in your competitor). Second, you may be increasing your competitor’s web traffic.

In the first instance, as explained above, by merely visiting your competitor’s website, your competitor is likely capable of learning your location, gender, and age (inferred age group. Furthermore, your competitor might be able to identify your network.To combat this, you may consider using the TOR browser.  The TOR software prevents tracking by “bouncing your communications around.” Your internet traffic is carried over a network of different relays, which are hosted by volunteers globally. According to the TOR site, “it prevents somebody watching your Internet connection from learning what sites you visit, [and] it prevents the sites you visit from learning your physical location.”[5]

Private browsing does not anonymize your data. IP addresses and other related information may still be collected. Private browsing only prevents cookies from being stored once the browsing window has been closed. However, a private browsing session will not read cookies from other sessions.

The second situation is not nearly as avoidable. Increased visits to a competitor’s website will increase web traffic. The cookie stored from that site will inform Facebook and other advertisers that you have viewed that site. As a consequence, it will then serve up more ads related to that site/product. Additionally, the more visits a page receives, the more likely it is to show up in search results and other advertisements in general. If you want to research your competitor, this consequence is almost inevitable.


Data collection is here to stay. As privacy policies change, entrepreneurs should be mindful of their internet behaviors, and they should remain up to date on policies. While you can cover some tracks, some things are unavoidable and are the cost of doing business.



The Perils of Contracting Through Third Party Websites

MichelleSargentBlogPicThe internet has a wealth of resources for start-ups, from legal and venture financing blogs to form documents and state entity-formation pages. There are also many pitfalls to leveraging the internet that unknowing and uncounseled start-ups may not identify. One of these is the use of third-party websites to enter agreements.

For example, a start-up wants to use a contractor to develop some code for their mobile application. The start-up finds a website that allows parties to post a project that needs completing and has users make bids to win the project. (There are many such, or similar, websites, such as,, and The start-up posts its project, and an unknown individual somewhere in the internet world is contracted to complete it.

On its face, this is an easy and low-cost way of finding work for hire. But what may the start-up not consider: Is its intellectual property protected? Is any intellectual property created by the worker assigned to the start-up? What happens if there is a dispute about the completion or quality of the work? What is the worker’s classification for employment and tax purposes?

1.  Worker Classification

Any time a company wants to bring in somebody to do some kind of work for the company, the company must address how to properly categorize the worker for employment and tax purposes. As this blog has previously discussed, in a January 31, 2013 posting by Tasha Francis on Hiring Tips for First Time Entrepreneurs, there are important ramifications to getting the classification correct. And the correct classification may not be the one that matches the start-up’s financial capabilities or expectations. In using third-party websites to locate workers, the classification problem is magnified because (1) there may be no explicit agreement or provision stating how the worker is classified, and (2) it is possible the website loosely uses terms such as “employer” and “employee”—even when categorizing independent contractor relationships.

The purpose of these websites is to allow companies to hire remote, unknown individuals who specifically market and offer their services through the website to perform a specific project, on a specific time schedule, for a specific fee. This seems to create a quintessential independent contractor relationship. Many of these websites’ terms of service, however, explicitly refer to the company seeking services as the “Employer.” And the generic terms of service may not describe the nature of the relationship between the two parties contracting for work—unless the parties independently choose to enter into a separate independent contractor services agreement (the terms of which may not conflict with any terms in the terms of service). Although a start-up may ultimately be protected by the actual nature of its relationship with these electronically engaged workers, the terminology and lack of a defined relationship is just one of many ways relying on a third party may undermine the company’s legal rights and relationships.

2.      Intellectual Property

For many emerging companies, their intellectual property is the company’s main—and potentially only—valuable asset. And seeking to protect this IP is often one of the first steps taken by lawyers representing start-ups. Yet, when an uncounseled start-up engages through a third-party website, it is highly unlikely any of the typical IP assignment or proprietary information protections are contained in the default terms of the relationship.

Now we will return to the company that retains a contractor to develop a mobile application. On the one hand, the company must send proprietary information to the contractor describing its business, what its vision is for the mobile application, specifics for its user interface, the types of users it envisages, etc. Without any nondisclosure agreement, this proprietary information will not be protected from the contractor disclosing it to other parties, or developing the information in third-party products. On the other hand, the contractor will be writing code and developing an interface that will be an integral part of the company’s platform and business plan. But without any intellectual property assignment provision, any copyright and other intellectual property in the work performed by the contractor will not belong to the company. Although it may seem distant to the start-up now, any outstanding, unassigned intellectual property gives individuals a future opportunity to come back and seek something from the company.

It is highly unlikely that a third-party website will contain default provisions sufficient to protect existing company IP or to assign any contractor-created IP. On the contrary, many of these websites include express disclaimers that they are merely acting as a portal and are not subject to liability for any claims under any IP laws. It does not, however, include any IP assignment language from contractors to users, or any default nondisclosure provisions. Therefore, unless a start-up independently identifies the need to draft a separate and independent agreement assigning any IP and protecting the confidentiality of any of its proprietary information, it leaves potential holes in its IP protections.

3.      Dispute Resolution

So what if the relationship created on this third-party website goes wrong, what can a start-up do? Here, there are two main considerations: Does the third-party website impose any dispute resolution conditions on users? And should litigation arise, where would it take place?

A brief survey of the terms of service of several of these third-party websites reveals the existence of mandatory dispute resolution or arbitration provisions. What do these mean? Say the start-up is unsatisfied with the quality of the work performed by a contractor and wants to renegotiate the rate it has agreed to pay for the services based on the poor quality: First, the parties may be required to attempt to independently negotiate and resolve any disputes. They may also, however, be prevented from renegotiating the fee for a project after it has begun. Therefore, the parties would be required to submit the dispute to the third-party website, which may grant itself the full power to resolve the dispute, including the determination of the documents that can be submitted in support of the dispute. Alternatively, the parties may be required to submit to arbitration in which the third-party website selects an arbitration team that makes a binding, irreversible decision on the dispute. In many of the terms of service for these websites, there are no express provisions providing for litigation between users. And insofar as the purpose of the website is to bring together individuals globally to provide services, there is the distinct potential that contractors may be judgment-proof—either geographically isolated or financially unable to be hailed to court in a foreign country to face charges for breach of contract or copyright infringement.

And what if the start-up wants to raise a dispute with the third-party website? There are the obvious broad disclaimers of all kinds of warranties, limitations of liability, and indemnification requirements. But the terms and conditions can also establish that the agreement is subject to foreign law with irrevocable and exclusive jurisdiction in foreign courts. For uncounseled start-ups, the potential pitfalls of being unable to dispute a contract—because litigating abroad is unfeasible—were probably never considered.

These provisions and examples reveal the unforeseen liabilities of entering into agreements through third-party websites. Even for start-ups that may benefit from legal advice in order to draft separate agreements dealing with IP and worker-classification issues, it is likely that the third-party website’s terms and conditions will seek to limit the effect of these agreements to the extent they may be inconsistent with the website’s terms and conditions. And so even extra-website attempts to contractually change dispute resolution procedures may be ineffective. Nonetheless, there are differences in the terms of service across these websites, with some providing for more default, start-up friendly provisions or options to protect IP, classify workers, or dispute services, and at the very least, start-ups may want to consider digging deeper before selecting a third-party service.