Big data as a term is a broad church that encompasses a range of innovation. The technology itself has evolved dramatically from the relational databases of the 70s, through to modern processing techniques for unstructured and structured data. Today,'big data' is often used synonymously with data analytics and data science, and can include anything from a whole project to just one aspect of a wider initiative. Effectively, it now covers all types of data analysis to derive insight, particularly where that data is constantly evolving or is large or complex.
The varied nature of big data, and lack of clear definition has led to uncertainty as to how innovation can be protected as intellectual property. For example, there is a common myth in technology circles that you can’t patent big data projects, which is false. Tackling this myth is important because, with a lifetime of 20 years and without having to evidence copying to demonstrate infringement, a patent is the strongest form of a protection for an innovation. This exclusivity can be critical in commercializing a big data invention, forming a business model, and even in helping companies secure funding.
Being technical, new and inventive
The criteria for every patent is that it must be technical, new, and inventive. Big data projects are often novel and inventive by nature, but it is commonly assumed they are excluded by failing the criteria of being technical, as they are typically based on software or algorithms. You may have heard that it is not possible to patent computer programs, as such – and it is true that this is a clause in the UK’s 1977 Patents Act. However, there is a nuance in the law that means that software can still be patented, as long as the invention can be shown to have a real-world technical effect.
Therefore, while you can’t patent a big data algorithm itself, you can protect aspects of the overall project. For example, if the functionality of the project provides an effect in the real-world, if it uses innovative technical ways of gathering the data for analysis, or if it makes the overall system work better.
Take, for example, an innovative system that gathers data to analyze retail footfall. The analysis itself can’t be patented. But the sensors and systems that retrieve and collate the data could well fit the criteria of being new, innovative, and technical.
Another example is a method of predicting bank fraud using data by analyzing the location of a transaction against the customer’s home address. This is also patentable in theory as it improves the functionality of the bank account in the real world.
Google’s patent for MapReduce, which it describes as a 'system and method for efficient large-scale data processing' is a real-world example of an innovation in collecting and analyzing data having been successfully patented.
Patents and big data provide both a risk and an opportunity
As mentioned above, Google holds granted patents on its MapReduce technology. Therefore, in theory, any big data project utilizing this framework - or any of its offshoots such as Hadoop - would notionally infringe those rights. There is a risk therefore that projects could be subjected to licensing costs or potentially halted entirely.
However, the reality is every big data project could be shown to infringe a patent right of some kind. The more important question businesses need to ask themselves is whether or not it will impinge business activities. What is the risk caused by the project infringing your rights, and is that manageable? For example, Google announced that it would not assert its rights against those using the Hadoop library.
What is more important for a business is the opportunities presented by patent protection. A patent application can demonstrate to potential customers, partners, investors, or buyers the innovation within a project. This makes it a valuable marketing tool and, even more importantly, provides leverage in collaborations and licensing opportunities.
With every project there is a risk and an opportunity, so businesses need to ask themselves whether the opportunities for patent protection are worth the cost. The important thing is that these companies consider patent applications, and do not reject them out of hand simply because the project is big data. It is also important to consider other IP rights that provide overlapping layers of protection for a company’s innovation in the big data space, such as copyright in the underlying code in software. However, if it is applicable, a patent is the strongest form of protection and offers companies the most commercial opportunity from their big data projects.