DataX Naming Matters

What’s in a name? For entrepreneurs, selecting the right name for a company or product is vital. Great names can increase product memorability and trust, as well as sales. In short, naming matters. However, finding a fresh, unique name in oversaturated markets can be an extremely difficult, time-consuming process. Berkeley startup Naming Matters aims to change that. The startup provides visualization services, copyright validation using state and federal trademark registries, as well as connotation checks using machine learning, to help new companies avoid costly business and legal risks. However, these services break down when it comes to the tricky problem of common law trademarks, also known as “unregistered trademarks.” Recognized in the UK, Canada, and North America, companies can create trademarks without registering them; a history of operation alone validates the trademark.

This semester, Data-X’s Naming Matters team endeavored to solve this puzzle. A good solution to the problem would communicate a desired name’s risk level (in regards to the potential of encountering trademark litigation disputes), given the possibility of similar common law trademarks in existence, within the same industry. Rethinking the previous team’s solution, extraneous implementations of machine learning that only supplied additional complexity were removed. By implementing fuzzy search at key stages— using character-edit algorithms— this semester’s solution allows similar iterations of a potential name (including soundalikes) to be searched for as well, which constitutes a vast improvement upon currently available search solutions in the naming industry. Another big change to the last team’s approach is the implementation of a polished, intuitive user interface, to enhance the end user experience, as well as to allow for quick analysis and risk assessment. Lastly, most significantly, in order to identify risks for names of both products and services, the solution utilizes an open-source database with around 1 million unique records of business licenses across multiple states. The result is a much more streamlined and effective product, and a step closer towards solving the common law trademark puzzle.

https://github.com/kyupark94/NamingMatters