Traditionally, trademark searches are performed manually either by a trademark lawyer or the trademark owner. But software applications are becoming more common place to perform B2B services, and trademark searching is no exception. Software applications do not have the limitations that humans do and can perform the same task more efficiently and cost-effective.
Trademark conflicts are determined by applying the likelihood of confusion test. This test is composed of 13 factors, one of the most important factors being the relatedness of the goods or services at issue. The best way to evaluate the relatedness of goods factor is to look at prior court decisions where a finding of relatedness was made.
For example, consider “restaurant services” as the services for a proposed mark. There has been a lot of trademark litigation involving “restaurant services,” which courts have found are related to over 100 different goods and services. Theoretically, a human conducting a trademark search should search for the proposed mark not only in connection with “restaurant services” but also in connection with the over 100 different goods and services that were found to be related to “restaurant services.”
Practically, though, no manual search will do this because it would take days to complete this thorough of a trademark search and would not be cost-effective. If you estimate approximately 10 minutes per search when using the USPTO TESS database, conducting 100 searches would take about two, full business days to complete. If you also assume a cost of $50 a search, the cost would be about $5,000. Because of these costs in terms of time and money, humans cut corners when conducting trademark searches by relying on their intuition, which can lead to the wrong decision.
On the other hand, a computer will conduct the trademark search in connection with all 100 related goods or services, and will do it in minutes not days. The result is a more thorough trademark search at a fraction of the price of a manual search.