Determining which patents are valuable can seem a lot like picking winning lottery numbers. There were, as of 2014, approximately 2.5 million U.S. patents in force. The number of new patent applications has risen steadily to more than 411,728 each year, and if the current trend continues, about 70 percent of those applications will become patents. Like lottery tickets, most patents are worthless, but a precious few are worth billions.
Analytical valuation methods have become popular in distinguishing valuable patents from worthless ones. These methods estimate the value of a patent based on characteristics that have been correlated with the value of other patents in the past. I tested some of the most common patent characteristics used to predict value in a meta-analysis of prior empirical studies. I describe below how these analytical methods can be useful in picking a winning patent and when they are misleading.