4 ms·

Highly misleading title. Number of patents filed and cited is an outcome of many factors including innovation and company's focus on building a patent stockpile

by arocks 14y ago

Highly misleading title. Number of patents filed and cited is an outcome of many factors including innovation and company's focus on building a patent stockpile. Correlation doesn't imply causation. Companies like Pixar have only about 165 patents yet nobody would call them less innovative.

mjn 14y ago
I would guess it's actually the opposite effect causing these results: companies that take a shotgun approach to filing patents, which is typical of public companies, will rank lower on this patent-quality metric. The researchers ranked companies by average citations per patent, as a proxy measure for quality of their innovation. That would typically be higher for companies that only file a handful of key patents, like Pixar does, rather than companies like IBM that file tens of thousands of patents. But it still doesn't really say much about innovation either way. If company A files 2 high-profile and 0 obscure patents, while company B files 10 high-profile and 50 obscure patents, company A will have more citations per patent on average, but it's not clear that's measuring anything about innovation. If the proxy is useful for anything, it's a measure of how selective a company is in filing patents.
- nostrademons 14y ago
  The other effect is that people's work grows more specialized as the company grows bigger, and so each individual invention becomes more niche and less likely to be relevant to other people. PageRank is one patent, because at the time, Google was just Larry and Sergey (and owned by Stanford). That one component of the system might (hypothetically) have a team of 10 maintaining it now, and their work might (hypothetically) involve looking at obscure ways to weight links differently or block people gaming the system. Such patents wouldn't be useful unless you had a working PageRank system already.
- rgbrenner 14y ago
  From the article: "An IPO didn’t affect the rate at which a company obtained patents, but Bernstein found that public companies’ subsequent patents were far lower in quality, as measured by how often each patent was cited in other patent applications."

by arocks 14y ago

mjn 14y ago

I would guess it's actually the opposite effect causing these results: companies that take a shotgun approach to filing patents, which is typical of public companies, will rank lower on this patent-quality metric. The researchers ranked companies by average citations per patent, as a proxy measure for quality of their innovation. That would typically be higher for companies that only file a handful of key patents, like Pixar does, rather than companies like IBM that file tens of thousands of patents. But it still doesn't really say much about innovation either way. If company A files 2 high-profile and 0 obscure patents, while company B files 10 high-profile and 50 obscure patents, company A will have more citations per patent on average, but it's not clear that's measuring anything about innovation. If the proxy is useful for anything, it's a measure of how selective a company is in filing patents.

nostrademons 14y ago
The other effect is that people's work grows more specialized as the company grows bigger, and so each individual invention becomes more niche and less likely to be relevant to other people. PageRank is one patent, because at the time, Google was just Larry and Sergey (and owned by Stanford). That one component of the system might (hypothetically) have a team of 10 maintaining it now, and their work might (hypothetically) involve looking at obscure ways to weight links differently or block people gaming the system. Such patents wouldn't be useful unless you had a working PageRank system already.
rgbrenner 14y ago
From the article: "An IPO didn’t affect the rate at which a company obtained patents, but Bernstein found that public companies’ subsequent patents were far lower in quality, as measured by how often each patent was cited in other patent applications."