Can you predict which patents are high risk?

Patent risk managers would love to know which patents could eventually cause potential business disruption or could end up being licensed at high costs and which not. A number of prediction methods are currently used in patent risk management but none of them provide a high degree of certainty. Professor Colleen V. Chien of Santa Clara University School of Law recently reported research she did on this subject, providing a “prediction model” on how likely it is that a certain patent will eventually be litigated. This would make life of innovators, R&D institutes , operating companies as well as NPEs (patent trolls) a lot easier. If they conduct “patent clearance”, that is the search and analysis for patents the claims of which could hinder R&D activities or a product launch, it would make the clearance or freedom to operate procedure a lot more effective if they could “model” the likelihood that certain patents are more dangerous than others. Some products, especially in mobile technology, could easily be covered by over hundreds if not thousands of patents. However common knowledge is that only a very low percentage of patents are actually enforced, some say less than 3% of all issued patents issued.  Combined with the relatively high costs of patent clearance or patent risk assessment many companies and R&D institutes rather ignore patents than spend the time and money to find the “relevant” that is potentially dangerous, patents. Two groups of professionals would welcome the findings of Mrs. Chien. Risk managers, who are being paid to avoid risk, not ignore it, as well as valuation experts would have a new tool to predict which patents have a greater likelihood of being enforced and would therefore in general also have a greater value (e.g. by those acquiring the patent for licensing purposes).

She compared litigated and non litigated patents and researched the effect of “acquired characteristics” or events that happen after the patent has been granted. She shows that these so called “post-issued events” or “traits” of a patent improve insight into a patent’s worth. There are “intrinsic” characteristics  of a patent, those that are within the control of the patentee, (e.g. number and broadness of claims) and external traits,  events that occur after grant, such as whether the patent is being traded (transferred) or securitized, whether reexamination took place, or otherwise undergone investment of sort, etc.  Patents originally assigned to individuals and small companies are more likely to be litigated.

From older studies it is known that patents that end up in litigation differ from the patents that do not. Litigated patents have been differently prosecuted, may have be assigned to certain types of patentees, have more claims, more prior art citations and larger families and are generally older (and hence for which maintenance fees have been paid), to name a few.

Mrs. Chien however focuses more on the acquired characteristics of patents, decision points in a patent’s lifetime, such as:

  • changes in patent ownership
  • post issuance investment in the patent (maintenance fees, ex parte reexamination
  • patent collateralization
  • adjusted forward citations to the patent

The results of Prof. Chien’s research are, in her own words, “dramatic”, litigated patents do significantly differ after grant than unlitigated patents.

1. Litigated patents are more likely to be transferred and nearly four times as likely as unlitigated patents to experience a change in owner size.

2. they are hundredfold more likely  to experience ex parte reexamination than the unlitigated patents

3. they are maintained more times, on average, than are unlitigated patents

4. they are more often used as collateral

5. they are cited twice as many times

all of these differences are statistically significant.

Litigation is a costly enterprise. If Prof. Chien’s analysis can help identifying patents that can cause disruption of businesses, hinder R&D or can be subject to licensing enforcement actions  the IP risk management tasks will become a lot more effective and predictable and could significantly alter the way IP managers will conduct clearance and conduct risk assessment projects.

Click here for the full text of the research article

For a critical review of Prof. Chien’s results, see Patently-O and Prof. Lee Petherbridge Response, On Predicting Patent Litigation