Skip to main content
3 posts
Last seen: 05/26/2023 - 14:44
Joined: 03/16/2023 - 15:27
Discontinuity of the Number of IPC Technology Main Groups per Patent in 2013

Dear PatentsView team,

I'm using information on IPC technology main group contained in the ipcr.tsv file to trace the dynamics of the number of IPC main groups assigned to patents (See figure below). I notice a dramatic change in 2013 as the average number of IPC main groups per patent takes off around that period. Does anyone know what may be driving this discontinuity? Is there major change in patent examination practice around that period of time? 

I know the CPC scheme was first introduced and used by the USPTO in 2013. Does this relate to it? If so, how?

Thank you very much!

Discontinuity in the avg. number of IPC main groups per patent in 2013

Role: moderator
Last seen: 09/21/2023 - 16:12
Joined: 10/17/2017 - 10:47
CPC and IPC concordance

Hello Gavin, 
We are in agreement with your hypothesis that the upward spike in IPC patents was related and due in part to the beginning of CPC which started in January of 2013. CPC was made to work in partnership with IPC and was utilized by ECLA and USPTO beginning in 2013 which likely increased the ease of the pathway from national patent grant classifications to international classifications. 


Last seen: 05/26/2023 - 14:44
Joined: 03/16/2023 - 15:27
Time Trend of Average Number of CPC Main Groups

Hi PVTeam,

Thank you very much for your reply!

On a related issue, the average number of CPC main groups per patent (based on the CPC current file) was gradually increasing since 1976 but took off dramatically around 2013 when the new CPC scheme has been introduced. If my understanding is correct, the current CPC classifications for patents before 2013 are back-engineered, correct? But how much would you say of this increasing time trend as merely an artifact of the transition from USPC to the CPC scheme?

In other words, were patents in much earlier years assigned to fewer CPC main groups merely because the data team has not finished the back-engineering of CPC or there has been less effort spent to classify those earlier patents under the new CPC scheme? Is there documentation about how the back-engineering or reclassification of those earlier patents into the CPC scheme is accomplished?