As several outlets have noted, the USPTO has extended the AFCP 2.0 program (see here).

The program is generally viewed positively by applicants. There has been much discussion about the best strategy for utilizing the program, and some number crunching sources have even provided hard data showing how an AFCP 2.0 request can improve prosecution. Also, those providing examiner statistics often indicate the results of AFCP requests for a particular examiner.

But the thing with such data analytics is that sometimes data can be a self-fulfilling prophecy, with the tail wagging the dog. Use of the data itself can warp reality. Consider the AFCP 2.0 program.

If your trusty examiner statistics show you that the examiner is extremely difficult, and assuming most applicants use such data (admittedly not true, yet), applicants will likely stop filing AFCP 2.0 requests for that examiner in all but the clearest of cases. Then, if that examiner is evaluated by the USPTO management on whether he or she is effectively responding to AFCP requests, the data could be within the bell curve of normal. On the other hand, if applicants take the exact opposite approach and inundate the examiner with AFCP requests (and the examiner unreasonably refuses a high percentage), the examiner now will look to be much further outside the normal curve. Assuming USPTO management uses examiner statistics to review examiner performance, the later approach provides some hope that the examiner will be incentivized to change behavior.

Granted this example is extreme and somewhat silly (making applicants waste their time in a futile exercise), it illustrates how examiner data usage by applicants can perpetuate undesirable examiner behavior if they merely follow conventional thinking, but also how applicants might take unexpected actions to highlight extreme examiner behavior.

So, think about your strategies and how they may have long term unintended consequences. And remember that for every measure, there is a counter-measure. The patent examination system is not static - rather it evolves over time as the actors gradually change behavior.