Today at the talk on Measuring Success at IPBC Boston, I promised to post my notes on objectively monitoring prosecution spending. The main message that arose from all four panelists: data alone is not enough; the data is simply a tool and requires skilled hands to use effectively.

Where is the Money Going

Any company from a StartUp to a F500 spends most of its prosecution budget on legal fees for preparing and prosecuting applications.After reviewing over 10,000 assets, I have seen assets prepared by countless prosecutors and law firms across the planet and have learned one undeniable truth: weak prosecutors neuter the protection of strong inventions, while great prosecutors are worth every penny. A patent owners needs to replace substandard prosecution counsel, immediately. If you can’t tell the difference between great and weak prosecution (note: size of the law firm means nothing in this regard), then here are some indicators.

Assessing Prosecutors

First, take a look at the work product. Read the issued claims. Do they map on the intended products that the application was intended to read on? If the answer is no, then little else matters. But this is not a reliable exercise to objectively compare prosecutors.

The goal then is to obtain objective data on the performance of those prosecution dollars. The objective data is then used to: 1) assemble a feedback loop that enables prosecution measurement, 2) facilitate a decision making process that uses the objective data in connection with prosecution goals and strategy, 3) communicate expectations based on the data, and 4) create accountability against those expectations.

What Data to Collect?

Today’s data-rich world offers an abundant amount of prosecution data, much of which was not even available fifteen years ago. There is too much data. The trick is in collecting objective and meaningful data that facilitates IP strategic decision making.

Here are a few examples of Good Data.

1) Time to File, calculated as the number of days from an outside counsel receiving a disclosure to filing the regular application. Theoretically, this number should be lower, but not zero; a good prosecutor needs some time to perfect good work product. But I have seen prosecutors take 180 days to file an application from receiving the disclosure. It happens a lot. In the old days, this was bad prosecution. With the new first to file system, this is simply unacceptable. What I like about this number: it is measurable, objective, and meaningful to the work product. It can be used to compare individual prosecutors, averaged for a firm and all its prosecutors, and therefore used to compare prosecutors and/or firms. Even more uses for this number are outlined below.

2) Number of days of Patent Term lost due to Applicant delays per issued patent – attributable the law firm. This is another objective and very meaningful number since one can readily argue that every day of patent term lost hurts the patent owner.

3) Number of extensions during prosecution of a patent. Ideally, there should not be any, but even good prosecutors will find an instance or two where it makes sense to delay a month. That being said, fourteen months of cumulative extension in a single application smells bad. Again, this number can be averaged per prosecutor and/or per firm for comparative purposes.

4) Total dollar amount in paid extensions. Some clients refuse to pay if extension fees are law firm based. When totaled for a firm, this is a telling number. I have seen firms 20X their neighbor firm for the same client. For a large client that uses multiple law firms, the firms stacked against each other will vary more widely than one might imagine in these last three stats.

5) Ratio of in person or telephone interviews to the number of Office Action responses (not including restrictions). In reviewing thousands of assets, I can say that the best prosecution occurs when the practitioner and Examiner get on the phone and/or meet in person. There is minimal prosecution, fewer office actions and amendments, far less estoppel, etc. Every million dollar patent that I have sold was prosecuted with a phone call or in person interview. While one can set aside some variance for style, a ratio below 15% is generally bad. The best prosecutors average above 50%.

6) Normalized stats via If you have not seen, then take the free trial and explore what it is there. The data don’t lie. It is tough to take some stats such the number of Office Action Responses or number of amendments on their own given how much effect an Examiner or Art Unit can have. Thanks to Patent, we can normalize the data for a prosecutor or firm against an Examiner or Art Unit. For example, when one prosecutor or firm averages 4.5 responses per allowance for a set of cases in a first Art Unit that averages 5 responses per allowance, compared to another firm’s 4 responses per allowance for a set of cases in a second Art Unit that averages 3 responses per allowance, normalizing the data shows that the first firm is actually doing better. offers abundant data that can be collected in normal prosecution for sake of comparing prosecutors; I’ll leave it to you to figure out which data you deem best for your program.

Putting all of these numbers together will collectively generate objective feedback on how your prosecutors are doing. You may be amazed when you see the stats. Crappy prosecutors will not like this exercise since it creates accountability and they will get exposed for substandard work. Great prosecutors will only be affirmed.

Bad Data

I left off many common numbers often used to assess prosecutors. My least favorite is cost per application. I strongly advocate finding a good prosecutor and paying them well. Irrespective if Jane is at a large law firm or working out of her home office, her data on the above stats will shine – except cost per application (where it may actually be less for the latter.) Flat fee billing is a red herring since it is a cost control and accounting measure; it is not meant for delivering patent quality. Just hire great prosecutors with integrity, and pay them well. Any professional services falls under the axiom of you get what you pay for. One does not want cheap legal representation in a defense lawyer, or a patent prosecutor.

I also left off number of words in a claim and a few other commonly overused stats since these are ultimately snake oil stats. Let expert eyes assess a claim relative to company or competitor products and impacted revenues, not easily duped metrics.


1) Show it to your prosecutors. If you have multiple outside counsel, let them know where they stand relative to the competition. Provide a quarterly ranking using the stats. This affirms the good behavior, e.g., prosecutor Jane was #1 out of 16 practitioners in terms of time to file, while flagging and punishing bad behavior, e.g., prosecutor Jim was #16 out of 16 practitioners for taking extensions. Put him on the spot; let him know. I’ve seen this done and it was VERY effective every time.

2) Use the data to throttle provision of new applications to outside counsel. Time to File is especially useful here. A slow acting firm sitting on old disclosures simply receives fewer new disclosures, and their billable hours suffer proportionately. In reality, a large firm likely has multiple prosecutors working on a large client and there is probably one slow actor, which either the firm or the client can reprimand or jettison as desired to preserve case flow for the others.

3) Use as a metric to replace underperforming prosecutors. A bad prosecutor will be bad on most of the stats. Release the worst one and get a new one, or add more work to the prosecutor or firm with the best data. Create a prosecution meritocracy based on objective data. Repeat as desired.