Jan 30, 2019Legal
Patexia Insight 55: How to Differentiate Your Prosecution Practice? (A Case Study of Apple)

A few weeks ago, we discussed how patent prosecution practice is becoming commoditized. Since then, a number of firms and attorneys have contacted us and explained how their work and its quality sets them apart from all others.  They claim their clients do not look at their service as a commodity, but they admittedly had similar concerns and wanted to know how to better differentiate their practice and communicate to prospective clients.

We decided to take a deeper dive into this and try to explain the best practices and ways to highlight that for new clients in the context of a Case Study related to Apple and its top five firms (out of almost 50) that have been handling approximately half of its patent prosecution load over the last five years: Meyertons Hood, Womble Bond, Treyz Law Group, Morrison Foerster, and Kilpatrick Townsend. Similar studies can be performed for all other companies at both attorney and law firm levels.

With the advancement in technology and computing power provided by cloud platforms, analyzing data in bulk has become both feasible and affordable. This type of competitive analysis helps identify valuable statistics and correlations when evaluating outside counsel.  Previously, many organizations had not been leveraging this type of analysis due to the  difficulty, time, and cost associated with such an undertaking. Advances in technology have made these concerns obsolete.

We looked at all of Apple’s utility patent applications (US Only) issued or abandoned over the last five years (i.e., January 1st, 2014 through December 31st, 2018). We decided to exclude design patents, as they are somewhat different in terms of both prosecution time and allowance rate (Sterne Kessler seems to be the firm handling most of Apple’s design patents). Over this period, Apple obtained 10,476 utility patents through many counsels working for almost 50 different law firms. Some of these firms filed only a single application while others such as Meyertons Hood handled more than 1,800 utility applications. The following table summarizes the top five firms that obtained about 49 percent of all utility patents for Apple in the last five years.

 

EntityIssued Patents (2014-2018)Office ActionsExtensionsPendency (Days)
Apple

10,476

3.11.71103.8
Meyertons Hood1,7372.51.4951.4
Womble Bond1,1622.81.51070.6
Treyz Law Group7873.31.31062.5
Morrison & Foerster7323.92.21341.2
Kilpatrick Townsend6642.71.6941.4

 

While obtaining a patent has been historically prestigious, we in the IP business recognize that more than 90 percent of patents are expired and worthless and it is not generally that difficult to obtain a patent if that is the only goal of the applicant. However, obtaining a high quality patent with a set of broad claims that has already been challenged by the examiner with several strong pieces of prior art - and can be enforced without high risk of invalidation - is not easy at all.

Here at Patexia, our Data Science team looks at more than 20 different signals extracted from patents, examiners, PTAB, district court, and many other sources of data to assess the quality of patents as well as all stakeholders involved, including attorneys and law firms.

In the above table, we have summarized a few of these signals we picked for the purpose of this case study for the top five law firms working for Apple:

Activity: This is the simplest metric showing the number of patents issued. For example, based on our data, Meyertons Hood has obtained the highest number of patents (1,737 utility patents) for Apple over the last five years.

 

Allowance Rate (Application-Level): This metric measures the success rate at the application level. It is simply the ratio of issued patents to all applications. While Apple’s overall allowance rate has been 90% over the last five years, all these firms except Kilpatrick Townsend seem to have a higher allowance rate. This means most of the remaining firms Apple had used generally had a lower allowance rate (i.e., their allowance rate should be less than 90%). In other words, it seems that Apple is giving about half of its prosecution work to firms with better success rate than the others.

Allowance Rate (Claim-Level): This metric is calculated for issued patents only and is simply the ratio of claims allowed to total number of claims in the original application. While getting allowance for an application is generally considered a success, we all recognize that there is a difference when only one claim out of 20 claims is allowed versus another scenario where all 20 claims are allowed. Apple’s average Claim-Level Allowance is 91%. It’s in the top five, all but Treyz Law Group hover around 90 to 92 percent. The Claim-Level AllowanceTreyz Law Group is 84%.

Number of Office Actions: Back and forth between the examiner and the attorney through numerous Office Actions may result in higher quality claims. Weaknesses of the claims might be captured by the examiner through different pieces of prior art and addressed by the attorney through his/her responses to those office actions and possible amendment of the claim. However, on the downside, this usually means longer and more expensive prosecution.

 

The above chart shows the average number of office actions, the top five firms used while working on Apple’s patent applications over the last five years. We understand that different companies may have different strategies when it comes to leveraging various tools provided by the patent office to continue the prosecution if need be. That said, it is fair to assume that a single company should apply more or less the same strategy for their prosecution regardless of which firm or attorney who is handling it.

The above chart shows that Meyertons Hood is generally faster (with an average of 2.5 office actions) while Morrison Foerster seems to be the slowest with an average of about four office actions. The average for Apple has been about three office actions per utility application over the last five years.

Extensions of Time: Another metric that is an indicator of efficiency of the firm/attorney and also cost of prosecution is the number of times an attorney requests an extension of time. Oftentimes, extension request requires a fee payment ranging from a few hundred dollars to as much as $3,000 (for large entities). This could substantially increase the cost of prosecution, while at the same time causing delays. Sometimes it may become critical to come up with the best response to an office action. However, many unnecessary extension requests could be avoided to expedite the prosecution and reduce the cost. Our analysis shows that Apple on average has used 1.7 extensions per application. This is while Morrison Foerster has used an average of 2.2 extensions per application and Treyz Law Group has used 1.3 extensions.

Pendency: Another measure of efficiency is the overall pendency, or the time from filing to issuance. Our analysis shows that on average, it has taken about 1,104 days (three years) for Apple to obtain a utility patent over the last five years. However, Meyertons Hood seems to be quite efficient with about 950 days (well below the average for Apple). In the top five, Morrison & Foerster seems to be on other side of the spectrum with 1,341 days (44 months). It is quite possible that Apple uses different counsels for different kind of inventions/business units and sometimes the strategy for a particular business unit might be different. That said, this analysis can be further fine-tuned by IPC, USPC, Technology Center, or Art Units if need be.

 

Application to Patent Ratio for Number of Words in Independent Claims: Another interesting metric we have used to evaluate the performance of attorneys/law firms is the ratio of average number of words in independent claims of the original application (published application) to that of the issued patents. The change in the number of words per (independent) claim shows how much the claim was limited through prosecution. If the claims end up with no change, the ratio would be 100%. However, usually some limitation is added during the examination, which increases the length of the claim and as result, lowers this ratio from 100%.

Our analysis of Apple’s data shows Meyertons Hood together with Kilpatrick Townsend have the highest ratio (73%) among the top five firms. Apple’s average ratio for its utility patents has been 69% over the last five years.

We should add that there are usually many strategies used by different patent attorneys, and oftentimes we hear that they start with a very broad (and short) claim and wait for feedback from the examiner before they further adjust the claim scope. Sometimes they may use keywords that are not adding any limitation (e.g., “said”, “whereas,” etc.) but artificially increases the length of the claim. This strategy may visually influence the examiner’s first impression when he/she looks at the claim.

While we used this Case Study as a sample to show how data could bring to light many facts and details about a particular company, law firm, or attorney’s prosecution work, we should add that there are many other factors that may impact some of the statistics we covered above. For example, backlog in the patent office in a particular Technology Center or Art Unit may impact the examination timeline. Also while all examiners adhere to the guidelines provided by the USPTO, there is always slight variation in their styles which could affect the examination.

For the purpose of this study as well as our Annual Patent Prosecution Intelligence Report (2019 Edition), which will be released in March, we have mathematically removed the noise caused by companies, examiners, subject matter (art units), and law firms. That means, when we look at the performance of a particular law firm, using regression modeling, we remove the noise caused by different examiners, art units, and clients. This should help make the comparison fair and bring our calculations as close as possible to real performance values for all the stakeholders.

Big law firms have dedicated marketing budgets that they allocate to advertising through conference sponsorship, trade magazines, client visits, etc. However, smart firms now look at data analytics as the best way to emphasize their differences and stand above the competition whenever they have that opportunity.

Companies are also looking at patent analytics to gain insights about the performance, efficiency, and success of their outside counsels, leveraging the data to make better decisions or guide their counsels more effectively.

Patexia has been advising its law firm clients on how to improve their profitability using patent analytics while simultaneously working with corporations to help them better understand the performance of their outside counsel or select the right firm for their operation. In March, we are releasing our first Patent Prosecution Intelligence Report, ranking the top 500 law firms from a pool of over 5,000 law firms, as well as the 500 best-performing companies from a pool of more than 100,000 entities. This data covers the last five years (2014 - 2018) and over two million applications and 1.6 million patents. We have divided the analysis to cover both high-tech and bio-tech sectors. Stay tuned…

 

Disclaimer: The data for this study was obtained from public sources including USPTO, PTAB, and PACER. Patexia has gone to great lengths to provide valid and accurate analysis based on this data. However, Patexia does not guarantee 100 percent accuracy nor take any responsibility for possible losses caused by use of information provided in this study.

Share
7 Comments
Anonymous This is very interesting data but like many data analyses, it is not very useful in real life. The reason is that in house attorneys often assign work to different firms based on the technology or business unit the invention comes out of. Therefore different firms are likely to handle inventions with very different technologies. They will be categorized in different classifications and directed to different examiners in different art units or event tech centers where aggressiveness of the Examiners will vary. Furthermore, the different technologies will have prior art of different densities. A patent application directed to one technical field may have very little ("sparse") prior art whereas another patent application directed to a different technical field can have very crowded prior art. This understandably will impact the prosecutions, making the firm handling the second application seem inept compared to the firm handling the first application.
Jan 30, 2019
Pedram SameniYou are absolutely right about that. As we explained in the article, we have mathematically normalized the data to remove the effect of different examiners or different art units. We use regression modeling and since there are more than 10M patents and patent application, we can remove the noise related to different stakeholders when calculating the stats for a particular attorney or law firm. In other words, we try to find the outcome of prosecution for a particular application independent of the examiner. If there is enough data, we can calculate the impact of the examiners and remove that from the calculation. That's what we have done in all our performance measurements.
Jan 30, 2019
Anonymous i agree with the first comment. the only way to perform a truly fair analysis would be to evaluate different firms handling the same application (set of applications). of course, as each application is unique (or at least supposed to be), this is not possible.
one might consider it better to compare the performance of, say, a US firm applying for a set of US patents vs a European firm applying for the same set of patents in Europe, but then you run into the problem of the different rules of the respective patent offices affecting the results.
for each of the metrics above, it is impossible to separate the causes of each result (for example, an allowable rate of above 95%) without knowing the particulars of each case (quality of the invention, lenient examiners, lack of prior art in that area, narrow claims, superbly argued office actions, wide but well-written claims, etc.).
the fact that people are using these metrics suggests to me that people already basically accept the premise that patent prosecution services are commoditized. only if you assume that quality of claim formulation and office action arguments are the same does it make sense to base your decisions on things like allowance rates or speed of prosecution.
having said that, corporations have to base their decisions on something, and given the impossibility of performing a close analysis of every claim and office action an attorney has ever written, making the assumption that patent prosecution services are a commodity may be the only reasonable way ahead. however, when you have an invention in front of you that could be the next big thing, or you want to make a divisional application that targets a competitor's best-selling product, you will probably want to choose your attorney based on something more than a metric like allowance rate. the devil is in the detail.
Jan 30, 2019
Pedram SameniYou are correct. in general it is almost impossible to compare two different people / measurements, unless the condition under which the measurement is taking place is 100% the same: same examiner, same application, same rules and same time. However, with large set of data, you can predict the outcome very accurately. Let me give a very simple example to make my point: imagine we have two attorneys, A1 and A2. Attorney A1 works on 100 WIFI related applications for Broadcom, Qualcomm and Intel. He obtains allowance rate of 70%, 80% and 90%, respectively while working with the exact same examiner E1 for all 3 companies. Now imagine Attorney A2 has also worked on 100 similar WIFI applications for Broadcom and Qualcomm and under the same Examiner E1, he has got allowance rate of 70% and 80%, respectively. We can predict with a very good accuracy that if Attorney A2 works for Intel on similar WIFI applications with Examiner E1, we should expect an allowance rate of about 90% for him based on the stats we have from A1. In other words, with enough large set of numbers, we can model the behavior of different stakeholders under different conditions (e.g., different Art Unit, different examiner, different client, ...)
Jan 31, 2019
Anonymous I wrote the initial comment. I am still not convinced that your analytics is helpful in the real world, even if you are great at normalizing your data for different patent classifications, examiners, art units (which I honestly don't believe is possible). Let's take two firms: One that prepares short but narrow claims, another that prepares broader claims. The first firm will look great in your analysis because the narrow claims will grant quickly. Of course, the claims may be unenforceable, but that's not taken into account by your algorithm. Meanwhile, the second firm fights for broad claims and has to fend off multiple office actions citing different art,l and may even have to go to appeal. The end result of second firm will be a stronger, more valid, broader, and enforceable patent.
Feb 3, 2019
Anonymous "we should expect an allowance rate of about 90% for him based on the stats we have from A1."
Feb 3, 2019
Anonymous I do not see what is useful about this information (maybe I am missing something).
Feb 3, 2019