How a different approach to patenting can help to close the gap

Building a patent portfolio is an expensive endeavour. Businesses expect their patent teams to be efficient and make sound financial decisions. To stress this point, this article refers to patent portfolio managers as patent investment managers (PIMs). Today, PIMs can use data-driven analytics to improve the efficiency of established portfolios. However, they face an ‘information gap’ when deciding whether to file new inventions because they must make investment decisions years before the relevant data is available.

This article provides four principles for overcoming the information gap and improving the efficiency of invention-harvesting activity (if not company activity overall). The theory behind each principle is then explained using real-life examples from Uber’s self-driving programme, Uber Advanced Technologies Group (ATG).

Among other things, readers will learn why the patent team at Uber ATG stopped using invention disclosure forms (IDFs) and how the group leveraged its approach to invention harvesting for the benefit of other legal teams. Although topics such as invention harvesting may be as old as the patent profession itself, there is always room for improvement and innovation in how companies make patent filing decisions.

Introducing the information gap

Patent decisions are investment decisions. Patents, like any other assets, have a value and an acquisition cost. If the value exceeds the acquisition cost (and there is room in the budget), then the PIM should recommend a “buy” decision. Otherwise, they should pass on the opportunity. This decision for a single investment can be modelled as follows:

Net value (NV) = Patent value (PV) - Acquisition cost (AC)

The actual calculation of PV is well outside the scope of this article. As such, references to topics such as patent “value” and “quality” are used to illustrate concepts relevant to improving the efficiency of patent-harvesting activity. Such references should not be interpreted as statements about the actual value or quality of an issued patent, and this article should not be relied on in any hypothetical negotiation or in any attempt to establish a reasonable royalty.

Fortunately, a PIM does not necessarily need to estimate PV for each patent investment decision. Instead, they can improve investment performance simply by avoiding patent assets that are unlikely to contribute positive PV to the patent portfolio. For example, one can safely assume that the PV is equal to or less than zero if the claimed invention is never implemented or desired for implementation. Stated another way, a patent must claim relevant technology in order to have a chance at contributing positive NV to a patent portfolio.

Therefore, PIMs should prioritise investments in patent assets claiming relevant technology. Although great in theory, consider a PIM’s investment decision on an unfiled patent application. How can the PIM limit investments to patent assets claiming relevant technology in this scenario? After all, the relevance of the technology will not be known when the filing decision must be made. By the time the PIM has strong data on adoption of the invention, a time bar should have already occurred (eg, public use or sale) and the PIM will have missed the opportunity to make the filing decision.

This is the information gap. PIMs responsible for filing new patent applications do not have the information that they need at the time they must make patent investment decisions. Instead, they have two choices: file on everything and delay hard decisions until more information is known; or guess which technologies merit patent investment. Unless the PIM has an unlimited budget (a luxury not typically afforded to other types of investment manager), the first choice is not a viable strategy.

Unfortunately, predicting which inventions will be adopted is hard. Predicting how inventions will be adopted is even harder. Although not addressed in this article, patents claiming relevant technology can still have minimal NV depending on how the claimed technology is implemented.

Few US law schools teach lawyers how to predict which inventions will be adopted, and even fewer law firms specialise in such analysis. This is a somewhat frightening phenomena: PIMs are charged with making financial decisions without being fully trained on how to make those decisions. Accordingly, even though the first choice (filing on everything) is not a viable strategy, some PIMs will default to this out of fear of rejecting the wrong inventions and foregoing patent protection on innovations that might someday be valuable.

Some business school programmes, however, do teach students how to predict which innovative technologies will be brought to market. For example, entrepreneurship programmes teach students how to assess a new technology, identify what market need exists (or could exist) for this technology and build a business plan around it. Unfortunately, this can be a rigorous exercise, involving customer interviews, SWOT analyses and many iterations of review to hone in on the best commercialisation path for the technology at issue. Being realistic, a PIM cannot scale this exercise across a large number of different inventions.

The good news is that many PIMs can often skip this step. Why? Because their company likely has performed some of this analysis when making decisions about next-generation products. Although new inventive aspects of a technology may not have factored into such analysis or be a factor in driving customer demand for a new product, the decision to use the new inventions in a new product may suggest the potential for future technical relevance or that other companies will aim to do the same thing. More specifically, new inventions selected for inclusion in a new product may have more potential for technical relevance than new inventions detached from any such product plans.

Of course, it is possible that neither invention will yield positive PV. For example, many decisions to incorporate technology into a new product are trivial or based on reasons that do not contribute to PV. As such, mere selection for inclusion in a new product is not indicative of what the future PV will be. Instead, selection for inclusion simply indicates a higher potential for technical relevance or positive PV, whereas inventions rejected by the product team are less likely to have technical relevance in the future.

This brings us to the first principle for filling the information gap and improving the efficiency of the invention-harvesting programme.

Principle 1 – prioritise inventions that your clients are building into products

This point may seem too obvious to merit publication. After all, many PIMs are likely to believe that they prioritise such inventions. However, subjectively appreciating the importance of these inventions is different from prioritising them in a way that drives efficiency.

For example, PIMs cannot prioritise the inventions that their clients are building into products if they do not capture commercialisation decisions in the invention-harvesting process. Most organisations capture information relevant to decision making through IDFs, but these rarely ask about the company’s commercialisation plans. Instead, most IDFs merely attempt to identify upcoming disclosure dates. Some IDFs may ask for the inventor’s opinions about future commercialisation potential, but these opinions are a poor substitute for information about the company’s decision on whether to invest resources in commercialisation of the invention. To this point, it may not be possible for an IDF to adequately capture company commercialisation information because the forms are almost always completed by inventors, rather than the company personnel making resource decisions concerning the development and deployment of the invention.

The Uber ATG patent team does not require inventors to complete IDFs. Instead, it utilises passive monitoring to understand what teams are building and to harvest inventions that are currently being built. Passive monitoring works by tapping into the ways that engineers naturally communicate with each other. Most engineering organisations have a variety of ways to spread information and to inform teammates about new innovations and developments. Organisations can use distribution lists, document repositories, product or programme reviews, tech talks, science fairs and other ways to ensure that people are aware of the latest developments. PIMs can tap into these communication modes to identify new inventions worth harvesting.

Passive monitoring helps to capture business commercialisation decisions because engineering communications tend to focus on product-related conversations, naturally weeding out interesting ideas with less commercialisation potential. Product-focused inventions also come with better documentation, which can be repurposed and fill the place of the IDF. Although Uber ATG still files some patent applications without the benefit of business commercialisation decisions, our PIMs can treat these cases as more speculative and limit investment in such assets.

In addition to capturing commercialisation information in the harvesting process, PIMs can improve efficiency by deprioritising harvesting activities that are not informed by business commercialisation decisions. Ideation sessions, for example, are an easy target. These sessions promote the free exchange of ideas, no matter how bad the ideas are. As a consequence, they do not get the same critical review as an innovation being incorporated into a new product. At a minimum, PIMs should look to improve ideation sessions to yield more valuable inventions. For example, consider inviting marketing or product professionals so that they can contribute market data and customer feedback to the discussion.

Remember, prioritising inventions that the company is building into products is an approximation of a more rigorous commercialisation study. Thus, if an approximation is not available (eg, an invention reaches the PIM before the company has decided whether to incorporate it into a product), the PIM may want to delay the decision by filing a provisional patent application.

Alternatively, PIMs should be equipped to perform an independent commercialisation analysis on inventions that the business has not already reviewed for potential inclusion in a new product. For example, the PIM can listen to the inventor’s pitch, ask questions that focus on the potential for commercial adoption and assess whether the inventor was motivated by meeting a widely felt need. Even if the PIM does not have the benefit of the business’s decision to incorporate the invention into a product, they can make a stronger guess by asking rigorous questions concerning commercialisation potential.

Because prioritising inventions that the company is building into products is merely an approximation of a more rigorous commercialisation study, PIMs must also appreciate the limitations of this activity. Most notably, reliance on the company’s commercialisation plans ignores the differences between the company and its competition. For example, some of the company’s innovations are directed towards problems that are unique to the company and less likely to be adopted by other firms. Conversely, some may be more attractive to other firms than to the company itself, and thus the company’s commercialisation decisions may undersell the value of the innovation to the industry at large. By appreciating these limitations, PIMs can start to tailor patent investment decisions to the unique nature of every invention.

Stated another way, prioritising inventions that the company is building into products is a starting point, not a destination. In fact, PIMs can supplement such prioritisation by taking advantage of other business research to inform patent investment decisions. Examples of such business school research include studies on futurism, forecasting, scenario planning and adoption cycles. Unfortunately, there is no single research study or collection of studies that all PIMs can use to inform filing decisions. Instead, they should choose the studies that best reflect the company’s industry and future plans.

To illustrate how this works in practice, consider the Gartner Hype Cycle (see Figure 1). According to the methodology of this, a technology’s lifecycle includes five phases: an innovation trigger, a peak of inflated expectations, a trough of disillusionment, a slope of enlightenment and a plateau of productivity. Every year, Gartner publishes a list of technologies and identifies where it believes they fall on the hype curve.

Figure 1. The Gartner Hype Cycle

PIMs can use this methodology to assess where a new invention might fall on the curve and to identify any obstacles that may hinder its commercialisation. To be clear, PIMs should utilise the existence of the curve itself without necessarily relying on Gartner’s annual hype data – after all, PIMs will often need to make filing decisions on technologies before Gartner places them on the hype curve, and waiting for Gartner to release the relevant data may mean missing the ideal investment opportunity.

That said, the Gartner hype data raises an important point concerning harvesting strategy at Uber ATG. Uber ATG is developing self-driving technology and the 2019 Gartner Hype Cycle identifies “Autonomous Driving Level 4” as approaching the trough of disillusionment and being more than 10 years away from the plateau of productivity. This article will not (and cannot) comment on Gartner’s predictions concerning self-driving. It is true, however, that self-driving is a long-term research effort. Applying Principle 1 to such efforts is particularly difficult because it may be years before the PIM receives any actionable commercialisation information.

Applying Principle 1 to long-term development programmes

Generally speaking, identifying inventions with higher chances of technical relevance is easiest in industries where products launch quickly. Companies that launch products faster make product decisions faster, which gives PIMs more information when making initial filing decisions. In some situations, the PIM may even have product information before needing to make foreign filing or prosecution decisions.

At Uber ATG, the patent team does not enjoy such luxuries. The development of self-driving technology is a multi-year effort, and engineering teams often change designs when iterating over multiple generations of technology. Many different innovations are identified over the years. However, only some will be part of the ultimate product; the rest will be discarded.

So how can a PIM rely on the company’s engineering and product decisions when the technology keeps changing? The best the PIM can do is prepare for change. Knowing that some inventions will be discarded is useful even if the PIM does not yet know which inventions will be replaced. Figure 2 presents one example model that PIMs can use to optimise investment decisions for long-term development programmes.

Figure 2. A model for PIMs to follow

The model assumes that patent filings will fall into three categories based on the filing date. Starting on the left, filings in the problem identification phase are more likely to have technical relevance because early filings are more likely to lead to broader claims and because new technologies may build upon the inventions claimed in the earlier filings, increasing the long-term relevance of the earlier filings.

At the other end of the chart, the most recent filings in the solution identification phase are also likely to have technical relevance because, even though these filings will have narrower claim scopes, such claims should cover the final technical solution (and thus have the best chance of being incorporated into products). In the middle is the iteration phase; filings during this phase will have narrower claims than the problem identification phase, but they will not be as precisely directed towards the ultimate product as filings from the solution identification phase. As a result, filings during the iteration phase are less likely to have technical relevance.

Unfortunately, PIMs applying this model will encounter a variation of the information gap. Specifically, they will not know when the iteration phase and solution identification phase will occur until after they have passed. As a result, PIMs cannot reasonably ramp down filings during the iteration phase and then ramp back up during the solution identification phase.

Despite this information gap, PIMs can still use models such as these to make more efficient investment decisions. For example, once the iteration phase and solution identification phase have been identified, the PIM can still shift some resources away from the iteration phase filings. Depending on timing, the PIM can abandon cases in the iteration phase and spend more on continuations and foreign prosecution for cases in the problem identification phase and more on domestic prosecution and foreign filing for cases in the solution identification phase.

Principle 2 – focus invention harvesting on patent portfolio needs

Principle 1 (prioritising inventions that the company is building into products) helps PIMs to make decisions on individual patent applications. However, PIMs are responsible for building a portfolio that serves the overall needs of their business. This requires a top-down analysis to determine what the overall patent portfolio should look like to best serve the company.

Unfortunately, adherence to Principle 1 alone will not yield a portfolio that matches the top-down analysis. Prioritising inventions that the company is building into products helps PIMs to make decisions about which identified inventions to protect but does not help them to target inventions that are needed to satisfy the top-down analysis. PIMs are limited in this effort, as most do not direct research or choose where engineers will innovate. In fact, even if a PIM could direct research activities, the yield would be underwhelming unless the directed research was informed by market data, customer feedback and other indicators of future commercial success.

Thus, Principle 2 is to focus additional harvesting activity on specific technologies to bring the portfolio closer to matching the top-down analysis. At Uber ATG, the patent team supplements activities such as passive monitoring with active targeted harvesting to identify additional inventions in areas of importance, as reflected by the top-down analysis. If, for example, the PIM wants to increase filings in a certain technical area based on a top-down analysis of the portfolio, then they can engage with the appropriate groups to learn more about what they are building and plan to build next. When working with a targeted engineering team, the Uber ATG patent team looks for the same type of commercialisation evidence as it does during passive monitoring. In other words, the team adheres to Principle 1 when implementing Principle 2.

The key word here is ‘targeted’. The patent team selects only a few engineering teams at a time for this activity, rather than broadly harvesting across the entire organisation. Targeted harvesting has two primary benefits over broad harvesting: the more targeted the harvesting, the fewer engineers experience disruption from their day jobs; and limited harvesting means that the patent team will have the resources to provide a white-glove service to the inventors that have been disrupted.

Targeted harvesting complements the weaknesses of passive monitoring. This is not designed to capture all innovations at a company but rather to survey engineering and product activities. This approach works because it yields a high percentage of inventions that the company is building into products. If a PIM were to expand passive monitoring to try to capture all new inventions, then the additional benefit would be limited and the efficiency would be lost.

Combining passive monitoring and targeted harvesting is efficient because both activities serve modest goals: identifying a high percentage of quality inventions using as few resources as possible. These activities are not designed to identify all inventions at a company and they will fail if measured against that standard.

Such modest goals are a feature of the approach, not a bug. Identifying all inventions is simply not an efficient goal for an invention harvesting programme. Even ignoring efficiency, it is not clear how such a goal would serve any company’s strategic objectives. Yet this is the approach encouraged at many organisations. For example, some try to encourage invention harvesting by paying inventors for their IDFs, their filed patent applications and/or their issued patents. Some patent teams are even measured based on the number of inventions that they identify. Such metrics do not incentivise quality, technical relevance or achievement of overall portfolio goals reflected by a top-down analysis.

Instead, these metrics incentivise volume and often lead to the mass proliferation of IDFs. In addition to efficiency issues, mass reliance on IDFs may not even reap the intended outcome. IDFs rely on self-reporting by inventors even though inventors are not typically experts on patentability issues and the best inventors are often too busy to complete the forms. In fact, there may be a correlation between busy inventors and inventors whose inventions end up in products, in that the additional time required for product development support might be why some inventors are too busy in the first place. From this perspective, self-reported IDFs are arguably less likely to yield inventions that the company plans to build into products.

Further, if the PIM is selective in prioritising product-based inventions, the resulting high rejection rate means that many inventors are asked to spend time completing IDFs that ultimately have limited benefit to the company. Inventors who are frequently rejected may even become disheartened and stop submitting IDFs. When this happens, the patent culture within an organisation can deteriorate and undermine the patent team’s future ability to harvest the best inventions.

To make matters worse, typical IDFs are quite long and try to capture all the information that a PIM believes they could need. Although this seems great for the PIM, it places a significant burden on engineers to create the documentation. If the PIM’s goal is to create a patent portfolio that best serves the company’s objectives, a complimentary goal should be to build the patent portfolio in a way that allows engineers to spend more time building and less time completing forms. This brings us to Principle 3.

Principle 3 – do not waste engineering time

Patent harvesting is an exercise in customer acquisition. In order to harvest the best inventions, it is important to provide a quality experience to inventors so that they become repeat customers. Strong customer service can even reduce the need for financial incentive programmes because giving inventors time back can be more valuable than providing separate compensation for invention disclosures and patent applications.

Although techniques such as passive monitoring and targeted harvesting may require more input from the PIM, overall the company should save time given that engineers will no longer have to complete IDFs. The PIM can start the filing process using the information learned through passive monitoring and speed the inventor straight to the inventor interview with patent drafting counsel. If necessary, the PIM can also collect additional information from the inventor before or at the inventor interview. This is especially important where the existing documentation is silent on potential public disclosures.

Overall, this is a more enjoyable experience for inventors, who would likely rather talk and write on whiteboards than complete IDFs. In addition, passive monitoring rejects inventions without requiring any effort from the inventor – or even notifying them of the rejection. In this way, the perceived rejection rate is much lower, which improves engagement between the engineering and patent teams.

Of course, passive monitoring and targeted harvesting are just two examples of how a PIM can improve engineering engagement through customer service. For example, PIMs who are regularly engaged in passive monitoring should have a strong understanding of the technology at hand and can train patent drafting counsel on the technology before any inventor interviews. Such training leads to more productive inventor interviews and happier inventors, who tend to prefer talking to patent attorneys who already ‘get it’.

Principle 4 – leverage harvesting activities for the benefit of other legal teams

One important takeaway from Principle 3 is that efficiency should be measured by looking at all activities of the company. A PIM that saves time by pushing work onto the engineering organisation is not being efficient. An efficient harvesting programme does not just optimise for the needs of the patent team; instead, it limits the overall time and resources spent by both the patent team and the engineering team.

But why stop there? At Uber ATG, the patent team also serves as a resource to other legal teams that support the company. Thanks to the patent team’s harvesting activity, each PIM has a strong understanding of the organisation’s technology and engineering activities and has developed deep relationships with inventors across the organisation.

The Uber ATG patent team leverages these secondary benefits in several ways. For example, passively monitoring the company’s engineering activity allows the team to spot non-patent issues for other legal team members. If they then need more information from the client about a non-patent issue, the patent team can make the introduction and leverage its relationships with the engineering team to ensure a smooth handover. If necessary, the patent team can even train the other legal team members on the technology at issue. As a result, other legal team members can have more productive conversations with the client, similar to how training helps patent drafting counsel hold more productive inventor interviews.

Leveraging patent harvesting activities to benefit other legal teams also bestows benefits back on the patent team. In particular, if a PIM is responsible for making patent investment decisions based on an invention’s potential for future commercialisation, then they will be more effective if they understand the non-patent issues that can influence whether or how a product reaches market. For example, a patent attorney that understands regulatory issues concerning driver-vehicle interfaces can make better predictions concerning the future commercialisation of a new vehicle touchscreen invention. By participating in conversations about non-patent issues, the patent team can improve its overall ability to make strong patent investment decisions.

Going forward

This article has presented four principles for improving efficiency during the earliest stages of portfolio development – consider these as a collection of tools to improve the overall efficiency of a portfolio. As patent applications change through prosecution and mature into issued patents, additional tools may be used to prioritise patent quality and rightsize the overall portfolio. When these different tools are used together, the end result should be a portfolio full of high-quality, high-relevance assets.