Links

In a previous post, I outlined four ways that patents create value for their owners. I realize that this can be confusing, because patents have become integral to so many different business models and strategies. In fact, there has been such a profusion of IP/patent strategies and schemes and a so many flavors of innovation management, IP management (IPM), and intellectual asset management (IAM), that it is difficult to believe patents create value in only four ways. On the other hand, many people correctly point out that patents only create value in one way: by blocking others from doing something. The question is: do patents create value in one way, four ways, or a whole bunch of ways? In my mind, the answer to this question is more than academic. It really ought to form the basis for how any company or organization deals with intellectual property, and particularly patents.

Luckily, finding the answer to this question is not as difficult as it might appear. We simply have to return to what we know. A friend of mine is a professor of engineering at MIT and an expert in metals and structural failure. He likes to recount a story to his students that illustrates what I mean. On September 11, 2001, as he drove home from work he was pondering why the World Trade Center towers had completely collapsed. His first thought was that he would need to examine and analyze the details of the engineering of the WTC buildings in order to answer this question. But, he quickly remembered that he already knew everything necessary to understand what caused the collapse (and what didn’t cause it). When he got home he wrote out those things (he calls them “fundamentals”) in a list - things like the temperatures at which aircraft fuel burns and steel melts, and the weight of the airplane relative to the weight of the building. The list was surprisingly short, and formed the basis for a paper he wrote that has become the most highly cited on the subject and of a subsequent NOVA program.

My point is that it is easy to become confused about how to think about intellectual property. But a review of the “fundamentals” can be very helpful. So, here is my attempt to return to what we know.

What We Know (the “fundamentals”):

1. MECHANISM: First, we know that intellectual property provides the owner with the opportunity to prevent others from using or doing something in business. This is literally how one goes about enforcing the rights of a patent and is fundamentally a legal process. I like to think of this as the mechanism by which IP delivers value.
2. MOTIVATION: Second, that opportunity can produce different kinds of value for the owner. In other words, there are different benefits that IP owners can receive by preventing someone else from doing something. In my mind, these as the motivations for owning IP. In case you are wondering, this is where my “Four Kinds of Patents Value” fit.
3. MODE: Finally, there are a lot of different contexts, business models, and strategies that these motivations can support or be a part of. To me, these are the modes in which IP owners leverage IP to pursue their business objectives. Although there have been a few traditional modes, there seem to be more and more lately, and they are probably limited only by the human imagination. Incidentally, this is really what most people are referring to when they talk about IP strategy (or “innovation” or “IAM”).

What it Means:

I believe that there are two conclusions from this that are of immediate and tangible value:

First, this clarifies why it is so important to involve different people with different skills, knowledge, and perspectives in IP management. Said another way, this makes it more obvious why companies that bring people from across the organization together to participate in IP management are consistently more successful.

Second, it is essential to remember that the mechanism and motivations of IP management don’t change over time (or, not much anyway), and they are not something managers have any control over. What managers have control over primarily are the strategies or modes they use to leverage IP.

The benefit of this realization is that it allows managers a relatively simple framework for evaluating any IP strategy. To produce value, an IP strategy (mode) must combine enforceable IP (mechanism) with the promise of generating at least one type of value (motivation).

Unfortunately, what I see too often is management that hasn’t identified its specific objectives (motivations). In other words, they haven’t decided specifically what types of value they want from their IP - usually they want to pursue them all. This kind of IP management is expensive and largely pointless mostly because it doesn’t provide any way to decide which IP to pursue (or not to pursue), or for that matter, when, where, and how much. And, isn’t making those decisions mostly what managers do? It is impossible to visualize effective management in this situation.

I also think that the failure to define the core motivations for IP strategy explains why we are seeing such a strong shift towards what Markus Reitzig calls the “Full-fledged IP protection” strategy. He defines this as the pursuit of IP for “every possible minor invention in order to block entire technology spaces”, and his recent research shows that nearly everyone is moving in this direction. In other words, this is what “IP Management” has come to mean to most corporations. I have a friend who calls this strategy, “spray and pray”.

This is not only expensive and unfocused, but it also shifts the competitive efforts increasingly toward speculation and guesswork. It is also one reason why many view IP strategy as a “dark art”. Personally, I am deeply skeptical about this as a viable IP strategy for any company. I strongly favor making informed decisions supported by knowledge, and that kind of IP management takes place where mechanisms, motivations, and modes converge.

Patents create value for their owners in a variety of different ways. One of the biggest problems with nearly all patent valuation techniques is a failure to be explicit about what kind of value they are attempting to measure. It is essential to have an accurate model of value types in order to identify the source of the information that is needed, and also to formulate the correct approach for interpreting that information.

The Four Kinds of Patent Value

Patents basically allow the rights holder to pursue legal measures to prevent others from practicing the subject matter. This is the reason why many people are quick to point out that patents are only valuable in the way that they act as a blocking mechanism, and this is certainly true. However, although it is very important, it is not accurate to say that the value of a patent is purely a function of its enforceability. In practice, there are different ways to use patents to create value, and people are getting more creative about it all the time.

I group the major ways patents create private economic value into four dimensions.

1. Practicing
   
   The first and most obvious method for realizing economic value from patents is by producing and selling products which embody the patent. This is what is known as “practicing” the patent. Value derives from the producer’s ability to exclude competitors and therefore earn monopoly profits. In practice, this type of value is virtually impossible to measure directly particularly because it is nearly impossible to determine exactly how much value is attributable to the patent’s exclusionary rights versus all other factors such as superior design, branding, timing, market power, other intellectual property, and existing manufacturing efficiencies. Another problem is that the value that a company may ascribe to a patent in this context will not normally be the same as the price at which the patent would transact in the market.
2. Licensing
   
   An alternative to practicing a patent is to license its rights to another entity. Licensing value is realized when the licensee pays royalties and other fees to the licensor. Thus, value is created through royalties on the sales of any products that embody the product and therefore can be viewed as a function of the ultimate value from practicing the patent. This provides at least one way to observe directly the value specifically attributable to the patent in its ultimate application, and therefore addresses the problem with measuring practicing value. Licensing transactions also offer the opportunity to observe prices that occur in the market between willing entities. In practice, these are virtually impossible to observe because patent licenses are always confidential.
3. Litigating
   
   An entity that does not practice or license a patent can sometimes receive value by leveraging the special legal status of patents. Recently, some patent holders (often referred to as “patent trolls”) use patents as an asset in a threatened or real patent infringement suit. In this case, the patent holder does not intend to practice the patent, but seeks value primarily through settlements and court awarded damages. The majority of patent disputes are settled out of court and are confidential. But, even if known, settlements and awards cannot represent the “market” for IP as they do not occur between two willing entities and amounts are distorted by legal rather than commercial considerations.
4. Deterring
   
   “Defensive” patenting is the practice of filing patents for the purpose of providing a basis for counter-infringement claims in patent litigation. In a sense, it is the opposite of litigation value. As patent infringement litigation has accelerated, many companies have filed defensive patents to serve as a deterrant by increasing the cost to opponents of asserting patent rights. Obviously, measuring this kind of value is virtually impossible as its economic benefits are unknown except in rare cases.

One of the purposes of this taxonomy is that it helps us to understand the roles that IP plays in business strategy and therefore helps us to construct more coherent and effective IP management practices. Another benefit, is that it provides a way for better understanding patent valuation. Most patent valuation techniques suffer from a lack of clarity on this point. Because patents can create different kinds of value, we have to be clear which kind we are estimating when we do patent valuations otherwise our results will not be valid. In addition, being clear about the kind of value we are estimating makes it obvious what types of data are relevant. In subsequent posts I will show how this model of patent value is essential to almost everything we do in IP management.

NOTE: This is the third and final part of a three-part series. (Part 1 | Part 2)

In order to understand what information we need to observe to derive useful market pricing information for patents, we need to have a basic understanding of the ways that patents create value. Only then can we understand what types of transactions we are interested in observing and the correct way to interpret their meaning. In addition to this, before examining the value of patents it is essential to be clear as to what kind of value we are talking about.

Private Economic Value

Most patent valuation techniques are hampered by a failure to be sufficiently precise on what kind of value they are attempting to measure. The problem partly arises from the almost universal failure to distinguish between the scientific significance of a patent and its economic value. Even those that have made this distinction have failed to adequately distinguish between the private economic value they generate for their owners/licensees and the public economic returns they create for society. To produce the kind of patent valuation metrics described above, it is essential to have information that measures the private economic value patents create.

IP Market Signals

It is also essential to understand the way different actors in the IP supply chain transmit market signals about the value of patent rights. It is the final market for goods and services that ultimately determines the commercial value of patent rights. Therefore, it is only when the products which embody patents are commercialized and sold to final consumers that economic value is established. This insight allows us to eliminate consideration of both litigation and deterrant value in searching for an optic on the market for IP. From this perspective, it is the “Practicing” value that is the most direct measurement of patent’s ultimate value.

However, this is not the specific type of value that we are most interested in observing. Remember that the need described here is for visibility on the market prices for IP. This is because all of our management tools and instruments rely on access to this particular kind of value. Thus we are most interested in the market-clearing price for patent rights. Only this particular definition of patent value can provide the necessary input to enable the adaptation of existing business skills and mechanisms to the ‘idea economy’.

Since market transactions occur between willing and knowledgeable parties, the market-clearing price for IP will also be influenced by the supplier. When the creator of IP is internal to the same organization that commercializes the final product, it is virtually impossible to observe “market” pricing. Therefore, it is only when patent rights are transacted between entities, as in the case of licensing, that we can accurately observe the sythesis of the influence of all of the actors in the IP supply chain.

Finally, in addition to the influence of the actors, market transactions also compound critical information about broader market forces and other external factors such as macroeconomic fluctuations, changes in regulation, the impact of key litigation, and many other influences that bear on the pricing of the transactions. This underlines the point that the most relevant and valuable IP valuation data are transaction prices between parties, or in other words, licensing transactions. The pricing of these market transactions alone reflect the true fusion all economic factors, and therefore, are the correct target of observation for measuring the market value of patents.

Conclusions

Four key conclusions come from this discussion of patent value:

1. Our ability to manage IP is limited by our inability to reliably measure its value.

2. Licensing value is the only type of patent value that can be measured consistently and reliably.

3. Only licensing transactions offer a valid measurement of the distinct economic value attributable to patent rights.

4. Only licensing transactions provide the opportunity to observe the “fair market value” of IP.

Based on this, the most viable solution is to gain access to observe a large number of licensing transactions as they occur and accumulate revenue over time. As noted above, the challenge is that these transactions are strictly confidential, and this is the reason that previous attempts to access this data have been unsuccessful. The need, therefore, is for a solution that provides a way to observe the market for IP transactions (or at least a statistically significant portion) which does not compromise the confidentiality of the transactions. Second, this data must be analyzed in such a way that the results can be accurately generalized to the larger space of patents. If you would like to learn more about this, my research on “Patent Citations and Licensing Value” examines the extent to which this kind of approach could produce meaningful metrics.

In a previous post, I argued that there are a lot of examples where aggregate market data are useful despite the fact that they often obscure a lot of pricing variability associated with heterogeneous products. In other words, market prices are averages of aggregated data that actually obscure a lot of pricing variation. Nevertheless, these data are extremely useful for management, in making investments, and a lot of other areas. My point is that aggregated IP/patent transaction data would be very useful to a lot of people and organizations. Here is just one interesting example of the value of market data from my own background…

In a former career, I worked as an economist analyzing global trade in fruits and vegetables mostly in support of World Bank and USAID projects to develop agriculture in emerging economies. I worked on several teams that built models which were used to focus hundreds of millions of dollars of investments in agricultural production, packing, transport, and marketing of fresh products throughout the world. Our clients had a very simple need. They needed to know what products to produce, how much, and where to sell them to make the most profit. Naturally what they can produce is constrained by their climate and location. But there were always several different products they could produce effectively and it was critical to know which would be the best investment. The data that are available to build market profitability models aggregate data in ways that obscure huge variations in what are actually very heterogeneous products and markets. For instance, for pricing and volume data we would often have a category like, “Grapes: Fresh”. Just think about the variety of grapes that are offered and the different venues, and formats where you buy them. “Grapes: Fresh” doesn’t really capture much variety. Yet, we were able to use this data very effectively to predict the volume of product that our clients could profitably deliver into every market around the world, in any given week. Any grape expert can tell you that the pricing data we were using obscured the very important reality that it matters a lot what kind of grapes you deliver, and in what format, for determining the price. However, this didn’t prevent us from using this data and combining it with that expert knowledge to be able to answer our client’s most important question.

The same is true for IP. Just because a product is not perfectly heterogeneous does not mean that aggregated/averaged market pricing information is not useful. In many cases, this data is essential in establishing fluid global markets. I am convinced that the same will be true for patents and other intellectual property. Once we can finally observe the market prices for large numbers of transactions (licenses and sales), this data will greatly reduce the “friction” associated with these transactions and a fluid global market in IP will finally emerge.

Note: This is the second of a three-part post. (Part 1 | Part 3)

If there is such a great need for pricing information on the global IP market, then why has no one ever bothered to gather it together and analyze it? Unfortunately, this is a difficult question to answer, partly because different people give different reasons. Many understand that the fact that virtually all IP transactions are confidential makes it extremely difficult to gather this information. However, there are also a lot of misconceptions, particularly regarding patents, that are often cited as fundamental reasons why this kind of data wouldn’t be meaningful even if it were available.

One issue that has been widely blamed for the non-emergence of IP markets is the fact that IP assets are all unique by definition, and therefore transaction prices cannot be comparable. However, the attributes that distinguish patents within a definable group do not appear to have a distortive influence on the dimension of market value. Although patents within a specific technical area are all scientifically distinct in some dimensions, it is not evident that these differences are the driving factor in determining the market price. It is important to remember that what makes patents different scientifically is very different than what makes them different economically. The scientific and economic dimensions are less interconnected than most people think, and this is a major source of confusion about valuation.

More important, however, is this argument misunderstands the role of market data in many if not most contexts in which it is successfully used. First, it falsely assumes that other types of assets or products being traded in market are truly identical, or at least that their differences have a smaller impact on their economic value. In fact, there are very few products and markets where the products are truly homogeneous (sometimes thought of as “commodities”). Many things that appear to be commodities actually exhibit considerable variance along dimensions that have meaningful impacts on price.

Most market pricing data are aggregated, reported, and used in ways that obscure the impact of variability in attributes on pricing. An example of this is an automobile blue book. If you look-up the price for a particular car model and year you will see a high value (implying a car that has marginally higher than average value - usually associated with better condition or features) a low value (implying worse condition or features) and a mid point (implying typical condition and features). In reality, these numbers obscure the variability associated with these and a lot of other attributes such as where the transaction takes place, and even what color the car is. Obviously some variation in price is attributable to things like the color which are not visible in the market data (the blue book). Nevertheless this data is used to successfully enable what has become a very fluid market in used cars.

Looking at other types of assets such as real estate, industrial equipment, and labor reveals how much more variability is associated with other markets. Yet, the existence of the variability does not eliminate the usefulness of market data in negotiating transactions and providing the key input for sophisticated financing instruments such as securitization, insurance, and more. In fact, most market data that is used as the basis of a wide range of industries is aggregated, averaged, or summarized in ways that obscure significant variations in the attributes of the assets or products. Market pricing data are almost never generated in a context of perfect comparability. For this argument to be valid, one would need to demonstrate that the dimension of variation in the attributes within other markets has a proportionately smaller impact on aggregate market pricing data than it does in patents. However, there is really no evidence to support this.

Another version of this argument is that because patents are scientifically unique, only one or a few companies will be in a position to commercialize them. Therefore, as the argument goes, patents cannot exhibit a “market value” as the market for any given patent is too small to be generalizable. However, this demonstrates a misunderstanding of the concept of market value. Even if an asset has only one buyer, it will have a market value at which it will trade hands. In this scenario, the market may be small but this will only have the effect of giving the buyer more bargaining power in the negotiation. For all of these reasons, we can reject the argument that the dissimilarity of patents prevents their economic comparability.

Further evidence supporting this view is that there actually have been many efforts to aggregate comparables for IP transactions over the years, some of which are embodied in commercial services. A substantial amount of effort has also been invested in alternative ways of approximating market value. In a way, the legitimacy and comparability of the data have already been defined by the demand for it in specific contexts. This is not surprising as it has actually been the case in other areas where market data are now used actively. Ultimately, the valid interpretation of the data will always be defined by the use case and the need, and as I have already argued, there are a lot of valid potential uses for IP market pricing data.