The brokered patent market in 2021
A decade ago, we contacted the editor of IAM to ask if there would be interest in publishing a piece on a whopping 186 patent-buying opportunities covering 7,500 patent assets. The goal was to shed light on an opaque industry, bring transparency and promote the idea of buying and selling patents as a strategic risk-mitigation tool.
Back then, nothing like this had ever been published before in our industry. You are now reading our 10th such report.
Over the years we have expanded our data to include more than 13,000 buying opportunities, covering 250,000 patent assets. We have also used the data to make some observations about the patent world and even helped a few corporations to buy patents (see “How and why LinkedIn learned to learn patents”). To date, we have used this data to assist dozens of companies buy patents, provided expert testimony on patent market prices, supported academics in their research (An Empirical Look at the “Brokered” Market for Patents, Love et al SSRN-id3056444) and enabled defensive aggregators to better model their return on investment.
Economists apply this data to better understand innovation – and we have featured in The Economist. While we cannot predict where the next 10 years will take us, we anticipate that it will be at least as interesting. To you, our readers, some fans and some critics: we thank you. Without your opinions and feedback, we would not be able to produce such a comprehensive report. Now, on to the fun stuff.
This year we have included not only the latest market year information, but some comparisons to past years and thoughts on what has changed. For 2022, the big changes are:
- NPEs dominated patent purchasing during the pandemic, but operating company buying is resurging.
- Prices have declined, but not noticeably - although there is a clear fall over the 10 years we have been reporting.
- Some brokers are now offering licensing options, in addition to the traditional sale.
- You, yes you, are now the primary target of patent sellers. The most frequent word arising in patent summaries this year is ‘user’. Importantly, this is the first year that we have not seen a company as the primary target of patent sellers.
- Packages with Alice-affected technologies are selling at significantly higher rates than the rest of the market.
We will split the data up in multiple ways and analyse specific discounts to predict closing prices and actual dollars transacted. We will also discuss the impact of covid-19 on the market as, like everything else, the secondary patent market has taken a hit from the pandemic.
Figure 1. Cumulative sum of asking prices – brokered and tracked private market ($ billion)
Patent buying and selling is big; this analysis covers only a slice of that market. Many of the wide selection of assets available for purchase have transacted. Figure 1 shows the cumulative asking price of all of the assets that we track, which currently stands at about $38 billion. We have created tools to parse the assignment records for over 250,000 tracked assets and have identified sales of these amounting to over $10 billion in asking prices. Projecting through to the Q2 2021, we expect cumulative total sales to reach $11 billion. This is the sum of asking prices for the sold packages without adjustments or discounts to estimate closing prices.
This analysis generally follows the flow of a typical purchase process – including sourcing, asking prices, diligence steps, purchase closing, litigation, and then concludes with our estimate of the market size. We analyse the 2021 brokered market and provide insights based on 10 years of market tracking data, ending with our estimate of the current market size.
The shift from quasi-public/broker-assisted deals to private deals has been a constant over the past decade. We view this as a sign of a maturing market where more in-house teams are capable of closing their own deals. However, the skills of a broker remain important, especially for buyers and sellers without internal corporate processes for managing and marketing this type of transaction.
Brokers’ skills include:
- experience in identifying assets of value;
- setting buyer and seller expectations; and
- sales acumen.
Why work with a broker? Consider the advice from one of our law professors: “Yes, sure, you don’t have to be a doctor. You are smart enough to set your own leg. Why would you want to? Hire a professional.”
Brokers leverage their networks of connections to consistently find the most diverse sets of patents for sale from a variety of sellers – and they know who is buying. They are able to align the expectations of both buyers and sellers, which is one of the most difficult hurdles in getting deals done. Without these skills, many patent packages would never be reviewed in the first place and would wither on the vine.
Over the past decade, brokers have had to adapt to the marketplace. They have diversified their services and now offer licence options that were unthinkable for many buyers five years ago. Brokers sell through multiple channels that simply did not exist 10 years ago - such as fixed-price patent auction programmes - and will also assist in private transactions that are not offered to buyers at large. As the market evolves and becomes more complex, brokers are generally well-suited to adapt; but not all of them survive the changes.
Brokers with five or more packages
During the past decade, the number of brokers listing at least five packages for sale has fluctuated. The peak occurred in 2016, with 28 brokers listing five or more assets (73 total brokers), but this has since dropped significantly down to 17 in 2021. This seems to have been the new normal for the last four years (see Figure 2).
Figure 2. Number of brokers listing packages
The total number of brokers that listed any number of packages for 2021 fell to 40 (in 2020, it was 41). (Note: the 2021 market year is 1 June 2020 to 31 May 2021.) The ‘for sale by owner’ category has performed consistently over the past 10 years and in 2020 rose to a high of 6.8% of all listings. However, in the 2021 market, this plummeted to 1.1%, likely due to decreased sales activity for operating companies during the pandemic. We will continue to monitor this segment to ascertain whether it rebounds in the coming years.
Additionally, over the last decade, established brokers have continued to become more dominant, with a higher concentration of the packages attributed to them. In the 2016 and 2017 markets, about 20% of all brokers brought 10 or more packages to market, whereas now it is 30%. In 2021, 12 of the 40 brokers took 10 or more packages to market (12 of 41 last year). Table 1 lists the brokers who proffered five or more packages to market during the 2021 market year. These accounted for 91% of all packages listed by brokers (a slight increase from 89% last year).
Table 1. Brokers listing five or more packages – 2021 market year
|Adapt IP Ventures|
|Darwin IP Advisors|
|Dynamic IP Deals|
|Morgan, Lewis & Bockius|
|Red Chalk Group|
|Rui Zhi Ventures|
Figure 3 illustrates the success of individual brokers (each represented by a dot). The X-axis displays the number of packages brought to the market during the measurement period. We used the five-year period of 2016 to 2020 calendar years for this analysis, in order to provide a better sense of overall sales rates. The Y-axis represents the sales rate or closing rate for that broker (the percentage of packages sold). Positions up and to the right in the figure indicate brokers that are doing better. As shown in Figure 3, some brokers were particularly successful (green circles) or unsuccessful (blue circles). While we expect these numbers to increase slightly as 2020 calendar year listings have more time to sell, overall brokered packages listed from 2016 to 2020 have sold 21% of the time (25% for 2016 to 2019).
Figure 3. 2016-2020 broker sales rates by number of listed packages
Brokers in the green circles are doing better, brokers in the blue circle are experiencing challenges
Total sales in a given year
With the benefit of 10 years of listings and sales data, we have adopted a new approach to exploring the brokered market sales each year. We attempted to estimate the amount transacted each calendar year by taking into account variable discounts from asking prices to sale prices, depending on the amount of time that a package was on the market. To do this, we first looked at the number of packages that transacted each calendar year. The upper portion of Figure 4 reveals the total number of packages transacted per annum (Note: we possessed only partial market data for 2011 to 2013 and sales were identified through the USPTO assignment database as of 20 August 2021).
To determine the estimated dollars transacted (lower portion of Figure 4), we investigated how long each sold package was on the market prior to sale and, depending on that interval, discounted the asking price from anywhere between 25% and 55%. This variable discount is more nuanced than the flat discount of 30% to 35% we have previously used. We built the variable discount rates based on our experiences buying for our clients, as well as in market surveys.
Compared to Figure 1, the transacted dollars are significantly lower as Figure 4 represents the brokered market and therefore does not include private and large transactions. Additionally, the discount was then applied to reduce the asking prices totalled in Figure 1 to estimated transaction prices.
With the covid-19 pandemic, we expected to observe a significant fall in both the volume of packages transacted and the total dollars transacted in the 2020 and first half of the 2021 calendar years. This did occur; but notably the drop in transacted dollars was more significant than the decrease in package volume. The packages that transacted on the brokered market tended to be smaller and were on the market longer. It is possible that the lower-priced deals were easier to close because they typically require less internal approval on both the buy and sell sides. Despite the market taking a hit, covid-19 did not wipe out the market completely during this period; packages still transacted.
Figure 4. Actual sales by calendar year of sale
Grey shading indicates partial data for the year
Other market opportunities
Since we started tracking the market in 2011, new ways to buy or sell patents have continued to emerge. Some methods combine the current skills of brokers and transaction platforms, while others follow completely new models. The following are the most noteworthy:
- IAM Market is a market provided by IAM, where sellers can list their patents and anyone can browse the packages, contact the sellers and close patent purchases.
- IP3 by AST is a fast-close patent buying programme. Sellers list their assets for a set price and AST member companies decide whether to purchase, all on an accelerated schedule. AST continues to attempt new models for buying patents and we expect additional leadership from them here.
- Brokers are offering more and more options to ‘license if you don’t want to buy’. At least three of the top brokers are presenting such options.
- RPX has switched to almost exclusively obtaining licences for its members, rather than buying the patents outright.
This report addresses brokered market patent transactions. Thus, licences and private buying programmes such as IP3 are excluded from our dataset. However, it is important to keep in mind the continuing diversity in the market.
Packages listed on the brokered market in the 2021 market year rose by 17% (Table 2), rebounding from the 14% drop in 2020. The number of US-issued assets listed similarly increased, though the total quantity of assets decreased as packages included significantly fewer international assets. Overall, the volume of 467 packages in 2021 was greater than either the 2019 or 2020 market, but lower than other market years. Unsurprisingly, 2016 was the high watermark for listings, with 650 packages listed. However, both the increase in private transactions since 2016 and larger deals (excluded from the base dataset as they are harder to compare) have picked up much of the apparent fall in the number of packages listed.
Overall, the market continues to provide buying opportunities in a diverse range of technologies, affecting various products and focus companies. There are assets available to fill business needs in almost any high-tech category. When we receive a package, we review the materials, along with any assets highlighted by the seller, to categorise it according to our taxonomy of technical areas. We have developed a two-tiered classification taxonomy with 17 general technical categories and 108 sub-categories. We continue to modify this taxonomy as new technologies arrive on the brokered market and supplement them with machine-learning classifications.
Table 2. Brokered patent market contents
|Market year||2021||2020||2020-2021 % change|
Figure 5 shows that software packages once again comprise the largest portion of the market. With the exception of the 2020 market year, software packages have accounted for the plurality of the market, with a high point of 53% in 2014. Over the last 10 years, software packages made up 41% of all packages. This is unsurprising, as software packages usually sell better than other categories.
Figure 5. Packages distribution by technology group
Additionally, we have seen a consistent rise in the ‘other technology’ category from around 7% of the market in 2014 to 32% in 2021.
This is due to:
- automotive, energy and medical device packages increasing; and
- several brokers substantially increasing their volume of ‘other technology’ packages.
We continue to monitor the ‘other technology’ category, but as the packages in this space have experienced low sales rates, we do not believe that these packages are contributing significantly enough to the market to place them in new technology categories.
As mentioned in the introduction, this is the first year that we have not seen a company as the primary target of brokered packages. Users are now the primary target. The word cloud in Figure 6 provides a way of visualising the focus of the brokered patent market. The relative sizes of the words highlight the hot companies, technologies and products-identified marketing materials provided from the broker or seller, especially any evidence of use (EOU) materials (these are like claim charts, but usually less detailed).
The biggest technology companies (Google, Apple, Microsoft and increasingly Amazon) continue to be the favourite targets of patent sellers’ EOU materials. In the 2021 market year, ‘user’ became the most frequent word because of an increase in EOUs targeting consumer software over business and backend.
Table 3 displays the most common words by year for the past 10 years. If your company is in the mobile space, you likely have busy litigation and risk mitigation teams.
Package size (see Figure 7) continues to be one of the most consistent factors in predicting pricing and other deal dynamics. Prices per asset decline for larger packages, but these deals have higher close rates (at least for some assets). Year on year, we discerned little change in the distribution of package size, with a slight drop in single-asset packages.
Figure 6. Word cloud of most frequent words in marketing materials and EOUs
Figure 7. Distribution of packages sizes (total assets)
Table 3. Most common word in marketing materials and EOUs by year
|Year||Most common word|
Table 4. Asking prices in the 2021 market
|Asking price||Per asset||Per US issued||Per family|
To get a deal over the line, you must settle on a price. In over 10 years of market tracking, pricing has become more consistent because of the maturation of the market and the increased concentration of listings by knowledgeable brokers. Pricing analysis was an original driving force behind the first brokered patent market article and continues to be the most intensely discussed topic. Our goal is to boost transparency in pricing. Some people have applauded this, while others have viewed it as the worst thing to have happened to patents. Our opinion is that the lack of pricing transparency has historically been one of the largest sources of market inefficiencies and, like all inefficiencies, has benefitted a few individuals while hindering the market as a whole.
Pricing is complex in every industry; if knowing the numbers alone was sufficient, entire aspects of product marketing would disappear. If sellers are asking for more than typical market prices, they must be able to articulate why. Similarly, buyers ought to be able to explain why they may be offering a lower price. The numbers in Table 4 should be treated as a starting point to help guide a pricing discussion, not to dictate a final number.
We recommend referencing both average and median pricing first, before making some adjustments for specific factors related to the deal – while recognising that there is a long-tail distribution to pricing histograms, with pricing extending far out on the high end. Understanding what the characteristics of a typical patent are, and why your patents might differ, helps you to adjust the price from general statistics.
We have helped clients buy patents over a wide range of prices, from more than 10 times the median price per asset down to one-tenth of the median. In each case, we regard the prices as justified and, importantly, the deviations from the median were supported by deal-specific factors.
As for year-over-year trends in pricing, Figure 8 reveals the asking price per asset for the last eight market years. There has certainly been a drop in asking prices since 2014, but the fall-off is not as precipitous as some claim; patents continue to hold value. This dip in prices is even less pronounced if the dataset is limited to sold assets; the prices of packages that actually sold tend to be more consistent year over year. Some asking prices remain aspirational, but the deals that attract buyers tend to be more in line with market prices. The variance in pricing over the last decade has diminished significantly.
With reference to the 2021 market year, the pandemic continues to contribute to uncertainty. The average asking price per asset dropped 10% from $162,000 to $146,000, but the median increased by 1% from $107,000 to $108,000. Understanding the distribution of the asking prices explains these seemingly incongruous statistics. The distribution became more consistent, with a decrease in the highest asking prices (fewer aspirational high-side asking prices). This affected the average, but did not move the median (midpoint of the curve). It is important to remember that these are asking prices.
We also conduct other independent studies of closing prices and observe closing prices discounted between 30% and 35% off these asking prices. However, the discounts start low and increase the longer the assets remain on the market. So, if a deal is new, you can expect almost no discount, while older deals may allow for much higher discounts. For that reason, we have changed our procedure for estimating transaction prices by discounting from asking prices to be a variable discount, rather than a fixed discount (see above).
Figure 8. Asking prices by market year
Average - top and bottom 5% removed
Per-asset pricing by package size
Figure 9 shows the price per asset compared to the package size. Packages with more assets mean lower prices per asset. Overall, in the 2021 market year, we notice no significant changes from previous years. Keep in mind that the package size is one of the easiest and best predictors for the asking prices of patents.
Figure 10 presents the distribution of package asking prices in the 2020 and 2021 market years. There is a clear preference for deals in the $500,000 to $1 million range. Above this range, signing authority at companies starts to cap out (“I have to get permission from the CEO if I go over $2 million”), and people become less comfortable closing high-priced deals.
Figure 9. Per asset price by package size
Figure 10. Distribution of package asking price (top and bottom 5% removed)
Packages with pricing guidance
Since pricing is so significant when it comes to closing a deal, lack of pricing guidance is one of the most frustrating aspects of the brokered market. Imagine shopping for a home without listing prices? You would waste a lot of time looking at homes that you might not have otherwise considered, especially if you are a less experienced buyer and do not have a sense of which neighbourhoods are particularly expensive. This is also true for patent sales.
In the 2021 market year, only 47% of packages came with pricing guidance of any kind. We accept ‘market pricing’ as an answer to the pricing guidance question, but it is at best a partial one. The number of sellers who asked for market price was 5%, 22% gave a price range (eg, mid six figures) and 20% of sellers provided exact numbers for pricing guidance in the 2021 market. Packages with exact pricing go to the top of the list for some buyers.
How does supplying more detailed pricing translate into sales? Over 10 years of data, packages with pricing guidance sell 73% more often than packages without. The quantity of packages with pricing guidance that have sold is 30%, against 18% of packages without pricing guidance. These numbers are even better when limited to packages with asking prices in the near market price. Not only do these sell 79% more frequently, but they also sell faster. For these reasons, it is baffling that many packages are still accompanied by no pricing guidance.
Asking price by tech category
Technology categories continue to drive asking price variations, but this signal for specific tech areas is often buried in the noise of other factors. For the more specific technology areas, we use a normalisation procedure to calculate the impact on pricing. For more detail, please see the “2019 Brokered Market Report” in IAM 98. More generally, the major tech groupings have moved closer together in pricing over the past decade. Software-related packages have consistently had the highest asking prices, which was also the case in the 2021 market year. Notably, communication (eg, wireless and networking equipment) packages were priced significantly below the other categories in 2021. This was likewise true of 2019, though communications packages were more in line with the other categories in the 2020 market.
Tools and processes used to analyse the data
As the brokered patent market has matured, access to data has increased. However, the market remains fairly opaque. Therefore, to analyse the market we pull data from many sources, combining this with a proprietary set of tools that we have designed in-house.
Our data sources include our proprietary patent package database, the USPTO patent data (Public-Pair), the USPTO Assignment database, Cipher, Derwent Innovation and litigation data from DocketNavigator.
This data is then combined on both a per-patent and per-package basis, using tools that we have developed over the past six years. The result is a proprietary database of hundreds of thousands of records across nearly 500 fields. These tools are programmed in SQL, R, Ruby, Python, AppleScript and VBA, using ODBC to retrieve up-to-the-minute live data from our database. We also use business intelligence tools such as Tableau. We continue to expand our capabilities to sort, sift and visualise the data.
We also internally track asking prices, bidding dates and clients’ specific diligence decisions, and maintain a list of unique entities that are buying and selling with standardised names. We classify these entities by entity type, meaning that we have our own internal list of companies that we believe to be NPEs. Although this process is quite time-consuming, we believe that using real data to back up our conclusion is the best means of providing accurate analyses to our clients and lowering the barrier to entry for companies joining the market.
Figure 11. Sales rate by years from package listing (2015-2019 listings)
Asking price and impact of an EOU
The median and average asking prices are merely a starting point for pricing discussions. How do you determine and demonstrate that a specific package should move off that starting point and demand a higher price? It is effective to show that the technology is in current use by providing an EOU document.
Overall, the percentage of packages with an EOU has dropped slightly this year to 33% (down from 39%), and these packages continue to see price premiums. The exact premium for an EOU varies greatly by package size and is another opportunity to dig into normalised data. However, over the 2021 market year, we observed a 13% price premium for packages with seller-supplied EOUs over those without EOUs. This varies year over year; but, for all data from 2014 onwards, we observe a 28% price premium. We also know that sales rates are higher for deals with EOUs (see below), so the value of preparing EOUs is substantial.
This year, we continue to detect some decline in total transaction volume on the brokered market (see Figure 3). Over the last decade, brokered sales have comprised a smaller overall portion of the patent market for buying, selling and licensing patents, while private sales (not included in Figure 3) have increased. Sales of older packages continue, including some that are five-plus years old. Generally, highly sought-after packages move fast, followed by a long tail of additional sales, particularly to NPEs.
We started tracking sales data because we did not want to present buying opportunities to our clients that were no longer available. The actual management of this process is deceptively difficult. It is quite easy to check the assignment records for one US asset; however, there are multiple assets per package, internal transfers (eg, “Gizmo Corp” to “Gizmo Holdings Corp”) that do not represent a sale or ownership change, and the analysis must be undertaken for thousands of packages over a decade.
To manage this process, we wrote code to parse the USPTO assignment data and suggest potential sales, before performing a human review of these records. Our methodology considers a package to have sold if at least one patent in it is found to have an assignment corresponding to a sale. We then used the execution date for the earliest transacted patent in the package as the date of the sale (inclusion is limited to packages received by 31 May 2021 and to sales recorded with the USPTO by 20 August 2021). When discussing sales, we change our analysis to a calendar year view. This sample set includes packages that were analysed in previous reports and dates back to packages listed as early as 2011. It is worth noting that some sales may have closed but might not yet be recorded, namely for the partial data in the 2021 calendar year.
Similarly, packages that have recently hit the market have had less time to sell and are expected to have lower sales rates – thus they are not good predictors of sales rates. Our sales rate for 2020 calendar year listings stands at 6.7%, only a slight drop from the 2019 calendar year rate (pre-covid-19) of 8.5%, analysed with the same relative time on market. Additionally, we observed more older – discounted – packages sell in 2020 than in previous years, so this may have depressed the sales rate for 2020 listings.
With regard to those buying patents: operating companies appear to be making a comeback in 2021 (sales recorded before 20 August 2021) and, for the first time, were the buyers of more than half of the transacted packages. This is a massive rebound in operating company buying from 33% in 2020, which appears to have been heavily depressed due to the freeze on corporate budgets at the beginning of the pandemic.
To determine how well packages sell over time (Figure 11), we plotted the sales rate at different points in time after listing for all packages in the calendar years 2015 to 2019. Each calendar year has a datapoint for every year the packages were on the market (eg, sales rate for packages listed in 2016 after a year on the market). The sales rate caps out around 28% after six years on the market, but for most years it is around 25% after three years. The notable exception to this are 2019 listings, the darkest blue dots, with sales rates below the other years, probably because of depressed buying during the pandemic.
Note: our methodology necessarily underrepresents the overall value of the market. First, if the buyer does not record the transaction, the package will never have been sold, and there are often delays in recording transactions. Second, increasingly sellers are offering licence deals where no sale is recorded and those transactions will never appear in this chart.
Sales by receipt date
Buyers have an advantage when they act quickly. If you go antiquing on Saturday morning, you can have your pick of the lot: all of the options and all of the best deals are available. You may still find gems on Sunday afternoon, but it is harder. Similarly, on the brokered market we regularly note sales of older packages, but we have seen multiple clients lose out on deals because diligence took longer than the couple of months the deal was on the market. If you can act quickly, do so.
We know that corporate decisions often require multiple levels of approval, all of which take time. So how fast is fast enough? We analysed how quickly the sold packages listed in the 2020 calendar year transacted to estimate how much time was available for buyers to bid. Lower buying activity in the first half of 2020 meant that packages remained on the market slightly longer than in the past. Yet 60% of the sales from the 2020 listings occurred within eight months of receiving the package and 30% of sales happened within four months. It is likely that these sales were already in the process of closing a month earlier. Even more challenging, if it looks like a high-quality package, the package sells fast. Our advice is to make quick decisions on the hot packages by establishing effective processes to accelerate decision making on packages of interest.
We also embraced the opportunity to dig deeper into the full 10 years of package data in order to locate some trends for when packages sell that cannot be found by analysing one year alone. First, we verified the standard economic principle that lower prices seem to induce higher demand. Of packages that sold, those with lower pricing expectations sold faster. Additionally, highlighting packages’ value increases demand. Packages with EOUs also sold more quickly than those without EOUs. Therefore, you need to operationalise your buying programme and make your decisions quickly to secure the best opportunities.
Figure 12. Sales rate by EOU provided
Figure 13. Percentage difference between Alice-affected sales rate and total market sales rate
Why do some packages sell after multiple years on the market? It transpires that buyers can still close deals years after packages were originally listed. Certainly, nothing incentivises sellers with high pricing expectations to lower their prices like a few years with no interest from buyers. As previously discussed, the longer packages are on the market, the steeper the discount off the asking prices. In some cases, this markdown means that the value to the buyer finally exceeds the proposed prices.
The second reason concerns technology adoption. We know from market adoption curves that it takes years for new technology to have significant market adoption. Sometimes bleeding-edge technology comes up for sale before it is clear if it will be adopted. It is not possible for a broker or seller to create EOUs for these assets, because they are simply too new. Without a clear market use case, it is difficult for a buyer to obtain approval for these packages. Once market adoption occurs, the value proposition changes, especially for NPEs who want to assert an asset shortly after purchase. Packages with EOUs already have adoption, which could explain why they usually sell faster.
Sales when an EOU is provided
Not only do assets with claim charts, an EOU or an indication of use sell faster; they add value to a package. This directly translates into higher asking prices (and sales prices), as discussed in the pricing section. However, there is a third advantage to providing EOUs: these packages also sell more often. Figure 12 shows the sales rate for packages listed by calendar year, and packages with EOUs always had higher sales rates. Across listing years from 2014 to 2020, packages with EOUs experienced an 82% higher sales rate.
By combining the 82% increased likelihood of sale and the 28% price premium associated with having EOUs for packages listed in 2014 and later, the expected value of a package with an EOU is 133% greater than those without (2.33 times the value). The bottom line on providing EOUs is that, if you can build them, the expected additional value more than pays for the costs of doing so.
Table 5. Repeat sellers (sold in 2020 or 2021)
|Acer Inc||Harris Corporation||Panasonic|
|AJOU University||Hewlett Packard Enterprise (HPe)||PARC (a Xerox subsidiary)|
|Allied Inventors||Hewlett Packard Inc (HP Inc)||Sisvel International|
|Allied Security Trust (AST)||Himax Technologies||SK Planet|
|AT&T||IoT and M2M Technologies LLC||Technicolor|
|Collective Dynamics||KT Corporation||Telia Company|
|Daimler AG||Kudrollis Software Inventions||Xerox|
|Flextronics International||MTEK Vision Co||ZTE Corporation|
Life after Alice
The turmoil from the 2014 Alice decision has subsided and Alice-affected software and financial packages continue to thrive. Packages containing affected technologies have consistently sold at higher rates than the overall market, as disclosed in Figure 13. The amount by which Alice-affected packages have outperformed the overall market had increased since the decision and seems to have stabilised. Data from 2019-2021 illustrates that Alice-related sales are 40% more likely to occur than packages in the overall market. It appears that buyers and sellers have worked out what to look for in patents in Alice-affected technologies.
Tracking the behaviours of sellers – both individually and in the aggregate – can optimise buying activities. Knowing the types of assets on the market, and who is willing to sell them, enables faster review of listings. Strong knowledge of the market also provides opportunities for direct contact to sellers to transact private deals. This is especially true when interacting with repeat sellers, as these accounted for 54% of the package transactions in calendar years 2020 and 2021 (see Table 5). Keeping track of sellers’ listings, package sizes and asking prices can also assist negotiations: it is an advantage to know their negotiation parameters before sitting down at the table.
Similarly, for sellers, it is important to make it known that you are selling. Listing packages on your website, listing on the IAM Market, sending targeted email blasts and working with brokers all help to attract buyers. This approach promises success, because buyers may come to you rather than you spending time and effort finding them.
Figure 14. Distribution of seller type (sales year 2020-2021)
To analyse current sellers and buyers, we studied transacted packages received at any time up to 31 May 2021, and sold after 1 January 2020 (assignments were last checked on 20 August 2021). Figure 14 demonstrates that 70% of the sold packages in this timeframe came from operating companies – up from 63% last year. This was not particularly surprising, as operating companies file the largest number of patents.
Operating companies buying patents to mitigate risk may benefit from monitoring sellers, to ascertain who is instigating sales before those sellers fully ramp up their sales programmes. Ask yourself: “How can I mitigate the risk of these patents without purchasing them?” Consider licensing. As discussed above, some brokers are offering licensing options for this reason. Early licensing of seller portfolios may be significantly less expensive than obtaining licences after the assets have sold to assertion entities. Additionally, solutions including the LOT Network may help to mitigate NPE risk across companies that currently have no intention of selling assets.
Figure 15. Distribution of buyer type by sales year
It is equally important for sellers to track buyers. However, it is less obvious that buyers should also be tracking other buyers. Yet doing so can provide the advantage of knowing whether any of your competitors or other companies of concern are actively buying in the market.
How might the in-house teams at corporations view the market? Returning to the analysis of sold packages, you can see that your world is changing rapidly (as shown in Figure 15). Corporate buying pulled back significantly in 2020 to 53%, leaving the majority of package purchases to NPEs. However, operating companies account for 52% of the transactions we have observed in 2021. Defensive aggregator buying remains down, at 11% in 2021 and 10% in 2020, against 22% in 2019.
From 1 January 2020 to 5 August 2021, 107 buyers purchased 178 packages. Of this, 29 buyers acquired multiple packages (see Table 6), accounting for 56% of the package transactions. Last year’s analysis had 24 repeat buyers. As with our overall analysis, the buyer and seller analysis includes only the brokered patent market and does not include private purchases or the purchases from Provenance IP Group.
Table 6. Repeat buyers (sold in 2020 or 2021)
|Acacia Research Group||Helios Streaming, LLC|
|Allied Security Trust (AST)||Intel Corporation|
|Arden Innovations LLC||Invincible IP LLC|
|Beijing Xiaomi Mobile Software Co||Magnolia Licensing LLC|
|Bright Machines, Inc||Majandro LLC|
|Capacitant Innovations LLC||Maplebear Inc (Instacart)|
|Commscope Technologies LLC||Mila Co, Ltd|
|Crown Electrokinetics Corp||Mindset Licensing LLC|
|Drnc Holdings, Inc||Patent Asset Management Advisors, LLC|
|Emergent Mobile LLC||RPX Corporation|
|Etsy, Inc||Stripe, Inc|
|Golden Eye Technologies LLC||Uber Technologies, Inc|
|Google LLC||Valtrus Innovations Ltd|
The brokered patent market represents a large pool of litigation risk. This risk pertains to any package listed for sale as the seller is attempting to monetise its assets, but the risk is especially high for sold packages. Generally, the likelihood of a patent at random to be asserted is quite low. Mark Lemley opined: “At most only about two percent of all patents are ever litigated, and less than two-tenths of one percent of all issued patents actually go to court” (Rational Ignorance at the Patent Office SSRN-id261400). Yet 5.2% of listed packages are litigated after the listing date (see Figure 16). Out of all packages, 4.4% were litigated in multiple cases, with 2.5% being moderately litigated (two to nine litigations) and 1.9% being highly litigated in 10 or more cases. Assets do not need to be sold for a litigation to occur: a seller may choose to litigate while waiting for the listed assets to sell.
The litigation rate after a package is sold grows to 19% and rises to 40% if the package is sold to an NPE (Figure 17). Not only do the packages sold to NPEs have a high litigation rate, but additionally the vast majority were litigated multiple times, with 19% of NPE sales being moderately litigated and 15% highly litigated after the sale. Though, we did wonder why this number is not even higher. It would be reasonable to assume that an NPE is purchasing assets for the sole purpose of asserting them – likely through licensing/litigation campaigns against multiple targets. So, how are the assets from 60% of these packages not litigated? Some assets are more effectively asserted in licensing discussions and others are more efficient in litigation, so some of these packages may be purchased with that in mind. Alternatively, they may have been purchased to round out the licensing portfolio and might not be the crowned jewels of the assertion campaign. Lastly, these packages may represent the potential risk of future litigations. Either way, for both listed and sold packages, the general trend is that the more time has passed, the greater the litigation rates.
Figure 16. Litigation rate after listing by listing year
Figure 17. Litigation rate after sale to NPE by sale year
When focusing on the specific cases, we have found that assets on the brokered market have taken some fascinating paths. NPEs have bought assets from defensive aggregators and then asserted those assets against unlicensed companies. Moreover, assets originally sold to NPEs have ended up in the hands of operating companies (likely due to a non-litigated assertion), who subsequently used those same assets in counter-assertions against other operating companies. Once assets are listed on the brokered market, and especially once they sell, they become part of a pool of assets that are in active monetisation. Regardless of whether you are a patent buyer, monitoring who is buying patents is crucial to your patent risk mitigation strategy.
Full market size
We estimate the 2021 market size to be $166 million. This is not comparable to the predictions in previous reports, as we have completely changed our methodology for calculations. When applying the new methodology to the recent past, we obtain a comparable market calculation of $197 million for the 2020 market year and $211 million in the 2019 market year.
To reach this estimate, we used the observed sales that occurred in the 2021 market year and applied a variable discount rate of 25% to 55%, depending on the time it took for the package to sell. This is the same methodology used to generate Figure 4 but applied to market years rather than calendar years. If no pricing guidance was provided, the average asking price per asset for the market year of the sale was multiplied by the number of assets to determine the expected asking price, before applying the appropriate discount. We relied on average pricing rather than the median to model the market size, because we were modelling across many independent events instead of identifying a typical instance. Nonetheless, we still had to make adjustments for a few large deals.
We then used the market size to back-calculate the number of working brokers supported by the market. By applying an average commission rate of 20%, the revenue from this market for brokers amounted to $33 million. After estimating the average loaded labour rate per broker ($300,000 a year), we calculated that there are 110 full-time equivalent employees working as brokers. Assuming that three brokers work in each brokerage, this results in approximately 37 brokerages. Our data shows 40 brokerages that listed packages in the 2021 market.
Opportunities, insights and reflections
Over the past 10 years, the patent market has matured. Rather than the 186 buying opportunities we saw a decade ago, we now monitor over 13,000 buying opportunities covering more than 250,000 patent assets. There are regular buyers, sellers and brokers who know what a reasonable deal looks like and how to close it. This market continues to gain popularity with in-house attorneys as a place to buy and sell patents to further their business strategies.
Yet there are still inefficiencies that need to be improved. While more sophisticated market participants understand that patent pricing has a long-tail distribution and many factors can move a deal up or down in price, we still notice some inexperienced sellers asking for aspirational prices with little to explain why. Similarly, we have seen buyers say: “Our last deal was $X per patent so we can’t pay more than that.” For this reason, we continue to push for greater transparency, not only in public reports such as this, but also in our private reports where we try to bring more people round to the idea that sharing information helps them more than it hurts.
Additionally, the patent market has evolved to the point where it is not only a place to buy and sell assets. Buyers are licensing through brokers, which was unheard of five years ago, and the market now contains a wealth of data for additional business insights beyond buying and selling strategy. In-house counsel can utilise market data to inform strategy and risk mitigation decisions based on increased NPE buying in their technology area. They can look for deeper analysis of competitors’ and the general ecosystem’s licensing and sales activities that may rebound onto their companies. We have even seen the data used to demonstrate a change to a client’s risk landscape employed as support for an increased departmental budget request. The litigation rates for assets on the market, especially sold assets, are so much higher that you should be paying attention to this market even if you are not transacting. Those not participating will find themselves at a disadvantage.
Table 7. Summary of the data
|Summary of 2021 results|
|Annual sales||$166 million|
|Asking price per family||$330,000|
|Asking price per US-issued patent||$208,000|
|Asking price per patent asset||$146,000|
|Number of people employed as brokers||110|
|Sold package litigation rate (tt)||19%|
|All package litigation rate (tt)||5%|
|Brokered packages listed||467|
|Average number of assets per package||12.5|
|Median number of assets per package||4|
|Packages with 10 or fewer US-issued patents||72%|
All data is for the 2021 market year (June 2020 to May 2021) unless noted. Total tracking (tt) - listed June 2011 to May 2021
Some aspects of the market continue to baffle us. The fact that so few sellers provide an asking price when the benefits are clear, and that sellers are unwilling to offer even indications of use when the premiums on the expected value are more than 100%, are just a couple of examples. Nonetheless, we will continue our push for market transparency, efficiency and more data-driven decisions as this market continues to evolve over the next 10 years and beyond.