An automated system for estimating patent value

Patent valuation is becoming increasingly important as revenue generated from intellectual property continues to grow. Although there are several well-known methods for patent valuation – including the income approach, cost approach, market approach and option approach – it can be difficult to apply these in practice. This is largely due to the fact that there is insufficient empirical data to support these methods and they rely too heavily on subjective factors.

This article offers a new method – the patent valuation system – for estimating the profits generated by patents based on existing industry financial data and patent data, in order to minimise the subjective analysis involved. This method can be combined with any valuation method that requires an estimate of the expected returns generated by patents. Comparisons between the values generated by the patent valuation system and real-life values of actual patent transactions are carried out in order to gauge the system’s accuracy.

Currently, the income approach and the option approach are the most widely used patent valuation methods. These use the returns generated by patents to determine value. However, there are many difficulties in estimating patent value using these approaches (eg, estimating the portion of profit attributable to the patent). This article introduces a new and innovative method of computing this portion of profit automatically, using existing financial and patent data along with the discounted cash flow (DCF) model (one of the income approach models). To test accuracy, actual patent values generated by this method are compared against publicly available patent values.

Difficulties in estimating patent-generated profits using existing methods

Different types of patent owner

The existing patent valuation methods may not be suitable for all types of patent owner. For example, the income approach calculates the value of a patent based on the profit generated by the patent owner from manufacturing and selling products incorporating the patented technology. However, many patent owners today do not manufacture and sell products at all, making this approach inapplicable. Further complicating the issue, many patent owners that do manufacture and sell their own products not only generate income from the products they sell, but also receive royalty income from other manufacturers whose products incorporate the patented technology. The traditional income approach also fails to account for this added revenue.

Another issue relating to patent ownership is that an owner’s management practices concerning its patents can significantly affect their value. For example, if the patent owner mortgages its patents and defaults, the bank financing the loan may try to sell the patent for debt recovery at the value calculated by the original patent holder. However, a potential buyer may not accept this value unless its own considerations are also taken into account. Thus, the sale price from such a liquidation sale can diverge substantially from the actual technical value of the patents. Such subjectivity makes accurate patent valuation extremely difficult.

Figure 1. The patent valuation system method

Patent contribution to product

Even if profit data for actual products incorporating the patented technology is available, only a portion of the profit is actually attributable to the patent. Many other variables – such as marketing efforts or manufacturing efficiencies – can contribute significantly to the profit from a particular product. Also, multiple patents may cover the various technologies incorporated into a single product, requiring the portion of profit attributable to patented technologies to be split further. In most cases the patents covering the technology in a particular product include not only those owned by the manufacturer itself, but also patents belonging to competitors or non-practising entities. Further, a single patent may cover technology used in many different products. In such cases the profits from several different products, perhaps from many different manufacturers, must be combined. As a result, determining the value of a single patent becomes extremely complicated.

Predicting future profits

Because an accurate patent evaluation must account for future revenue generated by the patent, there will likely be a subjective component in projecting future profits. Future profit projections are normally determined by the patent owner, so they are often exaggerated.

A new method for estimating patent profits

In view of the difficulties in objectively assessing patent value as described above, any accurate, objective method must take into account the entire industry for the sector to which the patented technology is relevant and all participants – both manufacturers and patent owners – in that industry.

The new method, described in more detail below, does just that. It takes account of the market size, the total profit generated by all participants in that market, the total number of applicable patents and the technology lifecycle in that industry, while minimising the subjective variables from the analysis.

More specifically, the market for the target patent is defined using the data from all patent owners which are active in the market. The market size and profit from the market are then estimated using these entities’ financial data. From the total estimated market profit, the portion of profit attributed to all patented technology is then computed and divided among each patent, depending on the individual score assigned to each patent. The effective valuation period of each patent is then taken into account, which may vary from the patent’s legal lifespan of up to 20 years. Finally, the DCF method is applied to generate a patent value. Figure 1 summarises this method.

Market size and profit

To see whether market size can be estimated using market participants’ patents and financial data, we examined the correlation between the number of patents (number of live patents or number of filings) and the revenue of companies using statistical analysis.

The target companies were US public companies with at least one specific live patent. The industrial classification from NASDAQ was used to group the target companies in order to evaluate the differences between various industries. The time period used to calculate total revenue from the group (the sum of each company’s revenue in the group) and the total number of patents (the total number of patent filings and live patents from the group) was the five-year period between 2009 and 2013. Industries with fewer than 10 patent filings over a two-year period were excluded. Pearson correlation analysis was used to determine the correlation between the number of patents and the revenue from a total of 87 industries. The results of this analysis are set out in Tables 1 and 2. Table 1 illustrates that the number of live patents shows a stronger correlation to revenue than the number of patent filings.

About 89% of the 87 industries with 10 or more patent filings per year show a Pearson correlation coefficient greater than 0.3. Thus, one can conclude that there is a strong correlation between the number of live patents and revenue for the companies that comprise this 89%.

Table 2 shows the 10 industries with the highest correlation between the number of live patents and revenue, including the total number of companies per industry and the Pearson correlation coefficient.

Based on this analysis, it is safe to say that market size can be estimated using market participants’ financial and patent data. However, this valuation method may become unjustifiable for industries with weak or inverse correlation.

Market profit

The contribution of patents to market profit is only a portion of the total market profit. Other activities by market participants also contribute to total market profit, including sales, promotion, customer satisfaction, cost reduction and R&D activities. While the ratio of R&D activities to total corporate activities, R&D costs to total corporate costs or the total number of R&D personnel to total corporate personnel can each be used to determine the portion of total profit attributable to patents, the patent valuation system uses the ratio of R&D personnel to total corporate personnel to calculate the patent’s contribution to market profit.

Market profit attributable to each patent

The amount of each patent’s contribution to market profit can be obtained by dividing the total profit contribution of all patents by the total number of patents in the market.

However, each patent’s relative contribution to market profit can vary (some patents may be more important than others), so that each patent’s evaluation score is used to account for such variance. Many commercial database providers make available patent evaluation scores. The patent valuation system uses scores provided by ActionablePatents.com.

Patent life

Although the legal life of a patent can be up to 20 years, each technology has its own lifecycle, so the practical value of the patent may not extend to the end of its legal lifetime. To measure the effective technology lifecycle of a patent, this article extracts citation relationships between all patents in that technology area, using a median value of the time difference between cited and cited-by patents’ filing dates. The patent’s future contribution to market profit is then computed using the DCF model.

Case study 1

An example is set out in Figure 2 to illustrate this methodology.

Let us assume that nine companies (A to I) are participating in the smartphone market. Of these, A, B, E, F and G have publicly available financial data, whereas C, D, H and I do not. The rectangle reflects the smartphone market size for a specific year. Each company’s respective revenue for that year is depicted by the size of its plane.

Figure 2. Nine publicly and privately held smartphone companies

Market profit

For companies A, B, E, F and G, the patent valuation system calculates the revenue per patent [revenue / (number of live patents)]. To obtain a mean value for revenue per patent, R, the patent valuation system excludes outliers (the top and bottom revenue per patent values) and then calculates the mean. The total number of live patents in the smartphone owned by all nine companies is N. The total smartphone market size M can then be calculated as follows:

M=R x N

If the average net income after tax ratio for A, B, E, F and G is PR, then market profit, P, can be calculated as follows:

P=M x PR

Market profit generated by all patents

The patent valuation system calculates the ratio of total number of patent inventors to total number of employees per company – excluding top and bottom values – to obtain an average value CR. Market profit generated by patents C can then be calculated as follows:

C= P x CR

Market profit generated by an individual patent

The total patent evaluation score from all patents in the smartphone market is S and patent evaluation score for a specific patent is s. Market profit generated by that specific patent, v, can be calculated as follows:

v=C x (s/S)

Individual patent value

The patent valuation system determines the average patent life by using citation relationships between all patents in that technology area to obtain a median value for the time difference between cited and cited-by patents’ filing dates as the average patent life T. Patent value, V, can then be calculated as follows:

Patent Value(V)=

Case study 2

Table 3 (p63) compares the publicly available actual selling price of a patent and the result for the same patent as determined by the patent valuation system. We used the US Patent and Trademark Office Patent Assignment Database and ActionablePatents.com to compile a list of traded patents. However, the number of publicised traded patents did not match the database exactly. For comparison purposes, the price calculated by the patent valuation system was scaled up or down based on the number of patents in the publicised sales.

Subjectivity and objectivity

The results of the patent valuation system often show a good approximation of actual market price. However, in some cases the valuation result differs by as much as four times the known market sales price. This discrepancy may be due to numerous circumstantial factors – such as whether the patents are part of a standard or patent pool, whether the seller and buyer are in direct competition, the patents’ involvement in infringement claims or lawsuits, or the bankruptcy status of the patent owner. Nevertheless, the patent valuation system minimises subjective factors and bases the valuation on actual data. The automated nature of the patent valuation system also enables multiple patents to be valued quickly at low cost. For entities involved in IP financing or sales, the patent valuation system can serve as an effective starting point for negotiations.