Invention Rating Checklist
This is a reproduction of Investing in Patents by Russ Krajec. For the complete book, get it on Amazon.
An invention rating checklist is provided in Appendix A. A free downloadable PDF version of the checklist is available at http://investinginpatents.com/bonuses.
The checklist has several areas of interest, each of which has a score to enter 1 through 5. Each value has a description that helps rate the invention.
These metrics represent hard-fought data that come from many different sources, and all reflect years of experience.
The checklist does not guarantee success – not by any means – but it does reflect a set of core metrics that should be used to evaluate inventions and patents. The purpose is to cull out inventions that would have little commercial value in order to invest in ones that do.
These metrics are designed to be used in the invention process, but they also have the same applicability to evaluating existing patents, or claims of a patent that is being examined.
How to use the checklist.
The checklist is an attempt to quantify several subjective measures. Each person may have biases or viewpoints that may adversely affect their scores.
The best way to get a true sense of an invention’s value is to have a group of people perform the valuation, then aggregate the scores. When a score is consistent between reviewers, the score is likely close to reality.
When a score varies significantly between reviewers, there should be some discussion to figure out why. Often, one person may be seeing the invention in a much different way than the other, and a discussion will bring about a consensus to the most likely value.
Do not read too much into the total score.
The total score may be deceiving. Two inventions may have the same total score, but when an invention has a “1” for detectability and may be completely undetectable, that may kill the analysis.
There are some inventions that have such an enormous market potential that they would be good patents even with very poor detectability scores. But other inventions with higher detectability but low market potential would be poorer inventions.
The decision to invest in a patent is very much a business decision, and there are tradeoffs.
The checklist is one way to highlight the strengths and weaknesses of an invention so that those business decisions can be made.
Look hard at the inventions. This is business.
Inventors (and startup CEOs) are notorious for overselling their inventions. They have a deep love for their ideas and their businesses, and they can overlook even blatant problems.
This is business. Try to look at the invention with the completely dispassionate, unemotional view that the marketplace will have. The more realistic and honest the evaluation, the more accurate the assessment of the invention will be.
The real value of an invention or a patent comes from business execution.
The Checklist helps evaluating inventions or patents on the speculative value of the invention. The real value comes from actual execution of a business that brings the product to market. A high score only indicates that there is a better probability that an invention will be valuable compared to another invention. However, the real value comes from the hard work of the entrepreneur: successfully bringing a new product to market.
Novelty and Non-Obviousness
Novelty and non-obviousness are the two factors used by the examiners to determine whether or not an invention should be patented.
Novelty is relatively easy to understand: has anyone done this before? That answer is actually quite black and white.
But there is a scale of novelty. Not every invention is as unique as the next. Some are routine solutions to a known problem, which may border on obvious. Other inventions are solutions taken from other industries and applied to the problem. Still other inventions use science that is new to the industry or technology.
G. S. Altshuller analyzed 200,000 patents and assigned a “level of inventiveness” to each one.
Level 1 – obvious, routine solution.
Level 2 – solution not well known in the industry, requires creative thinking.
Level 3 – applies engineering knowledge from other industries or technologies.
Level 4 – improvement uses science that is new to the industry or technology.
Level 5 – new scientific phenomena are applied.
The results of Altshuller’s study found that the patents were
Level 1 inventiveness represents 32% of the patents, and can be thought of as taking less than 10 trials to create.
Level 2 inventiveness represents 45% of the patents, and take up to 100 trials.
Level 3 inventiveness represents 18% of the patents, and take up to 1000 trials.
Level 4 inventiveness represents a mere 4% of the patents, and take up to 10,000 trials.
Level 5 inventiveness represents less than 1% of the patents, and takes over 10,000,000 trials.
Something to note from the chart is that nearly half of the patents surveyed (45%) were level 2, and a vast majority (77%) were either level 1 or level 2.
When analyzing a large set of inventions, the inventions should have a similar distribution where less than 1% are level 5, and 77% are levels 1 or 2. Inventors consistently over value their inventions two or three levels higher than they really are.
As a comparison, Edison’s light bulb patent would have had a level 3 for novelty. The lightbulb filament famously took Edison 1,000 tries to perfect, which corresponds with level 3.
This does not mean that the patent was not valuable, but it highlights the fact that the value of the patent comes from business execution.
Non-obviousness is difficult to quantify.
When the examiners perform their search, they try to find an exact match between the claims and prior art. Often, they will not find an exact match but they will find something close, and then combine that reference with another one which, when viewed side by side, will be “obvious”.
For example, a claim might state that a product has a specific color, such as red. The examiner might be able to find the exact same product in another patent, but that patent might only say that the color is blue. The examiner would find another reference that says blue and red are functionally equivalent (or may just state it without justification) and reject the patent.
There are several arguments that are typically made to the examiner to show that the invention is non-obvious. The purpose of this book is not to go through an exhaustive list, but just to touch on this point:
The applicant must have a convincing story for the examiner to show unexpected results or some kind of “ah-ha” moment.
In the red product invention example, the argument could be that changing the color made it work better in certain situations that no one anticipated. Another argument may be that the industry was going in one direction and the inventor decided to go in the complete opposite direction.
Many inventions involve automating steps that are normally performed by hand (or could be performed by hand). In some cases, using a computer to do a repetitive task can have benefits that would never have been possible with paper and pencil.
Let’s assume that an invention automates a manual task, and that the manual task could be done ten times a day and result in 10 data points. By automating the task, a computer can generate 100,000 data points in the same time.
On the face, this could be construed under Alice to merely automate a previously manual operation.
The unexpected benefits may be that large scale trends can be uncovered when analyzing 100,000 data points rather than merely 10. The further benefits are that large scale automation may allow for more accurate data checking and comparisons, meaning that improper results or anomalies can be detected. All of these benefits would not be possible with a pure manual version.
Patent searches and non-obviousness.
A patent search is the preferred way to find out if an invention is novel, but it is still difficult to figure out if it is non-obvious.
The results of a patent search will be the closest prior art. In many cases, the examiner will reject the case and say it was obvious in view of the closest prior art. To overcome the rejection, the applicant needs to have a solid argument about why the invention is not obvious.
It is very hard to guess how an examiner will view the invention and how they will make an obviousness argument.
However, when the closest prior art is very close, the obviousness argument may be impossible to overcome. If a patent attorney performs the search, they will always be pushing to file a patent, while knowing that the patent would be incredibly narrow.
The only search that counts is the one the examiner does.
There is a point of diminishing returns on patent searches prior to filing. No matter how much searching is done before writing the patent application, the examiner’s search is the one that counts.
Rather than spend a huge amount of money on searches, just send it to the examiner and see what they say.
Detectability
Undetectable infringement makes a patent worthless.
Detectability is a measurement of how difficult it would be for someone to detect that an infringer actually infringes a patent.
Detectability is ranked based on difficulty. On the high end, some inventions are detectable based on casual inspection. Often, a competitor will advertise the feature as a compelling reason to buy their product.
Sometimes, detection might not be so obvious and would only be known after targeted investigation.
Sometimes, infringement is detected only by subject matter experts, and sometimes it may take insider knowledge about how something works.
Still other times, it may be utterly impossible to detect that someone infringes.
Level 1 – Undetectable without specific knowledge about implementation.
Level 2 – Detectable through experimentation by subject matter experts.
Level 3 – Detectable through targeted investigation and experimentation.
Level 4 – Detectable through inference based on observation.
Level 5 – Detectable through casual inspection/competitor advertising.
Lots of products are impossible to detect.
Let’s consider a software product. Many times, a software product executes deep in the bowels of a datacenter or inside a device. If there is no way to know that a competitor is running the same algorithm, the patent is completely undetectable.
Sometimes, hardware inventions are hard to detect. For example, an inventive clock circuit on an integrated circuit might only be detectable by reverse engineering the IC chip and removing each layer, then mapping the circuitry. This can be a painful and tedious process, but may be worth the effort for extremely high value inventions.
A chemical or manufacturing process may be completely undetectable when the process is performed inside a factory and there are no telltale signs of the process.
The more detectable, the better. When the invention is not detectable, it is almost always a good candidate to maintain as a trade secret.
Every word in the claim is important.
Remember to analyze the claims, not the description of the invention, when analyzing detectability. The claims are the strict legal definition of the invention.
Ask whether or not detection is possible for the exact words in the claim. In a method claim, could someone tell that a competitor did the exact method of the claim? In a product claim, could someone tell that a competitor's product was exactly like in the claim?
For something to be detectable, a person should be able to verify each and every element of the claim. If there is one small item, no matter how small, that is required in the claim but cannot be detected, the patent does not have much value.
Inventors like to brush off detectability by saying that they can infer that infringement has occurred. The inference is often based on the inventor’s belief that their way of doing the invention is the “best” and nobody would possibly do it any other way.
This argument does not hold water.
Infringement requires solid, provable evidence that holds up in court. If there is any way possible that someone could do the invention, even if it was not perfectly optimized like our invention, the patent is not enforceable.
Overcoming detectability issues.
One way to overcome a detectability problem is to look for artifacts or telltale signs of the invention.
For lots of software inventions, a good practice is to focus on the interfaces of the software product, rather than the guts of the product. A patent on a sophisticated algorithm that automatically underwrites mortgages, for example, may be completely undetectable.
However, an application programming interface that accepted a borrower’s credit card history and returned a yes or no decision on a mortgage would be perfectly detectable and enforceable. After making that change to the focus of the invention, we might do another patent search to determine whether or not we could get a patent on the application programming interface.
Actor Analysis
Another very important factor to consider is the actor in the invention. There are some actors that can be sued, and others that cannot. Further, when two or more actors are in a claim, there is a problem of divided infringement.
The actor analysis has several different levels, from bad to good:
Level 1 – Two or more actors, all of which are customers or a single actor who is immune from litigation.
Level 2 – Two actors, one is a deep pocketed competitor.
Level 3 – One actor who is a customer, but has deep pockets.
Level 4 – One actor who does not compete, but has deep pockets.
Level 5 – One actor who is a direct competitor and has deep pockets.
This analysis is not linear, meaning these levels are not evenly divided steps. They are arranged in order of preference.
Some business cases support a patent directed at a customer that has deep pockets when other factors are present. For example, a product sold to enterprise customers may be protected by claiming actions that the customer would perform. This situation may be somewhat acceptable because it might keep the customer from building the product in house.
It would be preferred to have the patent directed at a competitor, but sometimes that is not possible given the specific invention.
Direct infringement by a consumer or customer makes a patent worthless.
An example is helpful.
Consider an invention of heat-moldable insert for a shoe, such as a ski boot. The invention was originally conceived as heating up the insert in boiling water, forming it to the user’s foot, and inserting it in the ski boot.
Who is the infringer?
It is the customer. Can the customer be sued? No. Does the patent have any value? No.
A patent that claims a method of use, like our example above, may not be enforceable when trying to stop copycat importers from overseas. These kinds of claims are not very likely to be enforceable by the International Trade Council, which is the best way to hold up shipments of infringing goods.
There are many different ways to express claims.
In the example above, the claims are “method-of-use” claims. These tend to be one of the weakest types of claims but some of the easiest to get. In many cases, method of use claims can capture the business value of the invention succinctly and broadly.
Another way to claim an invention is a product claim. Product claims are some of the strongest claims because they are easy to detect and they capture the product that would be produced by a competitor.
In some cases, product and process claims can get intermingled, where the product claims require certain operations or installation conditions to be met. These claims can cause lots of actor problems as well.
How to overcome actor problems.
In the example, the claims can be rewritten to focus on the product produced by a competing manufacturer. In the example, the claims could be rewritten as a plastic component that gets soft at one temperature and hard at another temperature and has a specific shape.
The rewritten claims now focus on a specific actor that we are targeting. The business case for the patent is to stop competitors from manufacturing a competing product.
Software products often are difficult to patent because two or more actors combine. Sometimes it is very hard to write a patent claim where the consumer or end user performs an action, then the computer does something in response.
A best practice is to stand in the place of the device and write a method claim that “receives input from a user.”
Another aspect of a software product is that various pieces are often performed by different parties. As software platforms evolve, there are services that specialize in the most mundane aspects of a solution, such as software-as-a-service providers for authentication, payment processing, or whatever.
Be mindful of the fact that software platforms and services are ever changing, and focus the invention on a relatively narrow value added element or point of novelty. As a product goes to market, some aspects may be ‘outsourced’ to other providers who have an API specifically designed to do some aspect of the product.
The best way to use actor analysis is to reform an invention to focus on the point of novelty that the products or services provide. Ideally, the point of novelty is also the customer benefit.
Protecting the razor/razor blade product.
Two actor claims are those that require two different people to perform the invention. One actor may be a supplier of a non-reusable part (e.g. a razors) while another may be a supplier of the reusable part (e.g., the razor blade). These types of patents are notoriously hard to enforce, and therefore tend to be worthless.
A better way to focus the claims in the razor example would be to focus on the mechanism that attaches the razor blades to the razor. Two patents (or at least two sets of claims) would be written: one describing only the mechanism on the razor side, and a second describing only the mechanism on the razor blade side. In this way, the two patents could be used to license razor blade manufacturers to provide compatible blades, or to license razor manufacturers when the company wishes to provide the blades.
Design Around Analysis
The baseline for all economic valuation of an invention is the best alternative to the invention. The easiest way to find the best alternative is to attempt to design around the invention.
Once the invention is identified, the inventor (or sometimes a separate set of inventors) will try to “design around” the invention.
This should be viewed from the technology standpoint, by looking at different technical solutions to the same problem.
It should also be viewed from the business standpoint, by trying to envision different ways a customer might solve the problem – or ways that they might live with the problem.
There is ALWAYS an alternative to the design.
How many times does a CEO/inventor tell people “there is nothing like it”? This is just not true – or if it is, it proves that nobody wants the product.
In some cases, the best design around alternative is actually better than the invention. By going through this analysis, it may better hone the business value of the invention. With some inventions, the invention may have advantages in one set of circumstances and disadvantages in others.
Once the best alternative is identified, the economic advantages of the invention should be determined. The economic advantage calculation should be as detailed as time may allow. In many circumstances, a back-of-the-envelope calculation may be sufficient to bracket the value of the invention.
The design around analysis considers two factors.
A comparison of the invention to the best design around alternative:
Level 1: Best alternative is superior in cost/performance to invention.
Level 2: Best alternative is equivalent to invention in cost or performance.
Level 3: Best alternative is missing important features of invention.
Level 4: Invention has 5x performance advantage over best alternative.
Level 5: Invention has 10x performance advantage over best alternative.
A comparison of the true economic advantage of the invention:
Level 1: Invention no economic advantage.
Level 2: Invention has economic advantage of 1.5x.
Level 3: Invention has economic advantage of 2x.
Level 4: Invention has economic advantage of 4x.
Level 5: Invention has economic advantage of 10x or greater.
In these analyses, we are trying to establish the value of the contribution of the invention to an overall product.
Alignment With Internal Business Goals
Patents protect a business only when the patents are directly related to the products being produced. This analysis considers two factors: how close the product is to shipping; and how key the invention is to the product’s value.
The invention can be graded using a first set of metrics:
Level 1: No resources assigned to invention.
Level 2: Feasibility stage only.
Level 3: Resources are devoted to implementation
Level 4: The invention is likely to ship in the next product release.
Level 5: The invention is shipped or committed to ship.
The second aspect is how important the invention is to the products.
Level 1: Supplemental aspect of product.
Level 2: Improvement to product.
Level 3: Key feature of product.
Level 4: Important aspect of product strategy.
Level 5: Key aspect of product strategy and product roadmap.
The goal of this analysis is to ensure that the patents focus on inventions that are important to the products are going into the marketplace.
Alignment to External Business Possibilities
External business possibilities add value to a patent because they open the door to licensing the invention to other companies. The single best way to get acquired is by having products and patents that fit into a competitor’s portfolio.
This analysis considers three parts of the problem: how well does the invention fit with a competitor’s products and strategy; the outbound licensing potential of the invention as a standalone product; and the likelihood that competitors will adapt the invention as a standard.
For the product fit in the competitor’s portfolio, we consider:
Level 1: No known or expected competitor activity.
Level 2: Solid improvement to competitor’s product where different solutions exist.
Level 3: Address a shortcoming in a competitor’s product that a third party may implement.
Level 4: Addresses a shortcoming of competitor’s product in a substantial manner.
Level 5: Addresses direct need of competitor’s product and fits in competitor’s strategy.
For the outbound licensing analysis:
Level 1: No expected licensing potential – invention limited to our products/technology.
Level 2: Invention could be licensed as part of a package with other inventions.
Level 3: Invention as good value in other markets.
Level 4: Invention has very significant value in other markets (i.e., dramatic cost savings or improvements).
Level 5: Invention has significant value and is a key element of a licensable product for other markets.
For the standards-related analysis:
Level 1: Not expected to be used by competitors.
Level 2: Market has no preference for invention over competing options.
Level 3: Some competitors may adopt.
Level 4: Most competitors are likely to adopt.
Level 5: Likely to become standard in all competing products.