Patents are often misunderstood in the startup world. In this practical guide, Alchemist mentor Will Allen breaks down how founders should think about patents—as strategic assets with defined roles, real costs, and market-tested value.
My name is Will Allen. I’m a mentor with Alchemist (Class 39), and I've enjoyed working with the participating organizations in this cohort. I’ve looked through many of the company profiles and it’s energizing to see how much serious innovation is underway.
By background, I’m a former HP Fellow (I parted ways with HP in 2020; I was the first Fellow promoted in HP’s Printing business). Over my career, I’ve accumulated over 100 patents and have also served in intellectual property governance roles at both small and global entities, including deciding what to patent, developing patent strategies, and working closely with technical and legal leaders.
Standard disclaimer: I’m not an attorney, and I’m not giving legal advice. Patents are legal instruments backed by national legal systems. For legal guidance, consult qualified legal counsel. Here, I share practical opinions I’ve developed from experience about a useful way to think about patents. I emphasize startup contexts where focus and resource allocation are paramount.
This article mirrors the structure of my companion video, and is aimed primarily at Alchemist participants and other organizations that consider IP a critical to success.
Intellectual property is, broadly, knowledge that can be possessed in an advantageous way. Common forms include patents, trademarks, copyrights, trade secrets, and industrial designs.
In this article, I’m focusing on patents, specifically the category most people mean when they say “patent”: utility patents (functional inventions such as processes, devices, systems, methods, compositions of matter, etc.). There are also design patents (ornamental appearance and non-functional features) and plant patents (specialized and distinct in how they operate). Design and plant patents are outside of this article’s scope.
At a high level, there are three core ideas I want readers to walk away with:
Patents are property.
Each patent should have a role.
Patent value is determined externally.
Those sound obvious, but taking them seriously changes how you budget, prioritize, and communicate your IP strategy. A solid IP strategy is important for all organizations, and it’s critical for startups.
Patents have a cost/return profile over time. Patents aren’t documentation of “good ideas.” They are property rights created through a process that consumes time and money:
Time to invent, test, and document the invention
Resources spent to decide if the invention is worth patenting
Attorney time to draft and prosecute
Ongoing interaction with patent offices
Maintenance fees over the life of the patent (typically around 20 years from filing for US utility patents)
In practice, the monetary planning “unit” is blocks of $10,000. Over the life of a patent, total legal costs frequently end up between $15k - $30k, and that excludes invention and testing. For startups, that means every filing decision competes with engineering, sales, hiring, and all other expenditures.
Patents don’t give you permission; they give you the right to stop others
A patent does not grant permission to practice. It’s a right to exclude. A right to prevent others from doing something.
Precisely what you can stop someone from doing is defined by the patent’s claims, as interpreted under the law.
The two principal parts: specification and claims
A patent document has two principal parts:
A patent can have a brilliant specification section and still have weak claims. Conversely, it may have strong claims; each claim must be technically enabled by what is taught in the specification. Beyond enabling claims, additional information in the specification can be risky… it teaches others and may not be protected. This is why experienced patent counsel matters: claims are both definitional language and expressions of strategy.
If you remember only one sentence from this section, claims are the “perimeter” of a patent’s ability to be enforced to stop activities.
Ownership can be transferred, sliced, and monetized
Because patents are property, they can be (for example):
This flexibility matters for startups. Licensing can be a/the product. Licensing can be a bridge: build a business in one market while licensing patents in non-core markets.
Enforcement isn’t free
A patent doesn’t magically make someone stop. It gives you a legal instrument you may be able to enforce, with enforcement costs frequently running beyond one million dollars. This is a powerful reason startups should be clear-eyed about what they want a patent to do and what they’re realistically willing to spend to defend it.
An analogy: patents and music rights
I sometimes compare patents to music rights. They are not identical, yet the analogy is useful.
A songwriter owns rights to a work they created, and the market decides whether the song is valuable. Rights can be sold or licensed. Enforcement options exist, but they aren’t free. The key difference is that patent claims define the ilk of what’s enforceable, while a musical work is defined by notes and lyrics. If a claim mandates four technical elements, and a competitor’s design uses only three of them, there may be no infringement.
That “design-around” (purposeful or accidental) reality is not a flaw; it’s the system. It’s also why claims, and the strategy behind them, are the heart of patent value.
For example, suppose a competitor’s claims define a method for processing real-time video of the interior surface of a 3D hollow object into a 2D representation. The claims only cover a specific process of stitching overlapping 2D video frames into a flat composite image. It may be possible to avoid infringement by first creating a 3D model of the object's interior using publicly available techniques, and then flattening that model into a 2D representation. While the end results are similar, the paths to those similar results are different.
Startups don’t have the luxury of collecting patents “because IP is good.” Patents are expensive assets. Every filing has an opportunity cost. I recommend treating patents like hires: each one needs a job description.
A simple model I’ve found useful is a triangle of three patent roles:
Vertex A: Product differentiation
Use patents to protect product features you don’t want copied. This is the “I built something customers want, and I don’t want competitors to clone it” use case.
Differentiation patents tend to map directly to product characteristics and buyer value. They’re often the easiest to explain to investors and partners, provided you can articulate
how the claims translate into business value.
For example, a company patents a specific connectivity interface/mechanism and pairs it with branding/certification, creating an interaction control point.
Vertex B: Monetization
You may not want to manufacture a particular product at scale. You may want to license the invention to someone else and collect royalties. You can also license a patent for use only in non-core fields-of-use or in specific geographies, while you focus on your main market.
For some ventures, patents aren’t defensive; they are the product.
Vertex C: Freedom to operate (FTO)
This one is often misunderstood. Freedom to operate, as founders usually say it, means “I want to be sure I’m not infringing.” Strictly speaking, that’s an analysis done (frequently by hired experts) by examining other entities’ patents, not by filing your own.
A business “milestone”: upon market success, a business may become the target of IP litigation. If you have patents that another party values highly, you have leverage for a cross-license or other potentially favorable outcome. Having a curated portfolio of IP that is valuable to others can assist in moving conflict away from litigation and toward a negotiated, favorable outcome.
If you have no IP of value to the other side, you have fewer levers.
Most patents are not “triple-role”
It’s rare for a single patent to strongly deliver differentiation, monetization, and FTO leverage simultaneously. I have seen many patents take on strong dual roles. But planning as if every patent will do everything is unrealistic.
A patent’s role can change over time. A patent originally filed for differentiation might later become monetized (licensed out) after a business pivot. Or it may become leverage in cross-license negotiations. Roles can be dynamic, and it’s important to have clarity before spending resources to generate or acquire rights to patents.
This is the part startup executives and investors often dislike: early determination of a patent’s value (or a portfolio’s value) is an estimate.
Value depends on what the claims read on
The biggest driver of value is not the technical cleverness of the invention; it’s what the claims cover in the real world and how that coverage impacts other parties.
Put simply: what would this patent allow the assignee (owner) to stop?
If the answer is “not much,” or “only things with design-around alternatives,” then the patent may not be valuable in spite of impressive underlying technology.
Value depends on external perception
The second major driver is how outsiders evaluate the patent:
External perception is shaped by claim scope, enforceability, and the extent to which the patent maps to meaningful economic activity.
Robustness matters: patents can evaporate
A patent can be attacked. If relevant prior art exists (if it can be shown the patented invention wasn’t novel), then some or all of the patent’s claims can be invalidated. When that happens, what looked like a valuable asset can turn into dust.
If a patent is challenged and withstands attacks, that “battle-tested” status will increase its value significantly.
Sophistication is not a value by itself
I’m not anti-sophistication…I love exotic technical solutions. Many important inventions are sophisticated.
But I’m pragmatic about the market: if you patent a sophisticated way to do something, and someone else invents a simpler approach that performs just as well, then your sophistication doesn’t matter. Actually, the “sophistication” becomes unnecessary complexity…it’s a liability. What matters is whether the patent’s claims also block the simpler approach, and/or do the claims also read on the final result.
This is why I treat technical sophistication as secondary to:
A portfolio mindset helps
Because value is externally determined and somewhat speculative, it often makes sense to think about portfolios, not single bets. You do diligence. You pick intelligently. But you also accept uncertainty and structure your strategy so it will survive being wrong on any one investment.
Patent portfolio management is an investment portfolio problem, and there are well-understood methods to design a portfolio to meet a set of explicitly stated needs over time.
If you want a startup-oriented way to apply all of this, here’s what I recommend:
1) Design a portfolio to meet business needs.
Use portfolio management paradigms to balance meeting needs, risks, and uncertainty.
2) Write down each patent’s role before you file…better yet before you invent*!
Differentiation, monetization, or leverage. What job is this patent being hired to do?
*Yes, it is absolutely possible to write a set of claims you’d like to have, then invent a specification that supports them, and finally file a patent application on that invention. I’ve done this many times. It’s “engineering” patents to meet stated needs, such as engineering an apparatus to meet them. Novel invention is still required; however, it’s directed toward business needs codified in the form of claims. Reach out if you’d like advice on this subject.
3) Force the “claims reality” discussion early. What do we want to be able to stop and why? What are the likely design-arounds?
How hard is it to detect infringement?
Which parties does this patent read on (competitors, customers, other)?
4) Budget explicitly.
Treat filings, and potential contract and litigation activities, as a real line item competing with product marketing, competitive analysis, salaries, and other financial outflows.
5) Treat enforcement as an option, not an assumption.
“We have a patent” is not the same as “we can afford to litigate against a competitor.”
6) Align business, technical, and legal leadership.
Patents are technical in creation, legal in form, and business in use. If those functions aren’t aligned, you’ll either over-file, under-file, or file the wrong things. Those are undesired wastes of resources in the time-and-budget-constrained world of startups.
7) Keep a short list of “future topics.”
Having a strategic framework helps decide if an organization should patent an invention today, in the future, or never. Having a “later” bucket can make it easier to appropriately assign “not today” as a near-term state.
I’m currently CTO at Kaspix (link), where we’re building analog AI circuit generation technology that sits at the intersection of deep engineering, product deployment realities, and IP strategy. In that world, it’s easy to waste money on IP that looks impressive but doesn’t map to business outcomes. And it’s equally easy to under-invest and later discover you left core value unprotected.
My goal (as a mentor and as a technology executive) is to keep IP thinking practical: not fear-based, not checkbox-based, and not vanity-driven. Patents are tools. They’re
expensive tools. Used well, they can create a durable advantage. Used poorly, they burn cash and resources without impactful results.
I’d love feedback on which additional IP topics would be most valuable as a future article/video topic.
If you’re in Alchemist and want to discuss your situation directly, please reach out. I’m happy to help!
I first recorded a video (YouTube link) on this topic, speaking extemporaneously to a set of high-level topic slides I created. I then used ChatGPT to summarize a transcript of the video and produce a draft article. Finally, I edited the draft word by word to arrive at this result, which is 100% my content.