The Inventor’s Strategic Matrix
In over 25 years of working with startups and innovation projects, I see a pattern that repeats regularly: a lack of clarity about the real role an invention can play in a business strategy. If you are an inventor or entrepreneur driving an innovation, this model can help clear up some confusions before they become costly.
Inventors tend to build their expectations on a fairly simple model:
INVENTION → ??? → PROFIT!
The question marks in the middle, where the magic is supposed to happen, represent the business model. In most cases, no one examines it clearly before making decisions that later become irreversible.
There are two versions of this problem.
The first is the pure inventor who believes the value of what they created will assert itself on its own, that the world will recognize the brilliance of their invention, and that some form of reward will follow naturally. The reality is that the idea may be brilliant, but that does not necessarily mean it can sustain a business. Because having a good idea is not the same as having a good offering.
The second is the entrepreneur who develops an innovative product but, instead of understanding what they actually have, moves forward “as they see fit,” “by the book,” or copies the model of some successful case without seeing that it does not apply to them. I call these autoimmune ventures: they attack themselves without knowing it.
In the strategic direction of an innovation, we need to answer concrete questions. What aspects of our project will become inevitably exposed as soon as we demonstrate that the product works? What elements can we keep open, or even benefit from keeping open, without risking future monetization? Is our innovation a sustainable business on its own, or is it actually a complement to another commercial model?
The Inventor’s Strategic Matrix is the model I developed to bring clarity to strategic decisions in innovation projects and to answer these questions with confidence. It is the product of three decades of professional practice, crossed with lessons learned back in 2004 from my mentor Roberto Cunningham (Director General of IAPG), and the vision of figures like Jorge Sábato, Joel Spolsky, Frederick Lanchester and Simon Wardley.
The Matrix does not prescribe what to do: it describes the board. Once you understand the board, you can consciously choose which game you want to play. Before that, you are playing without knowing the rules.
The Model: The Inventor’s Strategic Matrix
The Matrix has three dimensions:
- The nature of the asset: how exposed it is by its own nature, regardless of what the inventor decides
- Technical protection: what technical mechanisms the inventor can apply to control access
- Legal protection: what legal instruments the inventor can apply to control use
These three dimensions are organized in a table where the rows are the levels of asset nature, and the columns are two independent groups: technical protection and legal protection.
| Asset nature | Technical protection | Legal protection |
|---|---|---|
| Level -1: Preexisting knowledge | None | None |
| Level 0: The inevitable | None | None |
| Level 1: Observable with effort | Obfuscation | Copyright / Patent / Open license |
| Level 2: Controllable | Friction / Control of indispensable resource | Trade secret / Patent / Contracts |
| Level 3: The intangible | None | None; it is cultivated and sustained |
Dimension 1: The Nature of the Asset
Intellectual assets are not all the same. They have a nature that ranges from more exposed to more opaque, and this nature does not depend on decisions made by the inventor, but is an inherent characteristic of the type of asset.
Level -1: Preexisting Knowledge
This is what existed before the invention. Science, culture, the collective knowledge accumulated by humanity. It cannot be appropriated by anyone. You cannot patent a law of physics, a known chemical reaction, or a collective cultural tradition. The corporate attempt to patent varieties of Andean quinoa or medicinal uses of turmeric illustrates this violation: it is biopiracy of knowledge that intrinsically belongs to Level -1.
Level 0: The Inevitable
This is the idea that is revealed simply by showing the product. It is not a decision to open up. It is an irreversible consequence of existing publicly.
The moment you show that your product works, you reveal that the idea is possible. This cannot be protected in any way, regardless of what the contract, patent, or confidentiality agreement says. Intellectual property laws do not protect an idea, but rather specific implementations.
If an initiative has a part that is exposed there, “opening” it is not an error or a surrender of intellectual property. It is formalizing what is already happening in practice. Any explanation you need to give for people to become interested or adopt the initiative falls into this territory. If you need to explain things for your project to be credible and trustworthy, what you have to reveal is necessarily Level 0. You have no way to protect it, and the best you can do is have your strategy incorporate it clearly.
Collaborating at Level 0 does not generate rights or obligations regarding what comes after. Being free as air, this level cannot sustain any business model, because you cannot generate income on something that circulates freely and does not require a transaction to obtain it. Whether the first person to present the idea or the last to learn of it, anyone can take it and develop something on their own in upper layers that can be protected and monetized. That future commercial structure may owe moral recognition to the original Level 0 work, but in practice (and by legal advice) the commercial model operates autonomously.
The upper layers can have a commercial model. An open Level 0 does not close that possibility: it enables it, generating the credibility and conversations needed to finance protectable intellectual property in upper levels.
The question that defines what is at Level 0 is simple: can someone seeing my product work for the first time infer the idea behind the product, and that my product proves that idea is possible? If the answer is yes, it is at Level 0.
Friendster was the historical example: it demonstrated in 2002 that a social network with real identity worked. The concept was revealed at Level 0 to the entire market including Mark Zuckerberg, who was a Friendster user when he launched Facebook.
Level 1: Observable with Effort
This is what requires analysis, reverse engineering, or specialized access to see. It is not revealed on its own, but it is not protected by architecture either. With time, resources, and sufficient technical knowledge, it is accessible.
Level 1 is where many inventors believe they are protected when they actually are not. The illusion of technical complexity is not protection.
The chemical formula of Chanel No. 5 resides here: although the brand is legendary, any modern laboratory can apply reverse engineering and identify its components. Chanel’s real protection is not in the formula but in the control of the supply chain of its specific ingredients, which is Level 2.
Level 2: Controllable
This is what the inventor can maintain in an environment under their direct control. It can be code that runs exclusively on their own server, a calibration database, or the parameters of an industrial process that never leave the plant. Here lies the real strategic decision: exposure is not inherent but depends on the architecture and technical discipline of the inventor.
Coca-Cola operates at this level. By keeping its formula as a trade secret under strict physical control, it exercises protection that has been in effect for over 140 years: much longer than the 20 years a patent would have lasted.
Level 3: The Intangible
The community, reputation, authority, positioning. This level is not patented or stored on a server. It is the only truly permanent asset, and it is built as a consequence of decisions made at lower levels. Unlike the other levels, it does not admit technical or legal protection barriers: it is only cultivated and sustained, and it can be destroyed from within by the poor decisions of those who manage it.
The network of users, journalists, and politicians on Twitter/X is the paradigmatic example. Microblogging is Level 0 and the code could be replicated in weeks, but the concentration of social capital on that platform is irreplicable overnight. Even after the most controversial decisions of its new management, the platform retains its gravitational weight over all the clones that emerged to replace it.
Dimensions 2 and 3: Protection Decisions
The horizontal axis of the model details the barriers the inventor can put in place. These are two independent groups of decisions.
Technical Protection
The architectural or physical barrier. It is divided into three mechanisms in increasing order of robustness:
Obfuscation. Hiding the source code, formula, or process. This is the weakest mechanism: it collapses with a single discovery event. A former employee, a leak, or advances in analysis tools can destroy it. Stac Electronics’ code depended on this and collapsed in a meeting room.
Access friction. Making access costly without controlling it absolutely. This is more robust than obfuscation but still temporary: it collapses when technological change reduces access cost to zero. The music CD required purchasing a physical object; Napster reduced that friction to zero in 1999 and the recording industry’s business model was never the same.
Control of the indispensable resource. The most robust mechanism. It does not hide information; it controls the physical infrastructure, hardware, or supply chain without which the asset cannot function, regardless of who knows the concept. Apple Silicon eliminated the informal “Hackintosh” market that existed in the Intel era. Not because any law changed, but because the resource (the processor) became inaccessible to third parties again.
Legal Protection
It operates independently of technical protection. Its main instruments are:
Patent. Protects the specific implementation. It requires revealing how it works in exchange for a temporary monopoly of approximately 20 years. It has jurisdictional scope: a patent in Argentina does not protect in Norway.
Copyright. Protects the specific expression, not the idea. It exists automatically in most jurisdictions and has a much longer duration than a patent.
Registered trademark. Protects the name and identity. Unlike a patent, it does not expire as long as it remains active. Aspirin is a registered trademark of Bayer in Argentina and Germany, but is a generic public domain term in the United States, a loss that occurred as war reparations in 1918.
Open license. Defines what others can do with the asset. A GPL license, for example, allows free use but requires that all derivatives maintain the same open conditions. It does not protect the idea but regulates its use.
The Cross-Cutting Principle: Protecting Means Delaying
No technical or legal protection on an asset is forever. Patents expire. Architectures are breached. Resources change. Protecting is simply buying time.
Once you assume that premise, the strategic decision reduces to choosing what you build with the time you buy. There are three models:
Exploitation and exit. You exploit the product to the maximum while protection lasts, without intending to extend the business beyond that limit. You collect the rent and when the barrier falls, you withdraw with the capital. It is a one-shot bet, conscious and legitimate.
The ladder. You use the time and capital your current protection gives you to develop the next product. You know your current patent will expire, but you plan that by the time it does, you will already have the next generation protected. This is the game of mature corporations.
The intangible moat. You use the advantage time to build a community, reputation, or market standard at Level 3. By the time competitors manage to replicate your technology, it no longer matters: your business value has moved to an asset that is immune to reverse engineering.
The Possible Games
Once positioned in the Matrix, the inventor can choose their strategy:
The patent game (Edison). Methodical registration, active licensing, and litigation when necessary. Requires legal backing and accepts that in 20 years the invention will be public. The goal is to extract maximum value during the exclusivity window.
The trade secret game (Coca-Cola). Strict opacity and permanent internal control. Unlimited protection over time, but total vulnerability if the secret leaks: there is no legal recourse to recover disclosed information.
The indispensable resource game (Apple). Control the physical or logistical bottleneck that enables access to the asset. It does not depend on secrets or anyone figuring anything out: it depends on a structural position that is difficult to replicate.
The community game (Linux). Deliberately releasing Level 0 and Level 1 so that adoption builds a Level 3 ecosystem that no competitor can copy. The business is not in the code but in the authority and services surrounding the community.
The hybrid game (Android/Google). Google open-sourced Android to generate mass adoption: any manufacturer can use the operating system freely. But Google Play Services (the app store, maps, Google login, behavioral data from billions of users) are completely proprietary. Manufacturers who want Android with the full experience must accept Google’s terms. The openness of the visible layer generated an invisible lock-in in the services layer.
In-Depth Cases
Stac Electronics vs. Microsoft: The Cost of Showing What You Shouldn’t
In 1993, Stac Electronics had created Stacker, the first real-time disk compression software for DOS. The concept was Level 0: anyone who saw it working could understand it was possible. But the specific algorithm was patented and the source code was protected by obfuscation: the distributed binaries did not come with source.
Microsoft wanted to include compression in MS-DOS 6.0 and began negotiations with Stac to license Stacker. During the process, Stac engineers showed source code in the meeting room, believing the context protected them.
The negotiation failed. Microsoft released DoubleSpace. The code was functionally identical to Stacker’s.
Stac before the meeting:
| Asset | Tech. protection | Legal protection |
|---|---|---|
| Real-time compression concept | None (Level 0) | None |
| Specific algorithm | Obfuscation (binary) | Patent |
| Source code | Obfuscation (binary) | Copyright |
Stac after showing the code:
| Asset | Tech. protection | Legal protection | Change |
|---|---|---|---|
| Real-time compression concept | None (Level 0) | None | No change |
| Specific algorithm | None | Patent | ⚠️ Technical protection lost |
| Source code | None | Copyright | ⚠️ Technical protection lost |
The legal protection did not change. The technical protection collapsed in an instant due to a human decision in a meeting room. Stac won the lawsuit and received $120 million, but Microsoft’s countersuit (based on a trade secret Stac had seen during the same meetings) left them in a position that forced a settlement and condemned them as an independent company.
The lesson is not not to negotiate. It is to know exactly what level you are exposing before entering the room.
Friendster: Validating the Market for Someone Else to Take It
Friendster demonstrated in 2002 that social networks with real identity worked. The moment the product worked publicly, that idea moved to Level 0 for the entire market. Friendster had a time advantage but did not build any indispensable resource at Level 2 nor manage to consolidate a community of ideological identification at Level 3.
The result: they validated the market, educated users on social network behavior, and paved the way for Facebook, which executed better on exactly the same Level 0 base. Friendster ended up selling its patents to Facebook for $40 million in 2010 and closed in 2015.
Twitter/X: When Level 3 Is the Only Real Asset
Technically, Twitter/X does not keep any secrets. Microblogging is Level 0 and the code could be replicated in weeks. Mastodon, Bluesky, and Threads did it. None managed to displace it.
Twitter’s real asset was never the technology. It was the irreplicable concentration of journalists, politicians, scientists, and public figures that turned the platform into the digital public square. Elon Musk paid $44 billion in 2022: not for the code, but for that network. And subsequent management showed that Level 3 can be destroyed from within: the most controversial decisions generated migrations that would not have been possible if the intangible had remained intact.
Conclusion
The Inventor’s Strategic Matrix does not solve the inventor’s problem. It makes it visible.
The question mark in the “invention, ???, profit” model does not disappear with the Matrix. But the Matrix allows you, for the first time, to ask the right questions about what is in that space.
The real problem is not protecting what you have. It is understanding what you have, what level it is at, and what you can realistically do with it. Edison knew what game he was playing. Coca-Cola knew what game it was playing. Linux knew what game it was playing. Apple under Jobs knew what game it was playing.
Stac did not know what game it was playing when it entered negotiations with Microsoft. Friendster did not know what game it was playing when it validated the market for others to take advantage of.
If this kind of clarity can help you untangle your project’s decision-making processes, let’s get in touch.
Appendix: 30 Cases for the Three Strategic Models
The purpose of this appendix is not exhaustiveness but calibration: to show that the three strategic models are patterns that repeat across completely different industries, eras, and scales.
Model 1: Exploitation and Exit
This model works when the inventor honestly recognizes that their advantage is temporary and scales their ambitions accordingly. It is not a strategic failure, but a conscious and legitimate strategy.
- The Million Dollar Homepage (2005). Alex Tew sold one million pixels at one dollar each to pay for university. The concept was irrepeatable by definition (there is only one “first site to do this”) and he knew it. He collected, closed, and did not try to build an advertising business.
- Flappy Bird (2013). Dong Nguyen created the game in two days, saw it reach 50 million downloads, and voluntarily withdrew it when public attention became unbearable. He deliberately left money on the table.
- Wordle (2021). Josh Wardle created the game for his partner. It went viral. He sold it to the New York Times for seven figures without intending to become a video game studio. He identified the moment of maximum value and exited.
- Hoverboards (2015). The holders of the first patents aggressively licensed during the demand peak and withdrew before the flood of Chinese clones collapsed margins. They played the time window with precision.
- Pet Rock (1975). Gary Dahl sold rocks with instruction manuals and became a millionaire in six months. He never pretended it was anything else. It is the purest case of the model: he invented a fad, exploited it, and exited.
- Fidget Spinners (2017). The demand peak lasted less than a year. Manufacturers who understood this exploited the window and exited. Those who believed it was a lasting category were left with inventory.
- Snuggie (2008). A blanket with sleeves sold via infomercial. The brand was squeezed until natural interest exhaustion without attempting diversification.
- Billy Bob Teeth (1994). Tom Shepard designed novelty dentures. Generated millions in the Halloween and novelty market without seeking expansion.
- The Rubik’s Cube (1980s). Erno Rubik massively licensed during the cultural peak of the 1980s. The cube faded, resurfaced decades later, and the licensing model sustained the business in both waves.
- Aereo (2012-2014). A company that retransmitted TV signals via miniaturized individual antennas, exploiting a legal loophole. Operated until the U.S. Supreme Court closed the loophole. It was a model built explicitly on a temporary legal window.
Model 2: The Ladder
This model requires long-term vision and the discipline to invest in the next step before the current one collapses. It is the game of organizations that survive decades.
- Gillette. From the double-edge razor to Mach 3, to Fusion, to ProGlide. Each product generation is launched when the previous patent is about to expire, maintaining exclusivity continuously since 1901.
- Intel. The “Tick-Tock” model: one year they optimized the manufacturing process (Tick), the next they renewed the architecture (Tock). They released a new generation just as the previous one was becoming commoditized.
- Pfizer. Viagra → Lipitor → COVID vaccines. Each megablockbuster funds research for the next. When one patent expires, the next product already has years of accumulated development.
- Qualcomm. Built iterative patent portfolios on each generation of mobile telephony: CDMA, 3G, 4G, 5G. Each new standard incorporates its essential patents before anyone else can compete in that space.
- Dyson. From the bagless cyclonic vacuum to the bladeless fan to the supersonic hair dryer. Each product carries the same logic of radical redesign of an everyday object into a new category.
- 3M. Patent portfolio in adhesives and abrasives that is constantly renewed. Post-it was an accident that became a product; that ability to capitalize on lateral findings is part of the model.
- Corning. Pyrex in the 1910s, fiber optics in the 70s, Gorilla Glass in 2007, Ceramic Shield in 2020. Each specialized glass technology is born from the previous one.
- ARM. Each new version of its processor architecture is designed to be the standard for the next generation of devices. Their business is licensing architectures, not manufacturing chips, which allows them to be everywhere without manufacturing anything.
- DJI. Aggressive iteration of drones and camera stabilizers. When a competitor manages to replicate the current model, DJI already has the next one on the market.
- DuPont. Nylon in the 30s, Kevlar in the 60s, Teflon in the 70s, Lycra, Tyvek. A century of advanced materials ladder where each invention funds research for the next.
Model 3: The Intangible Moat
This model requires patience and the conviction that the community built is more valuable than any technical or legal protection. It is the most difficult to execute and the most difficult to destroy once achieved.
- Linux / Red Hat. The Linux kernel is free and open. Red Hat built a business of services, support, and certifications on that base. IBM bought Red Hat in 2019 for $34 billion. Not for the code, but for the community, reputation, and service contracts.
- Wikipedia. The technology of a wiki can be replicated in hours. The community of editors maintaining over 60 million articles with verification and neutrality standards is not. That is the real asset.
- Stack Overflow. A technical Q&A forum. The technology is trivial. The asset is the accumulated base of community-verified solutions and the professional reputation programmers built there over years.
- Airbnb. Accommodation booking technology is trivial. The real asset is the trust network between hosts and guests built over millions of verified transactions. Replicating that trust from scratch requires time that money cannot buy.
- GitHub. Git is an open version control protocol anyone can use. GitHub is the social network where programmers built their professional reputation and where the world’s most important open source projects live. Microsoft paid $7.5 billion for that intangible.
- Craigslist. 1990s technology, design unchanged for two decades, yet it remains the dominant local classifieds in the U.S. Its asset is local liquidity: the number of buyers and sellers in each city that makes the platform work. No competitor with better technology could replicate that critical mass.
- Waze. GPS routing is Level 0. Any map app can calculate a route. Waze’s asset is the real-time reporting from millions of drivers on accidents, police, and traffic closures, data that only exists if there is an active community generating it. Google bought Waze in 2013 for $1.1 billion for that data and community.
- WordPress. The content management software is completely commoditized. The moat is the community of tens of thousands of plugin and theme developers who chose WordPress as the platform for their businesses. Migrating that ecosystem is practically impossible.
- Roblox. The graphics engine is secondary to what Unity or Unreal offers. The real asset is the network of millions of creators who built games on Roblox and the millions of users who play those games. The platform is worth what it is because of the community that inhabits it.
- Strava. Recording a GPS activity is a function any sports watch or fitness app has. Strava’s asset is the social network of athletes: the segments with rankings, KOMs, clubs, reputation accumulated over years of activity. That does not migrate with a button.
Santiago Bustelo
July 2026