MASTERCLASS Patents for Entrepreneurs – Part 5: Patent Mining
Ola Wassvik covers how to extract patentable ideas and trade secrets out of a team, based on running the process with over a hundred companies. The core idea is that every solved problem creates IP, not every solution, since many solutions are just part of doing business rather than something novel. The golden rule is never to ask "what can we patent," since that only surfaces ideas from the handful of engineers who already think of themselves as inventors. Instead, ask "what problems did we solve that others still struggle with," which draws out ideas from the whole company, not just engineering.
Key basics
The workflow is to log a problem, capture how it was actually solved, classify it by layer (conceptual, product, system, or component), and then judge its visibility, since anything invisible to an outsider is a weak patent candidate. This shouldn't be a one-time exercise. Run it at the start of a project, partway through, and before going public, since most valuable IP surfaces during development and customer integration. It also works best assigned to someone naturally process-driven, like a product manager, rather than a fast-moving CTO who'll only do it once.
The four layers each need different people in the room: conceptual patents (customer-facing and business people, covering a whole business segment), product patents (design, manufacturing, QA, and operations), system patents (integration and ecosystem partnership teams, protecting how your product works with others), and component patents (core technology and systems architects, the layer engineers naturally gravitate toward first).
Strategy
To decide patent versus trade secret, use a reverse engineering test: could someone without your source code or manufacturing line spend a thousand hours and a million dollars to figure it out? If yes, patent it. If no, keep it secret, though this threshold matters less in a billion-dollar market than a million-dollar one. Beyond a simple yes or no, there are two more buckets: maybe later (hard to discover, low business value for now, worth revisiting) and strategic patents (hard to discover, but worth patenting anyway because of high legal value in a large market).
As a portfolio mix, start from a rough 50/50 split between patents and trade secrets, skewing toward more patents if your product is highly visible or physical, and more trade secrets if it's software-heavy. If you're licensing any part of your technology, expect that number to climb further, since anything you must disclose to a licensee needs patent protection first. Always log trade secrets you choose not to patent, since that record protects your own freedom to operate later.
About the Host
Ola Wassvik is CCO and co-founder of Lightbringer, a serial entrepreneur, and an M.Sc. in Engineering Physics. Before Lightbringer, he spent 20+ years as a tech entrepreneur, including as CTO and co-founder of Flatfrog.
Across that time he's managed over 300 patent families and helped raise more than €84 million for deep tech ventures, giving him a founder's-eye view of how patent strategy actually plays out in the room with investors, competitors, and boards.
Read more →
FAQ
What's the golden rule for running a patent mining session?
Never ask "what can we patent." Ask "what problems did we solve that others still struggle with." That question draws ideas out of the whole company, not just the engineers who already see themselves as inventors.
Who should be involved in patent mining?
The whole company, not just engineering. Conceptual patents need customer-facing and business people, product patents need design and manufacturing, system patents need integration and partnership teams, and component patents need core technical teams.
How do I decide whether to patent something or keep it secret?
Ask whether someone without your source code or manufacturing access could reverse engineer it by spending a thousand hours and a million dollars. If yes, patent it. If no, it's usually safer as a trade secret, though market size affects how much that threshold matters.
What's a good default split between patents and trade secrets?
Start around 50/50, then adjust. Skew toward more patents if your product is highly visible or hardware-based, and toward more trade secrets if it's software-heavy. Licensing any part of your technology tends to push the number of patents up further.
How often should patent mining happen?
Multiple times across a project, not once. Run it at the start, partway through development, and before going public, since most valuable IP surfaces gradually through development and customer work.
DISCLAIMER: THIS IS NOT LEGAL ADVICE. YOU SHOULD CONTACT A PATENT ATTORNEY IF YOU NEED A FORMAL ASSESSMENT OF PATENT INFRINGEMENT OR FREEDOM TO OPERATE.
.png)
.webp)