Patents Or Trade Secrets? 8 Questions To Ask

A s the cost of protecting patents escalates, many ventures ask themselves whether it would be better to keep their IP a secret. As part of a clear plan for building an IP portfolio that fits their needs and actually makes a difference, they should consider both, protecting their revenue streams and putting legal theory second. It isn’t about doctrine; it’s about protecting the elements that create revenue in the first place – now and when you enter larger rounds of funding. Start with your assets, not the law. Start listing the specific parts of your know-how that create revenue, increase margins or customer loyalty. Then decide how you try to suitably protect each element of your know-how. If a sufficiently sophisticated entity, such as a competitor, can analyse your products and has a reasonable chance of revealing your creation (for example, by analysing mechanical/electrical features, visible geometries, measurable protocols, parsing code, data traffic etc), there is a strong case for a patent. It offers exclusivity even against independent new developments, creates a licensable asset, especially in the case of standardisation and interoperability, and acts as a clear signal in the investor process. If details must be made public anyway (for example, through regulatory approvals with subsequent publication, open interfaces or open-source components), secrecy is also not a viable option.

For ventures now scaling, at least 80 percent of their value is ultimately going to rest on their IP, if all goes to plan. Yet it’s a far from certain outcome. Time and again, IP can appear strong on paper, but unravel at different points through the growth cycle. No one solution ever fits all. Instead, it depends on the decisions that the principals take to find the right combination of IP. In this book, IP leaders of today and tomorrow draw on real-life commercial scenarios to map out how ventures can make the right calls to turn their IP into a high-growth asset with the capacity to double or triple in value, not just once, but repeatedly.

You should always also be aware that you are not the only one on the planet who relies on smart people or sophisticated algorithms. There is a high population of that species out there, who are likely to make similar or even the same technical considerations as yours. It is particularly valid in a hyped field of technology with a far-reaching commercial impact where everyone is trying to solve essentially the same problem. In such a case, you are highly likely to see independent parallel developments that lead to the same solution of that problem. So you would be advised not to keep your creation secret beyond market entry but try to be the f irst one to stake your claim using patents. 
A secret, of course, is the one to go for if the subject matter is not patentable (due to its legal exclusion, for example, as an abstract idea related to software). Likewise, a secret is always advantageous when the element creating the value of the solution lies behind a firewall and cannot be reliably identified from the outside. These are typically processes or data values that are difficult to observe or reverse engineer, such as fermentation and coating processes, catalyst and temperature windows, proprietary datasets and recipes. 
In such cases, reverse engineering would be either technically unfeasible or commercially unattractive. Secrecy protection is particularly strong when the half life of the commercial relevance of your know-how significantly exceeds the typical patent term. Classic examples are beverage recipes or specialised process management. 
On the other end of the spectrum, if a patent application would reveal too much too early, keeping it secret may be the better option. Secrecy can help especially those who develop in short iterations and learn incrementally in retaining their competitive advantage rather than helping to put their competitors on a valuable learning curve. 
For example, you are testing ten variations of a solution. If you share each variant publicly, your competitors do not have to run those tests themselves. T hey simply adopt your best result or possibly even only the one that suits their needs better. If you keep your results to yourself, you retain this advantage. Cost and timing considerations also speak in favour of this approach: international reach can be achieved without the complexity of a broad patent portfolio, provided that the necessary technical and organisational protective measures are taken. This is particularly valid if you are operating in a market with short product life cycles where it is highly likely that your current solution will be outdated sooner than later. 
In less short-lived markets, patents are more suitable, especially for detectable implementations, such as mechanical structures, electric circuitry layouts, processing and communication protocols, and encodings that can be revealed by analysis and testing. In short, start with the value driver, choose the protection mechanism based on observability and replicability, and combine both where commercial logic dictates. 
 

Trade secrets only come to life with real measures 

Make no mistake, trade secrets and their legal protection only exist if they are backed by effective, verifiable measures, which the courts will expect. They won’t look kindly on any deficiencies. To start with, lean but genuine governance is sufficient. This includes, firstly, the systematic classification and labelling of sensitive information and, secondly, strict access control based on the need to-know principle with role-based rights. Thirdly, contract hygiene is a must: clear obligations for employees and contractors, robust non-disclosure agreements and, where legally permissible, pilot clauses on no reverse engineering. Fourthly, short, recurring training courses for your staff with documented participation are required.Fifthly, suppliers must be integrated via restrictions on use and data utilisation, audit rights and targeted segmentation in such a way that no single partner can reconstruct the whole.

Legal limits must, of course, be observed: whistleblowing and disclosures in the public interest are often exempt; reverse engineering of legally acquired products is often permissible unless it has been effectively restricted by contract. As noted, if practical reverse engineering is likey to reveal the secret, patent protection is usually the safer option. 
Leaders and executives are obliged to introduce, enforce and monitor confidentiality measures; persistent neglect can also be personally risky. A viable protection concept therefore combines legal, technical and organisational controls: information is classified, access is restricted, employees are trained and evidence is kept. Internal agreements should focus on confidentiality (while observing the limits of inadmissible non-competition clauses). External R&D and supplier contracts require modern provisions on confidentiality, IP ownership, scope of use and audit rights. The key point is that measures must not just exist on paper: they must work and be applied in everyday practice, and they must be verifiable.

Cost logic: high one-off costs v moderate ongoing costs

The costs of patent protection arise primarily at the beginning: conceptual and legal expenses are incurred for describing the invention and preparing the application, followed by official fees and representation in the examination procedure. With each additional jurisdiction, the costs of translation, representation and administration increase. International strategies spread the costs over time, but increase them overall. After the patent is granted, ongoing maintenance fees are added and increase over its lifetime. If the right is enforced or defended, procedural costs can be considerable. Patents are particularly suitable in terms of cost logic if the protected technology is detectable, imitation would be possible without disclosure, or investments and licenses require a formal exclusive right. 
In the case of protecting trade secrets, the main costs are not incurred on a one-off basis, but rather on an operational basis. There is no registration; what is crucial are appropriate, verifiable measures. These include governance and policies, clear contract design with employees, contractors and suppliers, access concepts and logging, training, as well as technical and physical safeguards. An initial programme incurs introduction and implementation costs; thereafter, there is a constant need for maintenance and monitoring. Here, too, relevant enforcement costs may arise in the event of a dispute. Trade secrets are cost effective if value drivers lie in processes, data pipelines, parameterisations or service recipes that are difficult to observe, and if the organisation has the means and discipline to permanently implement and document protective measures. So, bottom line: real trade secret protection is by no means free of charge. 
Viewed over several years, patent protection is characterised by high initial and country-specific additional costs, while secrecy protection incurs lower peaks but continuous operating and compliance costs. In practice, a hybrid approach often dominates: publishable principles are patented, while difficult to-detect detailed parameters, data sets and internal processes are kept as trade secrets. Which cost structure prevails depends on the technology, complexity, degree of internationalisation and the ability to achieve sustainable compliance.

Mini-cases: from global corporations to medium sized companies

Coca-Cola is a typical example of in-depth governance of trade secrets. Since 1886, sources have consistently reported compartmentalisation: no single employee knows both the recipe and the entire end-to-end production process; critical steps are separated from each other and ritualised. What works for a global corporation can also be effectively implemented in a medium-sized family-owned manufacturer that uses a basic recipe that is known only to the managing director and two family members. They mix it after work; the employees use the ready-made mixture the next day. This is effective secrecy by design: small circle, separation of steps, no way to reconstruct. 
Regardless of whether you run a global brand or not, the lessons for trade secret protection can be broken down: divide critical know-how into isolated areas; minimise the number of people with a complete overview; outsource steps where possible to sealed premixes or black-box deliveries; and prefer zero disclosure to a patent if your product is virtually impossible to reverse engineer. If you never disclose the recipe, you teach your rivals nothing. But keep in mind: you are not the only smart one on the planet. 
The trade secret philosophy can be scaled: protect what the market cannot see by limiting the circle of people who can see it, documenting this restriction and concluding contracts against misuse.

A decision tree for your IP strategy (which you can use tomorrow)

Ask yourself eight practical questions for each asset.

1. Detectability and reverse engineering

Is the asset recognisable from the outside or can it be traced through testing, disassembly or benchmarking? 
• Yes, easily visible or reproducible: patent. 
• No, not recognisable: trade secret. 

2. Replicability by third parties

Is independent redevelopment by competent third parties likely (state of the art, tools, team size)? 
• Yes, re-creation realistic: patent. 
• No, re-creation unlikely: trade secret. 

3. Lifespan and rate of change 

Is the economic value short-lived (months, rapid iterations) or long-lasting (years/decades)? 
• Long-lasting, rarely changed module: patent. 
• Short cycles / frequent adjustments: trade secret. 

4. Patentability and need for exclusion 

Is the asset (a) patentable (no grounds for exclusion, inventive step plausible) and (b) do you need strong exclusionary effect against everyone (core technology / platform)? 
• (a) yes and (b) exclusion desired: patent. 
• (a) no/uncertain and no external visibility: trade secret. 

5. Organisational capacity for secrecy 

Is reliable secrecy realistic (need to know, contracts, protocols, training)? 
• No, governance not sustainable: patent. 
• Yes, governance sustainable: trade secret. 

6. Disclosure and timing 

Is disclosure imminent (launch, approval, standardisation, publication)? 
• Yes, disclosure foreseeable, so apply in advance: patent. 
• No, no disclosure planned and low recognisability: trade secret

7. Budget and time 

Are the budget and schedule compatible with a patent application, including duration/costs? 
• Yes, resources available: patent 
• Budget/time limited; ongoing confidentiality measures feasible: trade secret 

8. Separability of external/internal (hybrid option) 

Can visible external components and internal parameters/recipes be cleanly separated? 
• Arguments in favour of hybrid: patent externally visible features; treat internal fine-tuning (parameters, tolerances, recipes, pipelines) as trade secrets.

The holistic IP stack that investors expect 

High-growth ventures rarely rely on a single type of protection. Successful portfolios combine the following: 

• Patents: if know-how is detectable and replication by third parties is possible, a strong exclusionary effect vis-à-vis everyone is required 

• Trade secrets: if know-how is difficult to reverse engineer, has a long half-life, is not (yet) patentable or too sensitive to disclose, and is quickly iterated. 

• Copyright and database rights: for code, documentation and data sets (as well as clear agreements with contributors). 

• Trade marks and designs: for market access and premium prices. 

• Contracts as a link: NDAs, assignment clauses, no reverse engineering where legally permissible, restrictions on use, rights of inspection. Document who owns what, the legal basis for use and how you can grant licenses. 

Five takeaways 

Start with the value driver, not the law 

Choose the protection mechanism based on observability and replicability. Visible and easily reverse-engineered features → more likely to use patents; internal, difficult to-observe processes/data with a long half-life or non patentable components → more likely to be trade secrets. 

Decide systematically (eight-question check) 

Check detectability, replicability, lifespan/iteration rhythm, patentability/need for exclusion, organisational capability of secrecy, disclosure timing, budget/time – and use hybrid solutions: patent externally visible features, keep internal parameters/recipes secret. 

Secrets only exist with governance 

Verifiable measures are mandatory: classification, access based on need-to-know, clean contracts (NDAs, possibly no reverse engineering), training, supplier controls. Observe legal limits (whistleblowing, permissible reverse engineering). 

Realistically assess cost logic 

Patents: high initial and additional country costs plus maintenance; worthwhile for external visibility, licence/ investor expectations. Secrets: no official fees, but ongoing organisational/compliance costs; advantageous for process/data values and disciplined implementation. In practice, the hybrid often dominates. 

Investor-friendly IP stack 

Value comes from a combination of patents (visible functions/protocols), trade secrets (server-side logic, manufacturing windows, datasets), copyright/database rights, trade marks/designs, and strong contracts and data provenance documentation. The goal is a portfolio that is monetisable and defensible.

Der Artikel wurde im Januar 2026 erstmals in: Winning with IP - Managing Intellectual Property Today, 2025/26 edition - Novaro Publishing veröffentlicht.

Header: AI-generated using ChatGPT (OpenAI)