Summary

This framework helps companies decide how to protect intellectual property by choosing between three primary strategies: patents, trade secrets, and open source release. The decision depends on four key variables: industry type, innovation characteristics (especially reverse-engineering risk), competitive goals, and budget constraints. Default recommendation: use a hybrid strategy that patents externally visible innovations while keeping internal processes as trade secrets, unless your competitive goal is ecosystem-building (which favors open source). [src1]

Constraints

An innovation audit (inventory of protectable assets) must be completed before this decision is meaningful — without knowing what you have, you cannot choose how to protect it
Patent applications are published 18 months after filing, permanently disclosing the invention — this disclosure cannot be retracted even if the patent is abandoned [src1]
Trade secret protection is permanently destroyed by a single breach — once information becomes public through any means, it cannot be re-classified as secret [src2]
Open source release under a permissive license (MIT, BSD, Apache) is irreversible — you cannot reclaim proprietary rights after public release
Decision-makers frequently underestimate the total cost of patent prosecution and maintenance — budget for $15K-$25K per patent over its lifetime, not just the filing fee [src5]

Decision Inputs

Input	Why It Matters	How to Assess
Industry type (pharma/software/hardware/CPG/services)	IP norms, enforcement costs, and regulatory requirements vary dramatically by industry — pharma relies heavily on patents while software increasingly favors trade secrets and open source	Ask: "What industry are you in?" Cross-reference with industry IP norms
Innovation type (product/algorithm/process/business method/data)	Determines patentability eligibility and reverse-engineering risk — physical products are easily reverse-engineered; internal algorithms are not	Ask: "What kind of innovation are you protecting? Is it visible in the product or hidden behind the scenes?"
Competitive goal (block/license/investors/standard/ecosystem)	Directly determines which IP strategy aligns with business objectives — blocking competitors favors patents; ecosystem-building favors open source	Ask: "What is your primary goal — preventing copying, generating licensing income, attracting investors, or building an ecosystem?"
Reverse-engineering risk (easy/difficult/impossible)	The single most important factor — if competitors can reverse-engineer the innovation, trade secret protection is worthless and patents are necessary [src1]	Ask: "If a competitor bought your product, could they figure out how it works?"
Budget and timeline constraints	Patent prosecution costs $10K-$25K per patent and takes 2-3 years; trade secret programs require ongoing investment; open source has near-zero protection cost but zero exclusivity	Ask: "What is your IP protection budget? Do you have 2-3 years before you need enforceable protection?"

Decision Tree

START — How should we protect this innovation?
├── Can competitors reverse-engineer this innovation from the product?
│   ├── YES — Easy to reverse-engineer (visible in product)
│   │   ├── Is it patentable (novel, non-obvious, useful)?
│   │   │   ├── YES → RECOMMEND: Patent
│   │   │   │   Reason: Trade secrets provide zero protection for reverse-engineerable innovations
│   │   │   │   Constraint: Budget $10K-$25K per patent; timeline 2-3 years
│   │   │   └── NO (abstract idea, natural phenomenon, or prior art exists)
│   │   │       ├── Goal = establish industry standard?
│   │   │       │   ├── YES → RECOMMEND: Open Source (Apache 2.0 with patent grant)
│   │   │       │   └── NO → RECOMMEND: Copyright + speed-to-market
│   │   └── Do you want licensing revenue from this innovation?
│   │       ├── YES → RECOMMEND: Patent (utility or design)
│   │       └── NO → RECOMMEND: Patent for blocking + trade secret for details
│   └── NO — Difficult or impossible to reverse-engineer
│       ├── What is your competitive goal?
│       │   ├── Block competitors / defensive positioning
│       │   │   ├── Can you maintain robust confidentiality controls?
│       │   │   │   ├── YES → RECOMMEND: Trade Secret
│       │   │   │   │   Reason: Indefinite protection without disclosure
│       │   │   │   └── NO (many employees, high turnover)
│       │   │   │       → RECOMMEND: Patent (defensive filing)
│       │   │   └── Manufacturing process or formulation?
│       │   │       ├── YES → RECOMMEND: Trade Secret (classic use case)
│       │   │       └── NO → Evaluate based on employee access scope
│       │   ├── Attract investors / increase valuation
│       │   │   └── RECOMMEND: Patent portfolio + trade secrets
│       │   │       Reason: Startups with patents are up to 10x more likely to secure funding
│       │   ├── Establish industry standard
│       │   │   └── RECOMMEND: Open Source (copyleft or permissive)
│       │   │       Reason: Standards require adoption; open source maximizes speed
│       │   └── Build developer ecosystem
│       │       └── RECOMMEND: Open Source core + proprietary extensions
│       │           Reason: Open core drives adoption while monetizing advanced features
├── OVERRIDE CONDITIONS:
│   ├── Pharma/biotech with novel compound → Patent required
│   ├── Budget < $5K total → Trade secret only
│   ├── AI/ML model weights or training data → Trade secret preferred
│   ├── Government contract → Check DFARS/FAR IP clauses
│   └── Multiple co-inventors across companies → Patent preferred
└── DEFAULT:
    └── RECOMMEND: Hybrid — patent visible aspects, trade-secret internal processes
        Reason: Lowest-risk approach combining enforcement power with indefinite protection

Options Comparison

Factor	Patent	Trade Secret	Open Source
Typical cost range	$10K-$25K per patent (US only); $50K-$100K international	$5K-$30K/yr (program setup + maintenance)	$0-$5K (legal review of license)
Timeline to protection	2-3 years (12 months with accelerated exam)	Immediate (once controls are in place)	Immediate (upon publication)
Duration of protection	20 years from filing date	Indefinite (as long as secrecy maintained)	None (no exclusivity)
Risk level	Medium (prosecution uncertainty, invalidation)	High (single breach destroys protection)	Low (no protection to lose)
Reversibility	Hard (disclosure at 18 months is permanent)	Moderate (can later file patent if still novel)	Irreversible (cannot reclaim proprietary rights)
Internal capability needed	Patent attorney, technical disclosure drafting	InfoSec, NDA infrastructure, employee training	Community management, license compliance
Best when	Innovation is reverse-engineerable, need licensing revenue or investor-grade IP	Innovation hidden from view, difficult to reverse-engineer, strong confidentiality	Want ecosystem adoption, standard-setting, or community development
Worst when	Short commercial life (<3 yrs), cost exceeds value, Alice exclusions apply	High turnover, many need access, innovation visible in product	Need exclusivity, licensing revenue, or competitive blocking
Hidden costs	Maintenance fees ($2K-$8K per window), international ($30K-$80K/jurisdiction), litigation ($1M-$5M)	Breach remediation ($100K-$500K), exit procedures, ongoing audits	Lost licensing revenue, competitor forks, community management ($50K-$150K/yr)

[src1, src3, src5]

Decision Logic

If reverse_engineering_risk = easy AND innovation is patentable

→ Patent. When competitors can reverse-engineer an innovation from the product itself, trade secret protection provides zero value. Patents are the only mechanism that provides legal exclusivity for publicly visible innovations. [src1]

If reverse_engineering_risk = impossible AND confidentiality controls are strong

→ Trade Secret. Internal processes, formulations, and algorithms that competitors cannot observe from outside the company are ideal trade secret candidates. Protection lasts indefinitely without the 20-year patent expiration. [src2]

If competitive_goal = ecosystem OR competitive_goal = standard_setting

→ Open Source. When the strategic objective is adoption, standard-setting, or developer ecosystem building, open source maximizes distribution speed. Use Apache 2.0 for maximum adoption with patent protection, or GPL/AGPL to prevent proprietary forks. [src4]

If industry = pharma/biotech AND innovation = novel compound

→ Patent (mandatory). Pharmaceutical regulatory exclusivity is structurally tied to patent protection. FDA market exclusivity periods are linked to patent listings in the Orange Book. [src7]

If competitive_goal = attract_investors AND company = startup

→ Start with provisional patents, expand to full utility patents at Series A. Startups with patents are up to 10x more likely to secure funding. Provisional patents cost $1.5K-$3K and establish a priority date with 12 months to file the full application. [src3]

Default recommendation

→ Hybrid: patent externally visible innovations + trade-secret internal processes. Most companies have both visible and hidden innovations. The hybrid approach maximizes protection breadth without requiring either full patent portfolio costs or perfect confidentiality across the entire organization. [src1]

Anti-Patterns

Wrong: Treating the patent-vs-trade-secret decision as binary

Companies assume every innovation must be either patented or kept secret, missing the hybrid approach. This leads to either over-spending on patents for innovations that could stay secret, or leaving reverse-engineerable innovations unprotected. [src7]

Correct: Decompose innovations into components and protect each appropriately

A manufacturing company can patent the machine design (visible to customers) while keeping the specific material composition as a trade secret. Software companies can patent the user-facing algorithm while keeping the training data and optimization parameters as trade secrets. [src1]

Wrong: Filing patents without a commercial enforcement plan

Startups file patents to "check the box" for investors without considering whether they would ever enforce them. An unenforced patent costs $15K-$25K to obtain and $2K-$8K per maintenance window — paying for a right you never exercise. [src5]

Correct: File patents only where you would realistically enforce or license

Before filing, answer: "If a competitor infringed this patent, would we sue? Could we afford to? Would we license it?" If all answers are no, consider trade secret protection instead. For investor signaling, file provisional patents ($1.5K-$3K) rather than full utility applications.

Wrong: Assuming trade secrets are free because there is no filing fee

Companies choose trade secret protection as the "cheap option" without investing in the confidentiality infrastructure required to maintain legal protection. Courts examine whether "reasonable measures" were taken — inadequate NDAs or lax access controls can void trade secret status entirely. [src2]

Correct: Budget $5K-$30K/year for a trade secret program

Reasonable measures include: NDAs for all employees and contractors with access, tiered access controls, regular employee training, exit interview procedures, document marking and classification systems, and periodic audits. Without these, you have information, not a trade secret.

Cost Benchmarks

Scenario	Patent (US Only)	Patent (US + 3 Intl.)	Trade Secret Program	Open Source
Single innovation, small entity	$8K - $15K	$40K - $80K	$5K - $10K/yr	$0 - $2K
5-patent portfolio, startup	$40K - $75K	$200K - $400K	$10K - $20K/yr	$2K - $5K
20-patent portfolio, mid-market	$150K - $350K	$600K - $1.5M	$20K - $50K/yr	$5K - $15K
Ongoing annual (per patent)	$2K - $8K/yr	$5K - $15K/yr	Included above	$0

Hidden cost multipliers: Add patent litigation insurance ($10K-$50K/yr for startups, $100K-$500K/yr for mid-market). Trade secret programs require exit procedure costs ($500-$2K per departing employee). Open source community management costs $50K-$150K/yr if the project gains meaningful adoption. International patent prosecution doubles the US-only cost per additional jurisdiction. [src5]

IP Strategy Decision Framework