Startup Patent Strategy

Timing is the most critical decision in startup patent strategy: THE PRIORITY DATE IS EVERYTHING: the priority date determines what constitutes prior art; any public disclosure BEFORE the priority date = prior art against your own patent; under AIA, you have a 1-year grace period for your own disclosures — but only in the US; no grace period in Europe, China, Japan, and most countries; RULE: file before ANY public disclosure in countries where you want protection; FILE BEFORE THESE TRIGGERS: product demo at a conference; pitch deck shared outside NDA; beta users (public use); press release or website launch; crowdfunding launch (Kickstarter); academic paper submission; any non-NDA conversation with investors or potential customers; PROVISIONAL APPLICATION TIMING: file a provisional application ($320 large entity / $160 small / $80 micro) to establish a priority date; provisional gives 12 months to file the formal non-provisional; provisional does NOT need to be perfect — but it must fully describe the invention; the non-provisional claims the provisional's filing date for any invention described in the provisional; FUNDING ROUND TIMING: Series A: investors expect some IP protection or at least pending applications; a provisional filed before investor conversations is often sufficient; a pending non-provisional adds more credibility; Series B/C: investors do more IP due diligence; actual issued patents or clear prosecution strategy expected; RULE: file provisional applications early and often — they are cheap ($320-$1,500 attorney preparation) and buy 12 months while the startup validates the product; DON'T WAIT for: the product to be 'finished' (file the MVP invention, improve in continuations); funding (the cost of a provisional is small compared to the value created); certainty (file now, optimize later).

Startups have limited IP budgets and must prioritize wisely: PATENTING CRITERIA — FILE IF: (1) the invention is NOVEL: would a prior art search reveal prior art? if yes, don't waste money on an application likely to be rejected; do a quick prior art search before committing to file; (2) the invention has commercial value: does this invention protect the core product feature? is this how the startup makes money? would a competitor copy this if there were no patent?; (3) the invention is NOT the secret sauce: trade secrets are often better for core algorithms (the secret sauce that makes the product work); patents require public disclosure after 18 months — the full technical description becomes public; (4) the invention is relatively easy to reverse-engineer: if a competitor can figure out how you do it from the product, patent it; if they can't figure it out from the product, trade secret it; PATENTING CATEGORIES FOR STARTUPS: CORE PRODUCT FEATURES: the specific technical implementation that delivers the value; novel algorithms with technical improvement framing (not just 'doing it on a computer'); novel hardware configurations; novel methods that improve technical performance; PLATFORM/API: APIs and platform architectures that competitors might copy; the interfaces others will build on; USER EXPERIENCE INNOVATIONS: novel UI flows with specific technical implementations; accessibility innovations; performance optimizations; DON'T PATENT: abstract business processes that lack technical implementation details; anything easily searched by competitors with straightforward prior art; features that will change dramatically in the next 12-24 months; features that are impossible to monitor for infringement (internal server processes); TRADE SECRETS vs. PATENTS: algorithms: if they run server-side and are undetectable from product behavior → trade secret; if detectable or easily reverse-engineered → patent; training data: trade secret (not patentable); model architecture: patent if novel and detectable.

Patent prosecution costs are manageable with the right strategy: ENTITY STATUS DISCOUNTS: SMALL ENTITY (≤500 employees; not university; not partner with large entity): 50% reduction on all USPTO fees; example: utility application basic fee $1,960 large → $980 small entity; MICRO ENTITY: 80% reduction; eligible if: (a) qualifies as small entity; AND (b) has ≤4 previously filed patent applications; AND (c) gross income ≤3× the preceding calendar year's median household income (~$135,000-$145,000 based on $45,000-$48,000 median HH income × 3); AND (d) assigned to an institution of higher education (or not assigned to such an institution); cost: utility application basic fee $1,960 large → $392 micro entity; PROVISIONAL APPLICATION STRATEGY: file provisional applications early ($80-$160 USPTO fees + $500-$1,500 attorney); the provisional establishes priority date and buys 12 months; use that 12 months to: validate the product; raise a funding round; decide which inventions merit non-provisional filing; a well-prepared provisional saves money by: reducing the work needed for the non-provisional; allowing focused investment after product-market fit validation; ATTORNEY COST REDUCTION: work with a patent attorney who understands your technology (reduces back-and-forth); draft a detailed technical memo before attorney meetings (reduces billing time); use less experienced attorneys at large firms supervised by a partner (rates $200-$350 vs. $500-$600+); consider boutique patent law firms specializing in your technology area; PATENT PROSECUTION ROADMAP: set a budget: early-stage: $15,000-$30,000 per patent application through issuance is reasonable; plan to file 3-5 applications in the first 2 years; avoid filing patent applications you cannot afford to prosecute through issuance; DEFENSIVE PUBLICATION ALTERNATIVE: if cost is prohibitive, a defensive publication prevents competitors from patenting the same invention (published disclosure = prior art); free to publish; doesn't give you a patent, but denies it to competitors; IP VALUATION CONSIDERATION: one strong patent is worth more than ten weak ones.

Patent portfolios play an important role in startup fundraising conversations: WHAT VCS LOOK FOR IN IP DUE DILIGENCE: pending patent applications (published or unpublished): shows the startup took IP seriously before launch; provisional applications: acceptable for seed/pre-seed, less so for Series A+; issued patents: highest credibility; patent portfolio in strategic technology areas; freedom-to-operate: no blocking patents from competitors; FTO opinion from qualified counsel; PATENT PORTFOLIO AS COMPETITIVE MOAT SIGNAL: investors want to see that the technology advantage is protectable; patent applications signal: 'we have something novel'; 'we've invested in protecting it'; 'competitors can't simply copy this without paying us or challenging the patent'; for B2B SaaS: IP ownership is particularly important — enterprise customers want confirmed IP ownership; PATENT PORTFOLIO IN M&A: acquired startups frequently have their patent portfolios separately valued in M&A; patents can represent 20-40% of acquisition value in IP-intensive sectors; even a small portfolio (5-10 solid patents) can significantly increase acquisition price; WHAT TO SAY TO INVESTORS: 'We have [N] provisional patent applications covering our core technology filed [date]'; 'We have [N] non-provisional applications pending at the USPTO'; 'Our [technology feature] is covered by [patent number] which issued [date]'; 'We have a freedom-to-operate opinion from [firm] covering our core product features'; 'Our patent strategy focuses on [specific technical improvements] that are fundamental to our competitive position'; WHAT TO AVOID: do not claim patents on obvious or prior-art-covered features; investors who do IP due diligence will find this; do not over-claim — 'patent pending' does not mean 'patent approved'; be prepared to show the actual applications/claims during due diligence; IP DEFENSIVE VALUE: a startup with a patent portfolio is much harder to bully in litigation; a threatening letter from a patent troll or competitor is less effective when the startup can respond with its own patents; FRAND/essential patents: in platform companies, SEP positions can be a revenue source.

The patent vs. trade secret decision is fundamental to IP strategy: PATENT ADVANTAGES: 20-year monopoly right (from filing date); public notice to competitors; government-enforced right to injunction and damages; licensable asset; assignable in M&A; PATENT DISADVANTAGES: full public disclosure of the technology (18 months after filing); expensive to obtain and maintain; takes 2-3 years to issue; can be challenged and invalidated; doesn't protect ideas only their specific implementation; TRADE SECRET ADVANTAGES: unlimited duration (as long as kept secret); no public disclosure; no cost to 'obtain' (just keep it secret); protects what patents can't (data, formulas, customer lists, algorithms); immediate protection (no prosecution delay); TRADE SECRET DISADVANTAGES: independent development by a competitor is a complete defense (no infringement); reverse engineering of the product is a complete defense; DTSA (Defend Trade Secrets Act 2016): federal civil action; actual + exemplary damages + attorney fees; but enforcement is expensive and difficult; no monopoly right — cannot block legitimate independent invention; DECISION FRAMEWORK: PATENT: when the technology can be detected in the product (reverse-engineerable); when competitors might independently invent the same thing; when the company wants to license the technology; when the company wants to signal innovation to investors/acquirers; TRADE SECRET: when the technology cannot be detected from the product (server-side algorithms; proprietary training data); when competitors cannot reverse-engineer it; when the competitive advantage comes from operational knowledge; when the technology changes too fast for the 3-year patent prosecution timeline to be useful; HYBRID STRATEGY: patent the detectable aspects (novel UI/hardware implementation); trade secret the server-side core (model weights; training data; proprietary datasets); EARLY-STAGE STARTUP DEFAULT: file provisional on everything that might be novel (cheap); decide patent vs. trade secret when converting to non-provisional (more information, more resources available).