Freedom to operate is the question of whether you can sell a product without infringing somebody else's patent. It is a commercial risk assessment, and it is distinct from every other kind of patent search.
The abbreviation is worth spelling out. FTO means many unrelated things outside patent practice, and in writing aimed at engineers, investors or executives, "freedom to operate" avoids a moment of confusion that "FTO" creates.
What FTO is not
Not patentability. This is the confusion that costs the most money.
| Patentability search | Freedom to operate search | |
|---|---|---|
| Question | Is my invention new? | Does my product infringe? |
| Looks at | Everything ever published | In-force claims only |
| Expired patents | Relevant | Irrelevant |
| Pending applications | Relevant as prior art | Relevant as future risk |
| Geography | Worldwide | Per market |
| Output | Novelty assessment | Infringement risk assessment |
You can have one without the other, in both directions. An invention can be perfectly patentable and still infringe — your improvement on someone else's patented device is novel and non-obvious, and building it infringes their claim. Equally, a product can have complete freedom to operate and be entirely unpatentable, because everything in it is old.
The date logic runs opposite. In a patentability search, older is worse — a patent from 1985 can destroy your novelty. In FTO, older is better — a patent from 1985 expired long ago and cannot be infringed.
Not an invalidity search either. Invalidity looks backwards from a patent to find prior art. FTO looks forwards from a product to find patents. Different starting point, different corpus, different output. See prior art for how the searches diverge.
What the search covers
In-force patents and published applications in every jurisdiction where the product will be made, sold, offered for sale or imported.
Patents are national rights, which is why FTO is always per-market. A US patent restricts nothing in Germany. A German patent restricts nothing in the US. A product manufactured in one country and sold in another needs clearance in both, because making and selling are separate infringing acts.
Published applications matter even though they cannot yet be infringed. An application publishing today may issue in two years with claims covering your product, and a design decision made now is much cheaper to change than one made after tooling. Applications are also a leading indicator of where a competitor is heading.
The eighteen-month blind spot is unavoidable. Applications publish eighteen months after their earliest priority date, so anything filed in the last eighteen months is invisible to any search. This is the single largest limitation on FTO and it cannot be engineered away — only accounted for.
How the search is constructed
Start from the product, not the patent. Break it into technical features, and for each feature ask what a claim covering it would look like.
Search by classification and by keyword. Classification catches patents that describe the same thing in different words, which is most of them. Keyword search catches recent filings that have not settled into a classification pattern.
Read the claims, not the abstracts. A patent whose abstract sounds alarming may have claims that are narrow and irrelevant. A patent whose abstract sounds unrelated may have a claim that reads squarely on your product. The claims are the only part that can be infringed.
Check status on everything you keep. A patent that looks like a problem may have lapsed for unpaid maintenance fees — 58.6% of US utility patents granted in 2014 never reached full term. Verify against USPTO Patent Center rather than a commercial database, which may lag by weeks.
From search to opinion
A search finds patents. An opinion analyses them. The distinction matters legally as well as practically.
| Search | Opinion | |
|---|---|---|
| Produces | A list of relevant patents | A reasoned conclusion on risk |
| Performed by | Search firm, in-house, or counsel | Patent counsel |
| Privileged | Not necessarily | Yes, if properly obtained |
| Rebuts willfulness | No | Yes, if competent and timely |
| Includes claim charts | Sometimes | For anything material |
An opinion analyses the material patents element by element, usually with a claim chart for each one that survives the first pass, and reaches a conclusion — non-infringement, invalidity, or a stated level of risk.
Non-infringement opinions and invalidity opinions are different documents. A non-infringement opinion concludes the product does not fall within the claims. An invalidity opinion concludes the claims should not have issued. Either can support a decision to proceed, and the first is usually cheaper and more durable.
Why the timing determines the value
An opinion obtained before the conduct rebuts willfulness. One obtained after a demand letter does not.
Willful infringement exposes a defendant to enhanced damages of up to three times the compensatory award under 35 U.S.C. 284. Since Halo v. Pulse the standard is discretionary, and courts look at the whole course of conduct.
A competent, written, pre-launch opinion is the strongest evidence that the conduct was not reckless. It shows the company identified the risk, took advice and formed a reasonable belief.
Section 298 cuts the other way as well. Failure to obtain an opinion cannot be used to prove willfulness. The opinion helps; its absence is not supposed to hurt. In practice, having one still matters more than the statute suggests.
| When the opinion was obtained | Willfulness value |
|---|---|
| At design, before commitment | Strong |
| Before launch | Strong |
| After launch, before any notice | Moderate |
| After a demand letter | Minimal |
| After a complaint is filed | Effectively none |
When to run FTO
At design, before the architecture is fixed. This is when a design-around costs a conversation rather than a retooling. An FTO that finds a problem at this stage is the cheapest possible outcome.
Before launch, updated. Between design and launch, applications will have published and pending claims will have issued.
Before a funding round or acquisition. Diligence will ask, and having the work done is materially better than commissioning it under time pressure while the deal is live.
Before entering a new market. A product cleared for the US is not cleared for Europe or China. New market, new search.
When a competitor's patent issues in your space. A targeted assessment of one patent is much cheaper than a full FTO and answers the question that has just arisen.
A worked example: what an FTO produces
A hardware product with four notable technical features, entering the US and German markets.
| Feature | Patents found | In force | Material after claim review |
|---|---|---|---|
| Sensor arrangement | 34 | 12 | 2 |
| Power management | 51 | 19 | 1 |
| Housing geometry | 18 | 6 | 0 |
| Communication protocol | 96 | 41 | 3 |
| Total | 199 | 78 | 6 |
199 patents surfaced, 78 still in force, 6 that actually need analysis.
The first cut is expiry. Over 60% of what the search returns is dead — either term-expired or lapsed for unpaid fees. That filtering is mechanical and fast.
The second cut is claim reading. Most in-force patents that look relevant from their abstracts have claims that do not read on the product. This is the slow part and the part that requires judgement.
Six patents get charted. Of those, suppose four are clearly not infringed on a straightforward reading, one is arguable, and one is a genuine problem.
The genuine problem, and the four options
| Option | Cost | Timeline | Certainty |
|---|---|---|---|
| Design around | Engineering time | Weeks, if caught early | High if the claim is narrow |
| Licence | Royalty or lump sum | Months to negotiate | High |
| Challenge validity via IPR | Six figures | ~18 months | Institution now ~37% |
| Proceed with a documented opinion | Opinion cost | Weeks | Risk accepted, willfulness mitigated |
Design-around is usually cheapest if the problem is found early, which is the entire argument for running FTO at design rather than at launch. The same finding after tooling is committed changes the ranking completely.
Note what the arguable patent costs. It does not block launch, but it requires a documented position, and that position is what the willfulness analysis will examine years later if the patent is ever asserted.
What FTO cannot tell you
Pending applications filed in the last eighteen months. Invisible, and potentially the most dangerous category, because a competitor who filed recently was working on the same problem recently.
How claims will change during prosecution. A published application with harmless claims may issue with different ones.
Whether a patent holder will assert. FTO assesses infringement risk, not litigation risk. A patent that reads on your product held by a company with no history of enforcement is a different commercial problem from the same patent held by an assertion entity.
Everything. A search is a sample of a very large corpus. Thoroughness reduces the miss rate; it does not eliminate it, and any opinion that claims otherwise should be treated with suspicion.
Before you commission an FTO
- Define the product precisely — features, markets, manufacturing locations. Vague scope produces vague results.
- List every jurisdiction where you will make, sell, offer or import.
- Run it early enough that a design-around is still cheap.
- Ask for claim charts on anything material, not just a list of patents.
- Have counsel produce the opinion if you want privilege and a willfulness defence.
- Verify status independently on any patent that drives a decision.
- Schedule a refresh before launch, because eighteen months of applications will have published since the search.
- Record the decision and the reasoning, because that record is the willfulness defence.