"Can I patent this?" is really four separate questions wearing one coat. US law puts four independent hurdles in front of every application, and an invention has to clear all of them. Failing any one is fatal, and they fail for completely different reasons: one is about the kind of thing you invented, one is about whether it is new, one is about whether it is a big enough step, and one is about whether you described it well enough. An hour spent understanding the four before you call an attorney will save you far more than an hour of that attorney's time.
One thing to get out of the way first: patentability has nothing to do with whether the invention is good, marketable, or profitable. Those are business questions, and they are covered separately in how to tell if your invention idea is actually good. Plenty of patentable inventions are commercially worthless, and plenty of extremely profitable products are unpatentable.
Hurdle 1: Eligible subject matter (Section 101)
The statute says a patent can be obtained for a process, a machine, an article of manufacture, or a composition of matter. That is broad, and for physical products it is almost never a problem. A mechanism, a circuit, a device, a material formulation, a manufacturing process — all comfortably inside.
The trouble is the judge-made exceptions: laws of nature, natural phenomena, and abstract ideas cannot be patented. The abstract-idea exception is what bites software and business-method inventions, following the Supreme Court's decision in the Alice case. Examiners apply a two-step analysis. Step one asks whether the claim is directed to an abstract idea — a mathematical concept, a method of organizing human activity, or a mental process. If it is, step two asks whether the claim adds something "significantly more" than the abstract idea implemented on a generic computer.
In practice this means a claim that recites collecting data, applying a known analysis to it, and displaying the result is in serious trouble, no matter how commercially clever it is. A claim that improves how a computer or a machine actually functions — reduced memory footprint, a new sensor fusion approach that makes a physical measurement possible, a control loop that changes what the hardware does — has a real path. Our guides to what software still qualifies and business method patents after Alice go through where the line currently sits.
Section 101 also carries a utility requirement, but the bar is minimal. The invention has to do something. Only perpetual motion machines and similar impossibilities fail it.
Hurdle 2: Novelty (Section 102)
Your invention must be new. Concretely: no single piece of prior art may describe every element of your claim. If one earlier reference — a patent, a published application anywhere in the world, a journal article, a product manual, a YouTube video, a trade show display, a sales listing — contains everything in your claim, the claim is anticipated and dead.
Two features of the US rule catch inventors regularly:
- The prior art universe is global and includes non-patent material. A crowdfunding page from another country counts. So does a forum post with a photo.
- Your own disclosures count against you. The US gives inventors a one-year grace period from their own public disclosure, sale, or offer for sale — but most of the rest of the world gives none at all. Demonstrating at a trade show or taking pre-orders before filing can preserve your US rights while destroying your European and Chinese ones. The specifics are in public disclosure and the one-year grace period.
Novelty is the hurdle you can genuinely assess yourself, at least roughly, by doing a proper prior art search. Note that this is a different exercise from checking whether you are free to sell — a distinction laid out in patentability search vs freedom to operate.
Hurdle 3: Non-obviousness (Section 103)
This is where most applications actually die. Even if no single reference shows your invention, you cannot have a patent if the difference between your invention and the prior art would have been obvious to a person of ordinary skill in the field at the time you filed. Examiners routinely combine two or three references and argue that a skilled engineer would have put them together.
Courts assess this through a framework known as the Graham factors:
- The scope and content of the prior art.
- The differences between the prior art and the claims.
- The level of ordinary skill in the field.
- Objective evidence of non-obviousness — the secondary considerations.
That fourth factor is the one inventors can influence and usually ignore. It includes commercial success attributable to the claimed feature, long-felt but unsolved need, failure of others who tried, skepticism from experts, unexpected results, and copying by competitors. If you have test data showing your approach performs surprisingly better than the obvious combination, that is evidence, and it should be in the specification. If three companies tried and abandoned this approach, document it.
The Supreme Court's KSR decision made obviousness rejections easier for examiners by rejecting rigid formulas in favor of common sense, so expect one. It is a normal opening position, not the end — see why the examiner rejects your claims for obviousness.
Hurdle 4: Description and enablement (Section 112)
The last hurdle is about your document, not your invention. Section 112 requires three things. The specification must enable a skilled person to make and use the invention without undue experimentation. It must contain a written description showing you actually possessed the full breadth of what you claim. And the claims must be definite — a reader must be able to tell what falls inside and what does not.
Practical failures here are mundane and avoidable. Claiming a broad range while describing one working example. Using relative terms like "substantially rigid" or "about optimal" with nothing in the specification to anchor them. Reciting a function without disclosing any structure that performs it. Leaving out the parameters that actually make the thing work because you would rather keep them secret — that is a direct enablement failure, and it is the one place where the patent bargain is unforgiving: full disclosure in exchange for the right.
Running the assessment yourself
| Hurdle | The question | What to do about it |
|---|---|---|
| 101 | Is this a machine, process, article, or composition — and not just an abstract idea on a computer? | Re-frame around a technical improvement, not a business outcome |
| 102 | Does one reference show every element? | Search by classification and by synonym; document your filing date discipline |
| 103 | Would combining two known things have been obvious? | Gather test data and secondary-consideration evidence early |
| 112 | Could a competent engineer build it from your text alone? | Describe alternates, ranges, and the parameters you would rather hide |
Work through those four honestly and you will already know whether the answer is likely yes, likely no, or genuinely uncertain — and "genuinely uncertain" is the case where paying a professional is clearly worth it. Then run the practical steps in the pre-filing checklist before anything gets submitted.
Projects House is a product development engineering firm, not a law firm. Where we help is the engineering half of a patentability assessment: establishing what is genuinely different about your approach, generating the test data that supports an unexpected-results argument, and producing a technical disclosure complete enough to satisfy the enablement requirement. The legal opinion and the filing itself belong with a registered patent attorney. To get the engineering groundwork done properly, reach us through the contact form.