Novelty and Obviousness: The Two Tests That Decide Everything
Every other requirement can be repaired by better drafting. These two are decided by facts that existed before the application was written, and they end more applications than all remaining objections combined.
An examiner reading an application for the first time is not assessing whether the invention is clever, useful or well made. The examination is a comparison against documents, and it proceeds through two distinct questions that are frequently confused with one another. The first is whether the invention is new. The second, much harder, is whether it involved an inventive step — whether a competent person already working in the field would have arrived at it without needing to invent anything.
The order matters because the tests use different evidence. The first permits only one document at a time. The second permits several, combined, provided there was a reason to combine them. That single procedural difference explains why an application can survive a search that turns up dozens of close references and still be refused on the strength of two of them read together.
Understanding novelty and obviousness as separate machinery, rather than as a single vague standard of newness, changes how an inventor reads search results. It also changes what is worth writing down at the drafting stage, because the evidence that answers the second test has to be present in the application from the day it is filed. It cannot be manufactured later.
How the two tests differ
- Anticipation
- One document, standing alone, must disclose every feature of the claim. If a single element is missing, the claim is novel — however close the rest of the disclosure comes.
- Inventive step
- Several documents may be combined, and ordinary skill may be added, but only where something gave a reason to combine them in that way.
- The skilled person
- A legal construct: competent, aware of everything published in the field, and possessing no imagination whatsoever.
Test One · Anticipation
Novelty Is a Narrow, Mechanical Comparison
Novelty is the more forgiving of the two, because its rule is strict in the applicant's favour. To destroy novelty, a single item of prior art must contain every feature recited in the claim, arranged as the claim arranges them. An examiner cannot assemble the invention from two sources and call it anticipated. If the claim recites five elements and the closest document discloses four, the claim survives this test outright.
What makes the comparison unforgiving in practice is the breadth of what counts as a document. The relevant body includes granted rights, applications published but never granted, expired rights whose teaching has long since entered the public domain, conference papers, product manuals, catalogue listings, dissertations and the inventor's own earlier disclosures. Publication in any language counts. A drawing counts even where the accompanying text says nothing about the feature, provided the feature is clearly shown. Rights that lapsed decades ago are as effective against a claim as a competitor's filing from last month, and they are the category first-time searchers skip most consistently.
Two further mechanics catch inventors out. The first is inherency: a document can anticipate a feature it never mentions, if the feature is a necessary consequence of what it does describe. A prior process that unavoidably produces a particular crystal structure anticipates a claim to that structure even though the earlier author never analysed it. The second is the effect of dates. In most systems the comparison is made against everything public before the earliest filing date being relied on — which is why the twelve-month priority window between a first filing and a fuller one is valuable, and why a competitor's application filed one day earlier but published eighteen months later can still be cited against you.
The practical response is to read search results element by element rather than impressionistically. Write the claim as a numbered list of features, then set each cited document against it and mark which features it actually discloses. Applications are frequently abandoned on the basis of a reference that, examined this way, is missing two of the six things that matter. Long-running advocacy work such as four decades of support for independent inventors keeps returning to that pattern: the reference that looked fatal in the abstract was, on a feature-by-feature reading, nothing of the kind.
Novelty asks whether one document already says it. Inventive step asks whether anyone needed to be told. The second question is where applications are lost.
The distinction that matters
Test Two · Inventive Step
Obviousness Is an Argument, Not a Measurement
The second test asks whether the difference between the invention and the prior art is one the skilled person would have made as a matter of routine. Because it turns on what a hypothetical person would have thought, it is the only major requirement decided by argument rather than by comparison, and it accounts for the majority of substantive refusals.
Most granting offices structure the reasoning the same way, whatever the local terminology. The examiner identifies the closest single piece of prior art, articulates the objective difference between it and the claim, frames the technical problem that difference solves, and then asks whether the skilled person, faced with that problem, would have reached the claimed solution using the other documents available. Each step is contestable. Applicants win at the first step by showing the chosen starting point is not the closest, at the third by reframing the problem the invention actually addresses, and at the fourth by showing the documents offered no reason to combine.
The recurring defect in these objections is hindsight. Once the solution is known, the route to it looks inevitable, and an examiner reading the application already knows the answer. The counter-argument is to insist on contemporaneous motivation: not whether the pieces could have been combined, but what — before the invention existed — would have prompted anyone to combine them. Where two references are drawn from unrelated fields, or where one expressly recommends against the feature the other supplies, that motivation is absent, and the objection weakens considerably.
Evidence
The Facts That Answer an Obviousness Objection
Because the second test is argumentative, it is answered with evidence rather than assertion, and several categories carry recognised weight. Unexpected results are the strongest: a measurable performance difference the prior art would not have predicted — a threefold increase in cycle life, a failure rate falling from one in twenty to one in a thousand, a reduction in assembly steps from nine to three. The figure has to be in the application as filed, which is the single most common drafting omission. Data produced two years later, after an objection arrives, is often inadmissible for this purpose in the territories that matter most.
Teaching away is the second category. Where a prior document explicitly warns that the approach the invention takes will not work, the skilled person had a reason not to take it, and an examiner asserting the opposite is arguing against the record. Long-felt need is the third: evidence that the problem was recognised and unsolved for a sustained period, while people with resources and motive were working on it, is difficult to reconcile with a claim that the solution was routine. Commercial success and independent copying by competitors carry weight too, though secondary to the technical evidence and only where the success can be tied to the claimed feature rather than to marketing.
Claim structure supplies the last line of defence. A well-drafted application does not stake everything on one broad claim; it descends through dependent claims that add features progressively, so that when the broadest formulation falls to a combination of references, a narrower one remains standing without new matter having to be introduced. That structure is decided at drafting and cannot be retrofitted, which is why the honest verdict on whether the invention justifies the filing cost is best reached before the drafting rather than after it.
None of this is exotic. Public discussion of invention tends to dwell on the flash of insight — a tendency visible in the way popular storytelling frames the moment of discovery — while examination rewards the far duller work of recording what was measured, when, and against what baseline.
The data that defeats an obviousness objection has to be in the application on the day it is filed. Later is usually too late.
On drafting order
In Practice
What This Changes About How You Prepare
Read the two tests backwards and they produce a short list of instructions. Search before drafting, because both tests are decided against documents that already exist and no drafting choice alters them. Record measurements as the design develops — before and after figures, failure modes, cycle counts — because those numbers are the only currency the second test accepts. Note which documents you considered and rejected, and why, since a contemporaneous account of why an approach looked unpromising is precisely the material that rebuts a hindsight argument later.
Then keep the file dates clean. File before disclosing, since an inventor's own demonstration or listing enters the prior art in many territories immediately and in others after a grace period of roughly twelve months. Where a first filing is used to secure a date, treat the twelve-month window as a working deadline for gathering the data the fuller application will need, not as a year of grace. Publication at around eighteen months from the earliest date is automatic in most systems, so the invention becomes public whether or not a right is ever granted.
The tests are also worth applying to yourself before an examiner does. Take the closest reference you found, state the difference between it and your claim in one sentence, and ask what a competent practitioner in the field would have done with that difference. If the honest answer is that they would have done the same thing, the useful move is to narrow toward the feature that genuinely surprised you, not to argue harder. The broader account of where projects stall, set out in the main report on invention attrition, treats that self-examination as the checkpoint most often skipped.
What remains after both tests have been applied honestly is a smaller invention than most inventors begin with, and a more defensible one. The narrowing is not a loss. It is the process by which a general idea becomes a specific piece of property — and specificity, as accounts like this record of one household device reaching production illustrate, is what survives contact with both an examiner and a manufacturer. Applied together, novelty and obviousness are less a pair of obstacles than a specification for the evidence an invention needs to carry.
The first test asks what already exists. The second asks what anyone would have done about it. Answer both on paper, early, and examination becomes procedure rather than verdict.
End of report