The Drafting Errors Examiners See Most Often

Words That Do Not Draw a Boundary

A claim must let a person skilled in the field determine, with reasonable certainty, whether a given product falls inside it. Language that cannot do that is indefinite, and indefiniteness is among the most frequently raised objections against unrepresented filings.

The usual culprits are comparatives without a reference point: "substantially rigid", "approximately vertical", "a thin membrane", "sufficient pressure", "an optimal ratio". Each is perfectly meaningful in a specification, where its job is to teach, and each is defective in a claim, where its job is to fence. Relative terms are not banned outright — many are accepted where the specification supplies a standard for measuring them — but a term used in a claim and left undefined anywhere else is a reliable objection.

The second variety is the missing antecedent. Claims use the definite article to refer back to elements already recited, so "the mounting bracket" is only permissible if "a mounting bracket" appeared earlier in that claim or in the claim it depends from. Referring to an element that was introduced in the description but never recited in the claim breaks the chain, and the objection follows automatically.

Both faults are fully curable by amendment, which is why they belong in the nuisance category: they cost a response cycle of three to six months and no scope. They are still worth eliminating before filing, because every avoidable objection in a first office action is examiner attention diverted from the arguments that actually matter.

Claiming More Than the Description Supports

The second family of faults runs in the opposite direction from timidity: a claim written broadly enough to cover territory the specification never taught. If a claim recites "a biodegradable polymer" while the description discloses only one specific polymer and gives no guidance on which others would work, the claim is unsupported. It claims a genus and enables a species.

Offices raise this as a written description or sufficiency objection, and it is a genuinely dangerous one because the remedy is narrowing. The applicant cannot cure it by adding the missing teaching — new matter is prohibited after filing — so the only route forward is amending the claim down to what the original description actually supported. Breadth written without support is therefore not merely useless; it consumes a response cycle and ends with a narrower claim than a careful drafter would have written on day one.

The competent version of breadth is structural. State the inventive principle at its highest defensible level of generality in the independent claim, then build dependent claims that narrow it in identifiable steps — by material, by geometry, by parameter range, by sequence. If the broad claim falls to prior art, the dependants survive it and prosecution continues from a lower rung rather than from nothing.

Flat Claim Sets and Broken Dependency

Most first-time claim sets are flat: fifteen independent claims that restate the same invention in fifteen slightly different ways. This costs money — most offices charge excess fees beyond three independent claims and twenty claims in total — and buys nothing, because all fifteen fall to the same citation.

A properly built set is a ladder. One or two broad independent claims, then eight to fifteen dependent claims descending in scope, each adding a single distinguishing feature. Different statutory categories are handled by separate independent claims — the apparatus, the method of using it, the method of manufacturing it — because a competitor who sells only components may infringe the method claim and not the apparatus claim.

Dependency itself is a frequent mechanical fault. A dependent claim must properly incorporate everything in the claim it references; claims that depend on a later-numbered claim, or that contradict a feature of their parent, draw objections. Improper multiple dependency is charged for or prohibited outright in many territories. None of this is conceptually hard, and all of it is easy to get wrong at eleven at night on a filing deadline. The broader distinction between what the different application types are for is set out plainly in the official summary of the application types and proceedings available to applicants.

Claiming a Result Instead of a Mechanism

A claim to "a device that keeps beverages at a constant temperature" claims an outcome. Outcomes are not patentable; the specific arrangement that produces them is. Written that way, a claim is either rejected over any prior art achieving the same result by any means, or read down to cover only the structures actually disclosed and their close equivalents.

That second consequence is the one that surprises people. Purely functional phrasing — an element defined by what it does rather than what it is — is construed in many territories as limited to the corresponding structures in the specification. A claim that reads as sweepingly general therefore turns out, at the moment of enforcement, to cover only the two embodiments that were drawn. Deliberate use of that construction is a legitimate strategy; accidental use of it is one of the most expensive drafting errors an applicant can make, because nothing in the office correspondence flags it.

The related trap is the single-embodiment specification: one product, described precisely, with no variants. It sails through examination and grants quickly, because narrow claims are easy to distinguish from prior art. Then a competitor changes the material, moves the pivot, or reorders two steps, and infringes nothing. This is why comparing self-preparation against a professional fee on grant probability alone gives the wrong answer: the fastest route to a granted patent and the route to a valuable one are not the same route.

Faults That Predate the Claims

Two further errors sit beneath the claim set and cannot be drafted around at all.

The first is misidentified subject matter. Utility protection covers how a thing works; separate design or industrial design registration covers how it looks. An inventor whose contribution is genuinely an appearance sometimes spends thousands pursuing utility claims that no examiner will allow, when a design registration — cheaper, faster, and entirely appropriate — was the right instrument from the start. The plain definition of what the utility right actually covers is worth reading before drafting begins; the official explanation of what a patent is and what it grants is a reasonable starting point.

The second is inventorship. Every person who contributed to the conception of the claimed invention must be named, and nobody else may be. Adding a financial backer who conceived nothing, or omitting a collaborator who solved the key problem over a weekend, is a defect that can invalidate a granted patent years afterwards. Correction is possible where the error was innocent, but it requires a petition and evidence, and the question tends to surface during due diligence, at the least convenient moment imaginable.

Both faults share a quality with every fault listed above: they are found late, by someone other than the applicant, at a point when the cheap remedy has expired. Products that reach the market with defensible protection — the wearable vibration device described in the coverage of one pain-relief invention is a workable example — generally got the foundations settled before anyone wrote a claim.