Doing It Yourself Against Getting Help With an Invention

What Each Route Actually Costs

Official fees are the smallest number in the exercise and the one most people fixate on. A provisional or equivalent placeholder filing typically costs an individual applicant somewhere in the low hundreds. A full utility application attracts a basic filing fee, a search fee and an examination fee, which together commonly land between roughly 800 and 2,000 in most major territories before any reduction. Many granting offices discount heavily for small or sole applicants — reductions of 60 to 80 per cent are common where the applicant qualifies — which can bring the official cost of getting an application on file and examined under 500.

Professional drafting sits an order of magnitude above that. A mechanically simple invention with a modest claim set is generally quoted between 5,000 and 8,000. Electronics, software-implemented methods and anything requiring a formal algorithm description run from 8,000 to 15,000 and upwards. Formal drawings add 100 to 400 per sheet. Prosecution — the argument with the examiner after the first rejection — is billed separately and typically consumes a further 2,000 to 6,000 across two or three exchanges.

The hidden column is time. An inventor drafting their own first application should budget 80 to 150 hours: reading the office's own guidance, learning claim syntax, running the prior art search, producing compliant drawings, and re-reading the whole thing against each statutory requirement in turn. At any plausible valuation of an inventor's own hours, the saving is far smaller than the fee difference suggests.

• Official fees: hundreds, and reducible for individual applicants in most territories.
• Drafting: thousands, scaling with technical complexity rather than with commercial value.
• Prosecution: unpredictable, and incurred on both routes once examination begins.
• Renewals or maintenance payments: due periodically for the whole twenty-year term, identical either way.

The Claim Set Is the Thing Being Bought

A patent application is usually eighty per cent description and twenty per cent claims, and the twenty per cent is the entire asset. The description teaches; the claims define what a competitor may not do. Everything a drafter is paid for concentrates in that final section, and the difference between competent and poor claiming is not stylistic — it is the difference between a monopoly on a mechanism and a monopoly on one commercial embodiment of it.

Consider the standard error. An inventor describes their product accurately and claims it accurately: the housing, the specific fastening, the particular polymer, the dimensions that worked. The claim is granted precisely because it is narrow. A competitor then reads it, substitutes a different fastening, and sells the same function without infringing anything. The patent is valid, enforceable, and commercially worthless.

A drafter's habit is the reverse: begin with the broadest statement of the inventive principle the prior art will tolerate, then add dependent claims of progressively narrowing scope beneath it, so that if the broad claim falls during examination or in litigation the narrower ones survive independently. Most offices allow three independent claims and twenty total before excess fees begin, which is the practical budget within which that ladder has to be built.

There is a second asymmetry. Claim language is a technical dialect with rules that are nowhere obvious: every element needs antecedent basis, relative terms invite indefiniteness objections, and functional phrasing can silently limit a claim to the structures disclosed in the specification and their equivalents. None of this is intuitive, and the cost of getting it wrong is usually discovered years later. The errors examiners see most often are almost all of this kind.

The Mistakes That Cannot Be Repaired

Most defects in an application are fixable. Formatting, drawing conventions, missing reference numerals, clumsy claim wording, even an entirely misconceived claim strategy can generally be amended during prosecution. Two categories cannot.

The first is new matter. Once an application is filed, nothing may be added to its disclosure. If the specification omits a feature — a step in the method, a variant of the mechanism, a range of values — that feature can never be claimed from that filing. It is not that adding it is expensive; it is that adding it is prohibited. The only remedy is a fresh application with a later date, in a system where dates decide priority. This is the single most consequential difference between a thorough specification and an adequate one, and it is invisible at the moment of filing.

The second is disclosure before filing. Describing an invention publicly can extinguish the right to patent it. Some territories grant the inventor a twelve-month grace period for their own disclosures; many operate absolute novelty, under which the first public description becomes prior art against its own author immediately. A trade stand, a funding page, a detailed post explaining how the thing works — each can be enough, and no fee reverses it.

Both traps are procedural rather than technical, which is why they catch capable engineers so reliably. The people who spend four decades working alongside independent inventors tend to report the same pattern: the irreversible mistakes are cheap to make, quick to make, and surface eighteen months later.

When Filing Alone Is the Rational Choice

Self-filing is not a failure state. There are configurations in which it is straightforwardly the better decision, and they share identifiable features.

The mechanism is simple and mechanical, describable in ten pages and a handful of drawings, with no software, no chemistry and no process parameters. The commercial ceiling is modest and known — a niche tool with a small addressable market, where a 9,000 drafting bill could not be recovered from licensing revenue even on optimistic assumptions. The inventor is technically literate, comfortable with statutory language, and genuinely willing to spend a hundred hours on it. Or the filing is a placeholder: a first application intended to secure a date while the design is still moving, with the full specification to be prepared professionally within the priority year.

That last case deserves emphasis because it is the most defensible use of a do-it-yourself filing — provided the placeholder is written as though it were the full document. A later application can claim the benefit of an earlier one only to the extent the earlier one actually described the invention. A three-page placeholder supports three pages' worth of eventual claims. What preparing a first application without representation demands is not legal talent so much as an unusual tolerance for procedural detail.

There is also a category the fee comparison misses entirely: inventions whose final form takes years to settle. Where the physical geometry is inseparable from the function, filing early and narrowly can be worse than filing late and completely — an argument that surfaces repeatedly in the case for inventor-led design in the built environment, where solutions are refined across whole projects rather than prototypes.

When Handing It Over Is the Only Sensible Answer

The mirror image is equally identifiable. Where the invention is a method with parameters, a software-implemented process, a formulation, or anything at the edge of what is considered patentable subject matter, the drafting problem stops being descriptive and becomes strategic. Eligibility objections are argued in a vocabulary that takes years to acquire, and the specification has to be built in anticipation of them.

Where money has already been committed — tooling ordered, a manufacturer engaged, a distribution conversation underway — the asymmetry becomes absurd. Protecting a 200,000 commitment with a document drafted over a weekend is not thrift. And where the invention will be licensed rather than manufactured, the claim set is the negotiating instrument. A licensee's counsel reads the claims first and the marketing second; a weak independent claim reduces a royalty rate before the first meeting ends.

Studies of granting office data have found repeatedly that unrepresented applicants obtain grants at markedly lower rates than represented ones, with much of the gap attributable to abandonment after a first rejection rather than to any substantive defect in the underlying invention. The application was not hopeless; nobody knew how to answer the objection, the three-month response window closed, and the case went abandoned. That is a procedural loss, and it is the most common way a good invention dies.

Between the two extremes sits the arrangement most independent inventors actually use: do the searching, the record-keeping and the specification drafting themselves, then buy professional review of the claims before filing. Structured help with patenting invention ideas is most valuable at exactly that seam, because the claim set is the part that cannot be improved after grant. Where that division of labour pays for itself, and where it does not, is set out in detail in the discussion of what a drafter's fee actually buys.

What no route offers is a way to avoid the decision. Inventions are described publicly, or they are not; applications are filed with usable claims, or they are not. The routine treatment of invention as an act of sudden inspiration — a habit visible in the wider cultural fascination with the moment of discovery — obscures how much of the outcome is decided by documentation choices made in the weeks after the idea arrives, by whoever is holding the pen.