An examiner's first substantive act on a new application is to read the claims, search the prior art and write a rejection. That last part sounds like failure and is not: something close to nine applications in ten draw a rejection at first action, and most of those rejections are ordinary, expected and survivable. This is not about those. It is about the smaller category of applications that were already lost before the file was opened — killed by something done, or left undone, months or years earlier.
Those failures are unglamorous and extremely repetitive. Five of them account for most of the wreckage: a disclosure made before filing, a provisional that described far less than its owner believed it did, a date that passed without anyone noticing, a claim set drafted to sound impressive rather than to be infringed, and a contributor whose name was never written down. Not one of the five is a question of inventive merit. All five are procedural, which is another way of saying all five were avoidable.
§ 01 Disclosure
The Sentence That Ends Every Market but One
Under most patent laws an invention cannot be patented if, before the effective filing date, it was patented, described in a printed publication, in public use, on sale, or otherwise available to the public. That last clause is deliberately broad. A demonstration video, a conference poster, a crowdfunding page, a spec sheet emailed to a retail buyer without a confidentiality agreement, a booth at a trade fair — each is capable of putting the invention into the prior art against its own inventor.
Some jurisdictions soften this with a twelve-month grace period for disclosures originating with the inventor. Most of the world does not. Many other territories apply absolute novelty: anything made available to the public anywhere, in any language, by any means, before the filing date counts against the application, including the inventor's own talk given the previous afternoon. The practical consequence is stark. An inventor who publishes first and files eleven months later may still obtain a patent in a grace-period territory and has, in the same act, permanently forfeited every absolute-novelty territory, and with it most of the licensing value a manufacturer would have paid for. The shape of that grace period and its limits deserves close reading before any demonstration is scheduled.
Two further details catch people out. The first is that a sale need not be public to count: a confidential supply agreement, with the invention itself never described to anyone outside it, can still start the clock. The second is that the narrow experimental-use exception — testing in public, under the inventor's control, for the genuine purpose of perfecting the invention — is a doctrine applied retrospectively by a court, not a permission slip granted in advance. It is a defence, not a plan.
§ 02 The Placeholder
A Filing Date Is Only Worth What the Document Describes
The provisional is the most useful and most misused instrument available to an independent inventor. It is never examined, never published and never becomes a patent. It expires twelve months after filing and cannot be renewed. What it does is fix a date — and it fixes that date only for subject matter it actually describes, in enough detail that a person skilled in the field could build the thing from the description alone.
This is where the damage occurs. A three-page provisional written in a weekend, describing the concept and its benefits but not the mechanism, buys a date for almost nothing. Twelve months later the non-provisional is filed with proper claims, and every claim reaching beyond what the provisional described takes the later date instead. If the invention was demonstrated in the interim on the strength of the provisional, that demonstration is now prior art against the claims it failed to support. The document did not merely under-deliver; it created the exposure. The reasons thin provisionals fail their owners are worth understanding before the cover sheet is signed.
A provisional does not protect an idea. It dates a description. Whatever the description leaves out is not late — it is simply not there.
The distinction that costs the most
Written properly, a provisional should be indistinguishable from the specification that follows it: full structure, alternatives, ranges, materials, failure modes, and the variants considered and rejected. That is expensive and slow, and it is the point. Accounts of ordinary inventors who reached a shelf — the MixAid story among them — tend to skip past this stage, because there is nothing to photograph. It is still where the outcome is set.
§ 03 The Calendar
Dates That Do Not Negotiate
Patent practice runs on a small number of fixed intervals, and almost all of them are counted from a date the inventor chose rather than one anybody sends a reminder about. Twelve months from a provisional to the non-provisional filing, and the same twelve months to file abroad under the international priority convention. Eighteen months from the earliest priority date to automatic publication. Thirty months from that same date to enter the national phase in most countries under the international application system. Three months to respond to an office action, extendable to six by paying fees that roughly double each month. Three months to pay the issue fee after allowance — that one is not extendable at all. Then maintenance fees at three and a half, seven and a half and eleven and a half years, each with a six-month grace period and a surcharge, and a term that ends twenty years from the non-provisional filing date regardless.
Missing one of these is rarely dramatic. Nothing arrives to announce it. The application simply goes abandoned, and reviving it means a petition asserting the delay was unintentional, a substantial fee, and, if enough time has passed, an additional evidentiary showing about the whole period of delay. Most inventors who look for help with patenting invention ideas make the call after a deadline has already slipped, at the point where the remedies are procedural and expensive rather than free. The pattern is consistent enough that people who have spent decades around independent inventors — as in this longer conversation about how first-time filers are guided — describe the calendar, not the technology, as the thing that separates outcomes.
There is one further trap in the same family. An applicant who does not intend to file abroad may request that the application not be published at eighteen months. If they later change their mind and file abroad, the office must be told within forty-five days, and failing to do so abandons the application outright.
§ 04 The Claims
Written to Impress Rather Than to Be Infringed
The specification explains the invention. The claims define what is owned, and they are the only part of the document a competitor's lawyer will read carefully. A claim is a single sentence built from a preamble, a transitional phrase and a list of elements, and its scope is decided by the smallest details of that construction. "Comprising" leaves the list open, so a device with the listed elements plus others still infringes. "Consisting of" closes it, so adding one component escapes the claim entirely. That is the difference between an asset and a decoration.
The characteristic first-time mistake is to draft for admiration. Every advantage of the product is loaded into a single independent claim, on the assumption that more detail means more protection. The opposite is true: each additional element is one more thing a competitor may omit in order to avoid infringement. The mirror mistake is drafting so broadly that the claim reads on prior art nobody looked for. Between the two sits a further hazard — functional language describing what a component does rather than what it is, which is construed narrowly against the specific structure disclosed, and is held indefinite where no structure was disclosed at all.
Nor is any of this easily repaired later. Nothing new may be added to a specification after filing, so an amendment can only be assembled from words already on the page. And a narrowing amendment made to get around prior art surrenders the ground given up: the claim can no longer be stretched to cover close equivalents of what was abandoned. Ambition at the drafting stage is cheap; ambition after a first rejection is paid for in scope. It also helps explain the gap between what an inventor imagines owning and what the file secures — a gap that the way science fiction shapes what inventors set out to build widens rather than closes.
§ 05 The Names
The Co-Inventor Nobody Wrote Down
Inventorship is a legal determination, not a courtesy. It attaches to conception — the formation of a definite and permanent idea of the complete invention — and it is assessed claim by claim. A person who contributed to the conception of even one claim is a joint inventor. A person who funded the work, built a prototype to instruction, tested it, suggested a market or supplied ordinary skill is not, however much of the effort they carried.
The reason this matters is ownership. Absent a written agreement, each joint inventor owns an undivided interest in the whole patent, and each may make, use, sell and license it non-exclusively without the consent of the others and without accounting to them for the proceeds. An unrecorded co-inventor is therefore not an administrative loose end. They are a co-owner who can, entirely lawfully, license the invention to a competitor. Inventorship errors made without deceptive intent can be corrected, but the correction is usually demanded by the other side during a dispute, at the moment when the least leverage is available. Because claims change during prosecution, the list must also be revisited: cancel a claim and a named inventor may cease to be one. The rules on who belongs on the application and who does not are unforgiving in both directions.
Every one of these failures is silent. Nothing rejects you, nothing writes to you, and the loss is discovered at the exact moment the right becomes worth something.
The common property of all five
Read together the five have an obvious shape. Each involves a decision taken early, on incomplete information, by someone reasonably assuming it could be revisited. Each turns out to be structural. And each is cheap to get right at the time and expensive or impossible to correct afterwards: file before you demonstrate, describe before you date, diarise every interval the moment it starts running, draft claims against the competitor rather than the customer, and write down who conceived what while everyone still remembers. The invention is the hard part. Nothing on this list is.
End of sheet 05