The reason a well-run company will not read an unsolicited idea has nothing to do with contempt for outsiders. It is that an unsolicited disclosure creates a liability: if the firm already has similar work under way, or arrives at the same solution independently two years later, the person who wrote in has a plausible story about where it came from. Every submission gate, portal and one-sided agreement exists to close that exposure. Read them as risk management rather than insult and the whole process becomes navigable.
What follows is the mechanics — who to write to, what they are assessing, what the paperwork does, and where an inventor's leverage genuinely sits. None of it is adversarial. Most of it is simply unfamiliar.
Targets · The shortlist
The Right Company Is Rarely the Biggest One
The instinct is to start at the top of the category, and it is usually wrong. Market leaders run the most defensive submission processes, carry the largest internal development pipelines, and have the least need for outside ideas. The productive target is the second or third tier: firms with existing tooling, existing shelf space and existing buyer relationships in the exact category, but a thinner product roadmap and a real appetite for a line extension.
Fit is mechanical, not thematic. The question is whether the invention can be made on machinery the firm already owns, sold to buyers it already calls, and packed into cartons it already specifies. A device that requires a new moulding process, a new distribution channel or a new safety approval is a project rather than an addition, and projects need a champion with a budget. Building a shortlist of ten to twenty-five candidates and reading their actual catalogues — part counts, materials, price points, the gaps in the range — produces better targeting than any amount of enthusiasm.
Two more filters are worth applying early, and both matter more than scale when approaching manufacturers cold. Firms that already carry licensed products from outside inventors have a route in place and people who know how to use it. Firms whose entire range is own-design usually have a culture that treats external submissions as noise. Both facts are visible from a catalogue and a few minutes of searching.
The best licensee is the company for whom your invention is a line extension, not an initiative. Line extensions get approved by one person; initiatives need a committee.
How to read a shortlist
Gates · How submissions arrive
Portals, Agents and the Paperwork at the Door
Most substantial manufacturers now publish an innovation or supplier-submission route, and using it is almost always better than emailing a product manager directly. The portal exists precisely so that a disclosure can be received without creating the liability described above, which means it is also the only channel where a submission gets logged, routed and answered rather than deleted by a cautious assistant.
The price of entry is a submission agreement, and its standard terms deserve slow reading. Common provisions state that the company accepts no obligation of confidence, that it may already be working on something similar, that it is free to use anything not covered by a granted right, and that any compensation is at its sole discretion. Some go further and assert the right to use the material outright. None of that is negotiable at the portal stage, and refusing to sign simply ends the conversation. What it does mean is that the only protection an inventor carries into that channel is the filing they made beforehand — which is why the sequence matters more than the wording, and why it helps to be clear on what a patent actually is and what right it confers before signing anything.
A pending application is usually sufficient. It establishes a priority date, gives a file number to cite, and converts a conversation about an idea into a conversation about an asset with defined boundaries. It does not confer enforceable rights until grant, and honest inventors say so rather than implying more than the file supports. Overstating the position is the fastest way to lose credibility with an in-house counsel who will look the file up in a minute.
Material · What to send
One Page, Three Numbers, No Adjectives
The document that gets read is short. A single page carrying a photograph of the working object, one sentence naming the problem, one sentence naming the mechanism, the filing status with its number, an estimated landed unit cost, and a suggested retail price. That is the whole of it. Everything else — drawings, test data, the development history, the video of the mechanism under load — belongs in a second tier supplied on request.
What a licensing manager is silently scoring is narrow and consistent: does it work repeatably, can we make it inside our existing cost structure, will it return at a rate we can absorb, does it fit a shelf we already own, and is the intellectual property clean enough to defend a price. Nothing in a submission should be doing work other than answering those five questions. Superlatives actively subtract, because they signal an inventor who has not costed the thing. The most persuasive line in most submissions is a measured number: cycles survived, grams saved, seconds removed from an assembly step.
Conduct · The long middle
Following Up Without Becoming a Nuisance
Silence is the normal response and it usually means nothing. Submissions sit in queues reviewed on a quarterly cycle, and the person who received the file may be waiting on an engineering opinion, a category review or a budget that reopens in a new financial year. A single courteous follow-up at four to six weeks, adding one new fact rather than repeating the original message, is the correct cadence. New facts are things like a granted claim, a completed durability run, or a second manufacturer's interest.
Run the shortlist in parallel rather than in series, and say so when asked. Sequential approaches to twenty firms at six weeks each consumes years of a term that is already running. Parallel approaches also change the tone of any eventual negotiation, because a licensee that knows it is one of several reads the term sheet differently. When the decline arrives, ask which of the five criteria failed — cost, fit, function, rights or liability. Answered honestly, that single question is worth more than the submission was, because it tells the inventor what version three has to fix. Profiles of the people who assess these files, including this long-form profile of how inventor submissions get evaluated, tend to describe the same unglamorous filtering: most declines are arithmetic, not judgement of the idea.
Approaching manufacturers well is largely a matter of removing reasons to say no before anyone has to say it. File first, cost the object honestly, target the firm whose factory already makes something adjacent, send one page, and treat the submission agreement as the price of a hearing rather than a trap. The broader arc that this stage sits inside — record, filing, build, approach, licence — is set out in the full commercial sequence from sketch to signed agreement, and the approach is the phase where sequence errors made earlier become expensive. It is also worth remembering how ordinary the source material usually is: accounts of how speculative fiction seeds real product ideas describe inventions that began as domestic irritations and ended as catalogue items, which is the shape almost every licensed product takes.
A decline that names the failed criterion is a specification for version three. A decline with no reason is the only genuinely wasted submission.
The question worth asking every time
File first, then knock