Patentability search services answer one question: has this invention already been described?
It is the cheapest question to answer and the one that most inventions fail. Novelty and obviousness defeat far more applications than subject-matter eligibility ever does, and both turn entirely on what already exists.
It is also the question people most often skip, because filing feels like progress and searching feels like delay.
What the search is looking for
| Ground | Test | What defeats you |
|---|---|---|
| §102 novelty | Anticipation | One reference disclosing every element as claimed |
| §103 obviousness | Would a skilled person have made the change? | Several references combinable with a reason |
Anticipation is binary and easy to assess. If a single document contains every element, the claim is dead as written.
Obviousness is a judgement and it is where searches become arguments. Two references that together suggest the invention may or may not render it obvious, depending on whether there was a reason to combine them and a reasonable expectation of success.
Prior art is broader than most inventors expect. Patents, published applications, journal articles, conference papers, product manuals, public demonstrations, sales — anywhere in the world, in any language, whether or not anyone read it. Including your own disclosures outside the twelve-month US grace period. See prior art.
Patentability is not freedom to operate
The most expensive confusion in this area.
| Patentability search | Freedom-to-operate search | |
|---|---|---|
| Asks | Can I patent this? | Can I sell this? |
| Covers | All prior art | In-force claims only |
| Expired patents | Relevant | Irrelevant |
| Can stop when | One good reference is found | Never — must cover everything |
| Typical cost | $500–$3,000 | $3,000–$15,000+ |
| Output | References and relevance | Claim-by-claim clearance |
You can pass one and fail the other. A patentable improvement on someone else's in-force patent is patentable and unbuildable at the same time — which is exactly the situation cross-licences exist to resolve.
Ordering the wrong one wastes the budget. A patentability search before a product launch tells you nothing about infringement risk. See freedom to operate.
What you are buying beyond the references
| Component | Value |
|---|---|
| The references | The obvious part |
| Coverage you cannot reach | Non-patent literature, foreign languages |
| Element mapping | Turns a list into an answer |
| A stated scope | Makes the result defensible |
| A dated record | Supports disclosure and opinions |
Coverage is the real purchase. Anyone can search Google Patents; few can search conference proceedings and Japanese-language art properly.
What a good report contains
| Section | What to expect |
|---|---|
| Scope statement | What was searched, and what was not |
| Databases and date ranges | Coverage |
| Classifications searched | CPC subgroups covered |
| Search terms and strategies | Reproducibility |
| References found | With bibliographic detail |
| Relevance assessment | How each maps to the invention's elements |
| Closest art identified | The references that matter most |
| Limitations | The eighteen-month blackout, at minimum |
A report that states its limitations is more useful than one that does not. Every search has gaps — unpublished applications, unindexed literature, languages not covered — and a searcher who names them has thought about them.
Element mapping is what separates a report from a list. Ten references with no analysis is a starting point; ten references mapped against your intended claim elements is an answer.
The search is not an opinion. A searcher finds references; an attorney assesses what they mean for patentability. Some engagements include both, at a higher price.
Briefing a searcher well
| Supply | Effect |
|---|---|
| Functional description, in detail | Better matching than a product name |
| The elements you expect to claim | Focuses the search on what matters |
| Field terminology and synonyms | Crosses the vocabulary gap |
| Known competitors and products | Starting points |
| Prior art you already found | Their hours go to what you could not reach |
| Jurisdictions that matter | Scope |
| What you have publicly disclosed | Affects what is art against you |
Sharing your own search results is the highest-value input. A searcher starting from zero repeats work you have already done; one starting from your results spends the whole budget on new ground.
Describe what it does, not what you call it. "Flow stabiliser" is your name for it; the prior art may call it a damping baffle, and the searcher needs both.
Worked example: a search and its consequences
A portable water filtration improvement. Free search first, then professional.
The free search
| Time | 3 hours |
| Tools | Google Patents, Espacenet, product listings |
| Documents reviewed | ~50 |
| Close references found | 2 |
| Conclusion | Promising, but the field is crowded |
The professional search
| Cost | $1,900 |
| Turnaround | 12 days |
| Scope | US, EP, JP, CN; patents, applications and non-patent literature |
| Documents in report | 23 |
| Closest reference | A 2013 conference paper, not a patent |
What the conference paper did
| Claim element as intended | Disclosed in the paper? |
|---|---|
| Multi-stage cartridge | Yes |
| Flow-rate compensation | Yes |
| Backflush geometry | No |
| Cartridge indexing mechanism | No |
Two of four elements anticipated by a non-patent document that no free database surfaced. The free search had found neither.
The invention was not dead — it was reframed. Claims directed at the backflush geometry and indexing mechanism survived, and those were the genuinely novel parts.
The $1,900 bought a narrower, defensible application instead of a broad one that would have met the paper during prosecution, or worse, after grant.
Search firms versus law firms
| Specialist search firm | Law firm | |
|---|---|---|
| Cost for the search | Lower | Higher |
| Search expertise | Their whole business | Varies |
| Legal opinion included | No | Usually available |
| Foreign-language capability | Often strong | Varies |
| Turnaround | Usually faster | Varies |
| Privilege | Not automatic | Attorney work product |
The privilege difference is worth understanding. A search commissioned directly from a search firm may not attract the same protection as one commissioned through counsel, which matters if the results are unfavourable.
Many practices commission through counsel for that reason, accepting the markup in exchange for the protection.
For an individual inventor pre-filing, direct is usually fine. The scenario where privilege matters most is a search that finds art threatening a patent you already hold.
Reading results without over-reacting
| What you found | What it means |
|---|---|
| One reference with every element | Anticipated as claimed — reframe or stop |
| Several that combine obviously | Difficult under §103 |
| Close but distinguishable art | Narrower claims likely available |
| Crowded field, nothing exact | Normal; claim precisely |
| Nothing found | Encouraging, not conclusive |
Close prior art is the usual result and not a failure. Almost every invention sits near something, and the useful output is knowing exactly where the boundary is before you draft claims.
The reformulation frequently comes out of the references. Reading what was tried and where it fell short is how the real unsolved problem becomes visible. See invention ideas.
Nothing found is the result to treat most carefully. It may mean the field is open. It may mean the search missed the vocabulary, or the reference is in an unpublished application.
Timing the search in the filing sequence
| Point | Search appropriate? |
|---|---|
| Idea stage | Free search — yes, always |
| Before a provisional | Free search minimum |
| Between provisional and non-provisional | Best time for the professional search |
| At non-provisional filing | Late but useful |
| After first office action | Reactive; the examiner has already searched |
| Before renewal | Different question — who practises the claims |
The twelve months between provisional and non-provisional is the ideal window. Priority is secured, the real money has not been committed, and the results can still reshape the claims before drafting.
Searching after filing is largely wasted for patentability, because the claims are fixed and the examiner is running their own search. What it does support is deciding whether to keep prosecuting.
What no search can cover
| Gap | Why |
|---|---|
| Unpublished applications | 18-month blackout — absolute |
| Non-publication request filings | Invisible until grant |
| Unindexed trade literature | Coverage is uneven everywhere |
| Products never documented | Still prior art |
| The examiner's own search | They may find what nobody else did |
The blackout is the limit that matters. Someone may have filed on the same invention last year, and nothing will show it until they publish.
Which is why a search reduces risk rather than removing it, and why any service promising certainty is overstating what is possible.
What the report cannot tell you
| Question | In scope? |
|---|---|
| Does prior art exist? | Yes |
| Is this patentable? | An assessment, not a determination |
| Will the examiner allow it? | No |
| Can I sell the product? | No — that is freedom to operate |
| Is an existing patent invalid? | No — that is a validity search |
| How should I draft the claims? | Attorney work |
A search reduces uncertainty rather than resolving it. The examiner runs their own search, may find different references, and applies obviousness as a judgement.
The distinction between a search and an opinion matters when comparing quotes, because a lower price frequently means references without analysis.
Turnaround expectations
| Stage | Typical |
|---|---|
| Brief agreed | Same week |
| Search conducted | 1–3 weeks |
| Report delivered | With the search |
| Follow-up questions | A few days |
| Rush service | Available at a premium |
Build the turnaround into the filing plan. A search commissioned two weeks before a provisional expires cannot change the claims meaningfully, which wastes most of its value.
Patentability search services: the checklist
- Run the free search first. It eliminates most inventions and costs only time.
- Search products and market listings too, since a product on sale is prior art whether or not it was patented.
- Commission the professional search before the non-provisional, where the real money starts.
- Do not confuse it with freedom to operate. Different question, different scope, different price.
- Specify jurisdictions and confirm non-patent literature is included.
- Brief with a functional description and the elements you expect to claim.
- Share what you have already found, so the searcher's hours go to new ground.
- Expect close art. Almost every invention sits near something, and knowing where is the point.
- Read the limitations section, and treat a report with none as incomplete.
- Treat "nothing found" as encouraging, not conclusive. The eighteen-month blackout means no search is ever complete.