The Six Searches — And Which One You Actually Need

The most common and most expensive error is treating a patentability search as an FTO search. They use different criteria, different scopes and produce different conclusions:

---

What Counts as Prior Art in India

India applies absolute novelty. Under Section 2(1)(l), a new invention is one that has not been anticipated by publication in any document or used in the country or elsewhere in the world before the date of filing. The search scope is therefore global and is not limited to patents.

Prior art includes:

• Indian and foreign patents and published applications
• Journal articles, conference papers, theses and dissertations
• Product literature — catalogues, brochures, manuals, datasheets, service documentation
• Websites, blogs, videos and archived web pages
• Standards documents and technical specifications
• Crowdfunding pages, product launch announcements and press releases
• Prior public use or sale of the product, anywhere in the world
• Oral disclosure in a public lecture or presentation
• Traditional knowledge, including material in the Traditional Knowledge Digital Library (TKDL) — a distinctly Indian and frequently decisive category

Section 13 governs the Controller’s own search for anticipation, covering both prior claiming in Indian specifications and prior publication in India or elsewhere.

The narrow statutory exceptions — Sections 29 to 34 — cover situations such as anticipation by previous communication to Government, public display at an exhibition notified by the Central Government with an application filed within twelve months under Section 31, and public working for the purpose of reasonable trial within one year under Section 32. These are far narrower than the grace periods available in some other jurisdictions, and they are not a substitute for filing before disclosure.

---

The Databases That Are Actually Searched

Indian

• InPASS — the Indian Patent Advanced Search System, covering Indian applications and granted patents
• Indian Patent E-Register — legal status, renewal position, oppositions, assignments
• TKDL — Traditional Knowledge Digital Library, essential for anything touching Ayurveda, Unani, Siddha, Yoga or traditional formulations

International patent

• WIPO PATENTSCOPE — PCT applications and many national collections
• Espacenet — the European Patent Office’s worldwide collection, with strong classification tools
• USPTO Patent Public Search
• Google Patents and Lens.org — broad coverage with full-text search and citation mapping
• National offices where relevant — EPO, CNIPA, JPO, KIPO

Non-patent literature

• IEEE Xplore, PubMed, ScienceDirect, SpringerLink, Google Scholar, arXiv, ResearchGate
• Standards bodies — BIS, ISO, IEC, IEEE
• Industry databases, trade publications and product archives

Specialist

• Sequence databases — GenBank, EMBL, for biotech and sequence listings
• Chemical structure databases — PubChem, Reaxys, SciFinder, for Markush and structural searching
• INPADOC for family and legal status data

A search restricted to InPASS and Google is not a patent search. Because India applies absolute novelty, a document published in a Japanese application or a German thesis destroys novelty just as effectively as an Indian patent.

---

Search Methodology: How It Is Actually Done

Step 1 — Invention disclosure and decomposition. The invention is broken into its technical features. Not “a smart water purifier”, but each element: the sensing mechanism, the control logic, the filtration arrangement, the fault detection method, the specific combination. The search is run against features, because a patent is granted for a combination of features, not for a product name.

Step 2 — Keyword generation. Synonyms, industry jargon, academic terminology, functional descriptions, British and American spellings, and terms as they would be translated in foreign filings. Most missed prior art is missed because the earlier document called the same thing by a different name.

Step 3 — Classification search. This is where amateur searches fail. The International Patent Classification (IPC) and Cooperative Patent Classification (CPC) systems index inventions by technical subject matter regardless of the words used. A classification search finds documents that no keyword string would ever surface. Relevant classes are identified, verified, and searched systematically.

Step 4 — Structured querying. Boolean operators, proximity operators, truncation and field-restricted searching, with the strings recorded so the search is reproducible and auditable.

Step 5 — Citation analysis. Backward citations from relevant documents, forward citations to see who cited them, and family analysis to catch equivalents filed in other jurisdictions.

Step 6 — Assignee and inventor searching. Known competitors, research institutions and prolific inventors in the field are searched directly.

Step 7 — Non-patent literature sweep. Academic, technical and commercial sources.

Step 8 — Screening and relevance grading. Results are screened and graded on the conventional scale — documents that destroy novelty on their own, documents relevant only in combination with others, and documents of general background relevance.

Step 9 — Feature mapping. Each proposed claim feature is mapped against each relevant prior art document in a matrix. This is the analytical heart of the exercise and the part that converts a list of documents into an opinion.

Step 10 — Iteration. The search is refined as the art reveals how the field describes itself.

---

The Patentability Assessment

The search produces documents. The assessment produces an answer.

Novelty — Section 2(1)(l)

Assessed against each prior art document individually. For an invention to lack novelty, a single document must disclose all the features of the claim. Combining two documents does not defeat novelty — that is an inventive step question. This distinction matters enormously in practice, because an invention frequently survives novelty and fails on inventive step.

Inventive step — Section 2(1)(ja)

India’s definition is demanding. It requires a feature that involves technical advance as compared to existing knowledge, or having economic significance, or both, and that makes the invention not obvious to a person skilled in the art.

Unlike novelty, prior art may be combined — but only where a person skilled in the art would have had a reason to combine them. The assessment considers:

• What the person skilled in the art knew at the priority date
• Whether the prior art contains any teaching, suggestion or motivation to arrive at the invention
• Whether the invention solves a technical problem in a non-obvious way
• Whether the result is a genuine technical advance or merely a predictable workshop modification
• Secondary indicators — long-felt need, commercial success attributable to the technical feature, failure of others, and teaching away in the prior art

Hindsight is the enemy of a reliable opinion. Once you know the answer, every route to it looks obvious. A disciplined assessment reconstructs the position of the skilled person before the invention existed.

Industrial application — Section 2(1)(ac)

Capable of being made or used in an industry.

The Sections 3 and 4 screen

This is the Indian-specific step that no international search report will do for you. An invention can be perfectly novel and inventive and still be excluded subject matter under Section 3 — a business method, a computer programme per se, a method of medical treatment, a new form of a known substance without enhanced efficacy, a mere admixture, a method of agriculture, or an aggregation of traditional knowledge. Running this screen early prevents a great deal of wasted expenditure.

Sufficiency and claim scope

Finally, the assessment addresses what can realistically be claimed in light of the art: the broadest defensible independent claim, and a set of narrower fallback positions that survive if the broad claim is attacked. This is where the search directly shapes the drafting.

---

What a Proper Report Contains

1. Invention summary and the feature decomposition used
2. Search strategy — databases covered, classification codes searched, search strings used, and the date of the search
3. Prior art list, with each document graded for relevance and its bibliographic and legal status data
4. Feature mapping matrix — every claim feature against every relevant document
5. Novelty opinion, with reasoning against each closest document
6. Inventive step opinion, addressing the closest prior art and any obvious combinations
7. Sections 3 and 4 analysis
8. Recommended claim scope, with fallback positions
9. Clear recommendation — file as is, redraft with a narrower scope, restructure the claims, defer and develop further, protect as a trade secret, or do not file
10. Stated limitations — including the eighteen-month publication blackout, language and literature coverage, and the impossibility of finding unpublished or undocumented prior art

---

Freedom to Operate: The Search That Protects the Product

An FTO analysis asks a completely different question, and the discipline is different.

Only live patents matter. Expired, lapsed, revoked and abandoned patents are free to use. Every candidate must be checked on the register for renewal status, oppositions and revocation proceedings.

Only the claims matter. Infringement is assessed against the claims, not the title, the abstract or the description. A patent whose description discusses your product but whose claims do not cover it is not an obstacle.

Territory is everything. Patents are national rights. An Indian manufacturer selling only in India need not clear a US patent — but the moment exports begin, the analysis has to be run again for each destination market.

Every commercial act is separately relevant. Making, using, selling, offering for sale and importing are each acts of infringement. Manufacturing in India for export can infringe an Indian patent even if the product is never sold here.

The outputs of a good FTO analysis:

• A list of live patents whose claims potentially read on the product
• A claim-by-claim, element-by-element mapping for each
• A risk grading for each identified patent
• Design-around options — technical modifications that move the product outside the claims
• Validity considerations — whether a blocking patent looks vulnerable on prior art, which converts a blocker into a negotiating position
• Licensing routes, where design-around is not commercially viable
• An assessment of remaining risk, stated honestly

---

Sector-Specific Considerations

Pharmaceuticals and life sciences. Markush structure searching, polymorph and salt-form art, formulation patents, and a rigorous Section 3(d) analysis on enhancement of known efficacy. Sequence searching for biologics. Patent family and legal status mapping across markets is essential because generic entry strategy turns on it.

Software and computer-related inventions. The determining question is usually not novelty but Section 3(k). Non-patent literature — open source repositories, technical blogs, standards documents and academic papers — is often more relevant prior art than patents.

Biotechnology. Section 3(j) exclusions for plants, animals and essentially biological processes, sequence database searching, and deposit requirements for micro-organisms.

Mechanical and electrical engineering. Classification searching is particularly productive, and older art matters — a 1970s patent that expired decades ago still destroys novelty today.

Traditional knowledge based inventions. TKDL searching is not optional. Section 3(p) excludes inventions which in effect are traditional knowledge or an aggregation or duplication of known properties of traditionally known components, and TKDL has been used extensively to defeat such claims internationally.

Startups and limited budgets. A staged approach works: a focused knock-out search first to see whether the concept survives at all, then a fuller search and opinion if it does, then FTO closer to launch. Spending the entire budget on a filing that a two-day search would have stopped is the worst allocation available.

---

Where Searches Go Wrong

1. Searching only Indian databases, when absolute novelty makes worldwide art relevant
2. Keywords only, no classification search — the single biggest cause of missed art
3. Treating a patentability search as an FTO search, and launching a product that infringes
4. Searching after the invention has been publicly disclosed, when novelty is already gone
5. Searching after the specification is drafted, so the art cannot shape the claims
6. Ignoring non-patent literature, particularly in software and life sciences
7. Ignoring TKDL for anything touching traditional formulations or practices
8. Failing to check legal status in an FTO, and treating an expired patent as a blocker
9. Assessing FTO against abstracts rather than claims
10. Treating “nothing found” as proof of novelty rather than as the limited finding it is
11. Not documenting the search strategy, so the work cannot be relied on or repeated
12. Ignoring the eighteen-month blackout and presenting the result as conclusive
13. Hindsight reasoning in the inventive step opinion, either overstating or understating obviousness
14. Running an FTO for one territory and then exporting into another

---

How Delhi Legal Company Conducts Searches and Assessments

• Structured invention disclosure — a confidential technical intake that decomposes the invention into claimable features before any searching begins
• Multi-database searching — InPASS, PATENTSCOPE, Espacenet, USPTO, Google Patents and Lens, with TKDL and specialist databases where the subject matter requires them
• Classification-led strategy — IPC and CPC codes identified and searched systematically alongside keyword and citation approaches
• Non-patent literature coverage appropriate to the field
• Feature mapping matrix and a graded prior art list, with the full search strategy documented and reproducible
• Written opinion on novelty, inventive step, industrial application and Sections 3 and 4, with a recommended claim scope and fallback positions
• Freedom to operate analysis with claim-element mapping, legal status verification, risk grading, design-around options and validity assessment of blocking patents
• Validity and invalidity searches for opposition, revocation and litigation support
• Landscape and white space reports for R&D and investment decisions
• Honest recommendations, including the recommendation not to file where the art does not support it

---

Frequently Asked Questions (FAQs)

1. What is a patent search?

A. A patent search is a structured investigation of patent and non-patent literature to identify prior art relevant to an invention. Depending on the question being asked it may be a patentability search, a freedom to operate search, a validity search or a landscape study, and each of these uses a different scope and different criteria.

2. Is a patent search compulsory before filing?

A. No. There is no requirement under the Patents Act, 1970 to conduct a search before filing. It is, however, the cheapest step in the entire process and the one that determines whether the rest of the expenditure is justified. Filing without a search is how applicants find out about the closest prior art from an examiner three years later.

3. What is the difference between a patentability search and a freedom to operate search?

A. A patentability search asks whether your invention is new and inventive enough to be granted a patent, and looks at all prior art worldwide including expired and abandoned patents. A freedom to operate search asks whether your product would infringe a live, in-force patent in a particular territory, and looks only at the claims of patents that are currently in force there. The answers are independent of each other.

4. Can I have a valid patent and still infringe someone else’s?

A. Yes, and this surprises people constantly. A patent gives you the right to stop others from practising your claimed invention; it does not give you the right to practise it yourself if doing so falls within someone else’s earlier claims. Improvement patents very commonly sit inside a broader third-party patent.

5. What counts as prior art in India?

A. India applies absolute novelty, so anything published in any document or used publicly anywhere in the world before the priority date counts. This includes foreign patents, journal articles, theses, product catalogues, manuals, websites, videos, standards documents, crowdfunding pages, prior public use or sale, and traditional knowledge recorded in the TKDL.

6. Does my own disclosure count as prior art against me?

A. Yes. An inventor’s own prior publication, presentation, demonstration or sale destroys novelty just as effectively as a third party’s. The exceptions in Sections 29 to 34 are narrow — principally display at an exhibition notified by the Central Government with an application filed within twelve months under Section 31 — and they rescue very few applicants.

7. Can any patent search be completely exhaustive?

A. No, and you should be wary of anyone who says otherwise. Applications are not published until eighteen months from the priority date, so recent filings are structurally invisible. There are also gaps in language coverage and non-patent literature, oral disclosures that leave no document, and classification errors that hide art in unexpected places. A search materially reduces risk; it does not eliminate it.

8. What is the eighteen-month blackout?

A. Under Section 11A, applications are published only after eighteen months from the priority or filing date. A search conducted today therefore cannot see anything filed in the preceding eighteen months. Any competent search report states this limitation expressly.

9. Which databases should be searched?

A. At minimum InPASS for Indian applications and patents, together with PATENTSCOPE, Espacenet, the USPTO collection and a full-text tool such as Google Patents or Lens. TKDL must be searched for anything touching traditional knowledge, and non-patent literature databases are essential in software and life sciences. Searching only Indian databases is inadequate given absolute novelty.

10. Why is classification searching important?

A. Because keyword searching only finds documents that use your words. The IPC and CPC systems index inventions by technical subject matter regardless of terminology, so a classification search surfaces documents that describe the same concept in entirely different language. Most missed prior art is missed for exactly this reason.

11. How is novelty assessed?

A. Against each prior art document individually. For an invention to lack novelty, a single document must disclose all the features of the claim. You cannot combine two documents to defeat novelty — combining documents is an inventive step question.

12. How is inventive step assessed?

A. Under Section 2(1)(ja), the feature must involve technical advance compared with existing knowledge, or have economic significance, or both, and must not be obvious to a person skilled in the art. Prior art may be combined here, but only where the skilled person would have had a reason to combine it. The assessment must avoid hindsight, which makes almost everything look obvious after the fact.

13. My invention is novel. Does that mean I will get a patent?

A. Not necessarily. Novelty is only the first hurdle. The invention must also involve an inventive step, be capable of industrial application, and — critically in India — must not fall within the exclusions in Sections 3 and 4. A great many novel and commercially valuable inventions are simply excluded subject matter here.

14. Will a search tell me whether my software invention is patentable in India?

A. A search tells you about the prior art. For software the more decisive question is usually Section 3(k), which excludes computer programmes per se, business methods, mathematical methods and algorithms. That analysis should be done alongside the search, because it frequently determines the outcome regardless of what the prior art shows.

15. How long does a patent search take?

A. A focused knock-out search can be turned around in a few working days. A full patentability search with a written opinion typically takes one to two weeks depending on the complexity of the field. A freedom to operate analysis takes longer, because every candidate patent requires claim-by-claim mapping and legal status verification.

16. What does a good search report contain?

A. The invention summary and feature decomposition, the full search strategy including databases, classification codes, search strings and the date of the search, a graded list of prior art, a feature mapping matrix, reasoned opinions on novelty and inventive step, a Sections 3 and 4 analysis, a recommended claim scope with fallback positions, a clear recommendation, and a candid statement of limitations.

17. What is a feature mapping matrix?

A. A table setting out each feature of the proposed claims against each relevant prior art document, showing which features are disclosed where. It is the analytical step that turns a list of documents into an opinion, and it is what allows the drafter to see exactly where the claims must be positioned.

18. When should I do a freedom to operate search?

A. Before committing to manufacturing, before product launch, before entering a new territory, and before an investment or acquisition. It should be repeated for each market you export into, since patents are national rights and clearance in India says nothing about clearance elsewhere.

19. What if the FTO search finds a blocking patent?

A. There are usually options. You can design around it by modifying the product so it falls outside the claims; you can assess its validity, since a vulnerable patent is a negotiating position rather than a wall; you can seek a licence; you can check whether it is actually in force and renewed; or you can confine your activity to territories where it does not exist.

20. What is a validity or invalidity search?

A. A search focused on finding prior art that predates a specific granted patent, in order to challenge it in opposition or revocation proceedings or to defend an infringement claim. It is deeper and more targeted than a patentability search because the target claims are already known and fixed.

21. What is a patent landscape study?

A. A broader analysis mapping who is filing in a technology area, what is protected, how activity has changed over time, and where the unprotected white space is. It is used for R&D direction, investment decisions, licensing strategy and competitor monitoring rather than for a specific filing decision.

22. Can I do the search myself using free databases?

A. You can run a preliminary check, and it is a sensible first step. What free searching generally misses is classification-led strategy, non-patent and foreign-language literature, citation and family analysis, and — most importantly — the legal assessment of novelty, inventive step and the Section 3 exclusions, which is where the actual answer lies.

23. The search found nothing. Does that guarantee my patent?

A. No. A nil result means nothing relevant was found within the scope searched, which is not the same as nothing existing. Between the eighteen-month blackout, literature and language gaps and undocumented prior use, a nil finding is useful evidence but never a guarantee.

24. Should I search before or after drafting the specification?

A. Before, always. The purpose of the search is not merely to decide whether to file but to shape what is claimed. A specification drafted before the art is known will usually claim too broadly, and the matter needed to support narrower fallback positions may never have been disclosed — and it cannot be added later.

25. What if the search shows my invention is not patentable?

A. That is a valuable outcome, not a wasted exercise. Options include narrowing the claims to a genuinely inventive feature, further development to create a real technical advance, protecting the invention as a trade secret where it cannot be reverse engineered, pursuing design registration if the value lies in appearance, or reallocating the budget entirely. Learning this before filing saves years and a great deal of money.

26. What does Delhi Legal Company charge for a search and patentability opinion?

A. It depends on the technical field, the scope of databases required and whether the deliverable is a knock-out search, a full patentability opinion or a freedom to operate analysis. We quote in writing before starting, and for early-stage clients we frequently recommend a staged approach so that the budget is spent in the right order.