Introduction to Global Patent Searching & Analysis
1. Global Patent Mining :
Tools & Strategies
2016
ŠJon Cavicchi
Professor of Legal Research
2. Learning objectives
⢠We only have a few weeks soâŚbe articulate
ď§ vocabulary & themes of patent searching
ď§ different ways patent data is used in legal and business settings
ď§ Sources & limitations of U.S., foreign, international and patent
family data
ď§ Spectrum of no fee to premium sources of said data
ď§ Professor Cavicchiâs iterative hybrid patent searching approach
transferable to scoping out any platform
ď§ Application of approach using the Thomson Innovation, Dialog
and free web platforms.
3. Three phases of patent
searches
Search Analysis
Report
Each search report is customized so
Course does not teach report production
Good search is foundation.
Search report walks through
sources & approaches
5. ⢠Before patent drawings were printed, examiners, while searching patents, had to look at the
original drawings in the draftsman's office in large portfolio cases
⢠There were 777 folio cases of original drawings which were safely removed from the draftsman's
office in the fire of 1877. But when the drawings, and especially the entire patents, were printed,
they were available for search in the examiners' rooms and soon in the new public search room.
The necessary new filing system was provided by shoe cases.
⢠The origin of the term shoe is lost, although every patent examiner knows where his shoes are.
Some have attributed the term to Thomas Jefferson, suggesting that he stored his patents in shoe
boxes. But we know that the drawings kept in the Patent Office in the preprinting days were of
varying sizes and were kept in portfolios from the earliest days. Shoes, as we know them, could
only have arisen after all patents were available in small, uniform sizes and could be kept in
small, uniform boxes.
⢠A complete inventory of the moveable property in the Patent Office was made in 1870, including
numerous portfolio cases, portfolio racks, portfolio drawers, and cases for models, 22,000
volumes of books and 300 spittoons, but no shoes. Augustus Burgdorf, livery stable operator,
undertaker and cabinet-maker of Washington, sold portfolios and cases to the Patent Office in
1878.
⢠The first known mention of shoes was on March 28, 1879, when he sold shoe drawers to the Patent
Office for $115. What were called shoe drawers would now be called shoes. Perhaps file cabinets
suitable for holding bundles of patents while allowing easy access to search through them were
already available before they were needed. Perhaps shoe shops of the day kept their supply of
ready-made shoes in wood cabinets containing numerous drawers of just the right size to hold
patents, and when the Patent Office wanted to order its first drawers, it ordered shoe drawers
from Augustus Burgdorf.
6. There are as many ways to
do prior art searches as there
are prior art searchesâŚ
Each search is unique
Dozens of searching platforms
9. AIA Statutory Framework
Prior Art
35 U.S.C. 102(a)
(Basis for Rejection)
Exceptions 35 U.S.C.
102(b)
(Not Basis for Rejection)
102(a)(1)
Disclosure with Prior
Public Availability Date
102(b)(1) (A)
Grace Period Disclosure by Inventor or
Obtained from Inventor
(B)
Grace Period Intervening Disclosure by
Third Party
102(a)(2)
U.S. Patent,
Published U.S. Patent
Application, and
Published PCT
102(b)(2) (A)
Disclosure Obtained from Inventor
(B)
Intervening Disclosure by Third Party
Application with Prior
Filing Date
(C)
Commonly Owned Disclosures
10. ⢠Big databases
ď§ EPO/INPADOC (77)
ď§ World Patent Index (41)
ď§ WIPO PATENTSCOPE
ď§ PCT (122)
ď§ National Authorities
ď§ Lexis Total Patent (100)
ď§ ESPACENET v5. - ?
ď§ WIPO PATENTSCOPE -?
NO WORLD PATENT DATABASENO WORLD PATENT DATABASE
11. Growth of EPO & WIPO CollectionsâŚa threat?
⢠Having been in this industry for over 10 years my sense is
that the free services have become an impediment to the
growth and development of more robust offerings and
analytic capabilities from the private sector.
ď§ Peter Vanderheyden, Former Vice President, Global
Intellectual Property, LexisNexis
12. Much of the worldâs patent
literature not searchable
14. Multiplicity of sourcesMultiplicity of sources
Types of sourcesTypes of sources
ApplicationsApplications
Multiple access points to same dataMultiple access points to same data
Who uses patent data sourcesWho uses patent data sources
Why use patent data sourcesWhy use patent data sources
Factors to choose access pointsFactors to choose access points
Search approachSearch approach
We will discuss todayâŚ
15. Small piece of patent informaticsâŚ
Finding prior artFinding prior art
patent as datapatent as data
non patent datanon patent data
Competitor intelligenceCompetitor intelligence
Statistical usesStatistical uses
Patent citations & innovationPatent citations & innovation
Patent mapping & analyticsPatent mapping & analytics
Identifying intellectual assetsIdentifying intellectual assets
Manipulate & present data in graphicalManipulate & present data in graphical
17. Who uses patent data?
⢠academics
⢠policy experts
⢠government officials
⢠commercial searchers
⢠research scientists
⢠inventors
⢠Engineers
⢠managers
⢠marketing
⢠IP attorneys
⢠patent agents
⢠patents searchers
⢠law firm
⢠corporations
⢠PTO examiners
⢠licensing pros
⢠technology transfer
pros
18. Why do they use patent data?
⢠Intelligence &
reconnaissance
ď§ watch technology
develop over time
ď§ id new opportunities
ď§ protect IP
ď§ save $ before
proceeding
ď§ who is active
⢠Focus on company or
inventor
ď§ what are they up to?
ď§ licensing
opportunities?
ď§ monopolies?
ď§ block monopolies?
19. ⢠how companies can tap the asset values in intellectual
property to achieve board and shareholder objectives, we
pay particular attention to the use of patents to:
ď§ Generate new revenues through licensing
ď§ Boost earnings per share and total shareholder return
ď§ Improve return on investment for R&D and seed
continuing innovation
ď§ Raise corporate valuations and enhance equity and
other financing efforts
ď§ Serve as the currency of mergers, acquisitions, and
joint ventures
âPatent data is the greatest source of
competitive intelligence on earthâ ~ Rivette
20. â˘Plot competitors' product strategies,
as well as ways to "patent-block" them
â˘Gain patent-protected entry into
lucrative but hotly contested markets
â˘Acquire exclusive rights to emerging
market-leading technologies
â˘Increase R&D effectiveness and avoid
infringement minefields
â˘Detect possible infringers, as well as
likely sources of licensing income
22. Patent search more than
searching patents
Patent search involves reference to
publications other than patents
23. Definition of âprior artâ is broadâŚ
⢠âThe existing body of technological
information against which an invention is
judged to determine if it is patentable as
being a novel and nonobvious inventionâ
⢠McCarthyâs Desk Encyclopedia of IP, 2d Ed.
⢠Reference to Patent Act and cases.
27. Patent Act
⢠Section 101 requires that an invention have
novelty & utility
⢠Section 102 defines novelty and creates
statutory bars
⢠Section 103 requires that the invention be a
nonobvious development over that prior art
28. Section 102
⢠Novelty provisions focus on certain events
constituting anticipation
⢠Statutory bar focuses on events that may
occur prior to the patent application
⢠Domestic or foreign: prior patent, printed
publication, public sale, use, invention by
others...
29. DisclosureâŚ
⢠Printed publication accessible to the public
anywhere in the worldâŚ
ď§ Manufacturers catalogs
ď§ Product specification sheets
ď§ Conference papers
ď§ Articles in journal
ď§ Academic theses in libraries
ď§ Web content
⢠Anywhere you find evidence of public use,
advertising, salesâŚ
30. Jorda to Ciba Geigy Pharma R&D
Five events that trigger patent search
⢠before filing a patent application
⢠before entering a new field of research
⢠before commencing manufacture of a new product
or use of a new process or improved products or
processes
⢠before taking a license under someone elseâs
patent
⢠before patent litigation
31. Patent Searching
Terminology is Messy!
⢠Lawyers, scientists, business, IP,
international development professionals use
labels without a common reference
vocabulary
⢠Leads to âsymantic disconnectâ
⢠Get as much detail as to the purpose and
nature of the search, analysis, reportâŚ
34. FREEDOM TO OPERATE
â˘a.k.a. âRight to Make Searchâ, âClearance Searchâ,
âRight to Use Searchâ, âJustification Searchâ
â˘Includes an infringement search in addition to
expired patents that may provide a safe haven for a
specific product, process, or serviceâŚaddress both
infringement and patentability concerns.
35. RIGHT TO USE
to determine whether a patent has
expired & can be copied with
impunity
36. VALIDITY
before taking a license or litigation to
determine whether a patent will
withstand attack
37. PATENT LANDSCAPE
STATE OF THE ART
⢠Patent landscaping uncovers:
ď§ Market opportunities for licensing
ď§ Competitive technology trends
ď§ Gaps in corporate patent portfolios
ď§ Opportunities for entry into new markets
ď§ Targets/risks for patent infringement litigation
ď§ Opportunities for redirecting R&D efforts
39. FAMILY/CORRESPONDING
⢠whether a corresponding patent has been
filed with a filing entity
⢠to map out the countries where the
competitors' patents or utility models form a
potential business hindrance.
⢠to reduce the risk of infringement cases
with potential claims for damages.
⢠to obtain information about changes in the
ownership of a patent or a utility model.
43. TYPICAL PATENT DATA PATH
⢠Source of Patents
ď§ PTO
ď§ sell the same data to anyone
⢠Publisher or producer
ď§ corrects, indexes, abstracts, packages data
⢠Vendor
ď§ factory outlet
ď§ department store
ď§ boutique
44. 3 TYPES OF PATENT DATABASES
⢠#1 Bibliographic
ď§ not full text
ď§ citation, dates, class, abstract
⢠Full Text
ď§ basic or enhanced record
ď§ Human OR Machine Translated OR Machine
Assisted
⢠Hybrid
ď§ full text plus abstract...
45. Online services organize data
differently--some models
⢠Trend is to âmenuizeâ and âtemplatizeâ
ď§ Innovation, TotalPatent and most others
⢠Lexis
ď§ Library, file, documents, segments, word
⢠Westlaw
ď§ Topical area, database, documents, fields, term
⢠DIALOG
ď§ file, record, (index, suffix, prefix, field), string
⢠QUESTEL/ORBIT
ď§ file, record, (field, index, qualifier), string
48. DO WE DO A SEARCH?
⢠Patentability Search
⢠Client wonât pay
⢠Client approves a very small budget
⢠âLet the Patent Office do the searchâ
⢠âLet the inventor do the searchâ
⢠âWe donât care if the patent is invalidâ
⢠Unintended downstream consequences?
49. Patent document as a commodity?
⢠A patent is a patent is a patent?
ď§ yes and no
ď§ same basic record
ď§ record indexing & output differs
ď§ quality issue
ď§ time issue
ď§ cost issue
⢠Always ask what a producer does to the
recordâŚâwhat is the value addedâ to your
service?
50. IN HOUSE OR FARM IT
OUT
law firms and corporate department
have no uniform practices
51. Factors to consider...
⢠Past practice
⢠Cost
⢠Hourly rate
⢠Control
⢠Liability & ethics issues
⢠Purpose of search
⢠Type of search
ď§ target searches
⢠Sophistication of in house staff
⢠Trust in manual & electronic searches
ď§ missing patents both online & at PTO
52. Cast of searchersâŚ
⢠Large Search Firms (Langdon IP &
Cardinal IP)
⢠Small Search Firms (DEMARCOIP)
⢠Solo searchers (Stephen Key)
⢠Technical Search Firms (NERAC)
⢠Unregulated collection of searchers with
varying degrees of technical and legal
background
57. Factors to consider...
⢠No standardâŚevery office has own approach
⢠LawyerâŚlibrarianâŚparalegalâŚsubject
specialistâŚinventor
⢠Depends on purpose of search
ď§ novelty & state of art may differ from validity or
infringement search
⢠Balance familiarity with subject matter with
searching skills
⢠Again--cost is a factor
58. WHAT IS THE STANDARD
OF CARE?
little guidance from the law
59. Some observations...
⢠No statutory duty
⢠No negligence or malpractice cases
⢠Conduct Rule 1.1 Duty of Competence?
⢠Rule 56 Duty of Candor
ď§ Bury a citation = hiding from examiner?
67. Web is both alternative & complimentary
⢠Depends on why you are doing search
ď§ some bibliographic only
ď§ some lack coverage
ď§ some admit missing records
ď§ some lack images
⢠May choose as document delivery option
only
68. NPL
⢠Patent and non patent literature
Thousands of
of non-legal print &
database sources
70. Defensive publicationsâŚ
⢠The IP.com Prior Art Database is an excellent solution for
companies who wish to publish their technical disclosures (defensive
publications) in a well-known, library indexed, publicly searchable
database specifically dedicated to the promotion and publishing of
prior art data.
76. A good prior art search is an
insurance policy to avoid
validity issue expenses
Professor Tom Field
77. Assumption for this class..
You will do the search from your
desktop BUT NOT using the Public
Search Room at the USPTO or its
tools.
78.
79. Approaches
⢠Case by case basis
⢠Totality of the circumstances
⢠No model approach
⢠No step by step approach
⢠No best practices per se
81. Recall & precision
⢠Recall describes the idea of all items which are relevant
(useful) to a query. In the real world, only a subset of the
relevant items are found.
⢠Precision describes the idea of only those items which are
relevant to a query. Again, in the real world, many items
which are matched by a query are not really relevant to the
question, although they might match the vocabulary.
⢠In information retrieval, there's a classic tension between
recall and precision. Specifying more recall (trying to find
all the relevant items), you often get a lot of junk. If you
limit your search trying to find only precisely relevant
items, you can miss important items because they don't use
quite the same vocabulary.
82. The issue of patents for new
discoveries has given a
spring to inventions beyond
my conception.
- Thomas Jefferson
83. Patent searchersâ conundrumâŚ
⢠Retrieve highly specific and tailored results
with as few marginal, irrelevant hits as
possible
⢠Rarely do searchers need to see every
related item
⢠Patent searchers hope to find nothing, but
by finding nothing they cannot be sure
whether or not the job is really finished
84. Problems with patent
searching
⢠Recall more important than highly tailored
results
⢠Precision not primary consideration
⢠False drops included as solution occurs in
millions of patent and non-patent
documents
85. Other difficulties
⢠KEYWORD OBFUSCATION
⢠Short, meaningless titles and abstracts
⢠Patent documents notorious for vagueness
⢠Language may be abstract - patent attorney is own
lexionographer
ď§ Frisbee = levitating disk
⢠Vocabulary may not be standardized or even exist
ď§ Kevlar = optically anisotrophic aromatic polyamide
dopes classed with synthetic resins and not tires or
bullet proofing
86. Subject descriptors
⢠Patent classification schemes
ď§ Bloated?
ď§ Out of date?
ď§ Patented invention may be in different
technology from that in which it is eventually
applied?
⢠Velcro = classed in stock materials while
applications found in medical and amusement
devices
87. New - CPC
⢠PTO and European Patent Office team
up on classification system
ď§ The National Law Journal
ď§ October 25, 2010
⢠The latest U.S. Patent and Trademark Office
innovation is a deal to work with the European
Patent Office on a joint patent classification
system aligned with global standards.
89. U.S. Patents Riddled With Mistakes,
Survey Finds
⢠An astounding 98% of approved U.S. patent applications contain
mistakes ranging from simple spelling errors to omitted claims.
⢠The mistakes were uncovered by Intellevate, the worldâs
largest patent proofreading organization. More than half of
the mistakes it found at its office in India were made by the
U.S. Patent and Trademark Office, according to Intellevate
chief executive Leon Steinberg.
⢠We find mistakes on everything from leaving out portions
of the patent claims to putting in the wrong drawings, to
spelling mistakes, Steinberg said. Spelling mistakes like
those can have significant consequences for both the
inventor and the consumer.
90. Step 1 :get to know the idea
or invention
⢠Spectrum from idea to reduction to practice
⢠Too little informationâŚtoo much
information
⢠Use the inventor or management to explain
the invention
⢠What is the inventive step?
⢠Is there other information you need?
⢠Ask about non patent literature in field
91. What type of inventions
patentable?
⢠Device
⢠Process
⢠Improvement
⢠Kit
⢠Machine
⢠Manufacture
⢠Composition of matter
or material
⢠Business method
⢠Design
⢠Plant
⢠Software
⢠chemical compositions
⢠mixtures of ingredients as
well as new chemical
compounds.
⢠Anything under the sun!
92. Step 2: what is the scope of
your search?
⢠What is the scope of protection you are
looking for?
⢠Scope will play role in formulation of
search terms and choice of databases
⢠Disclosed information is disclosed
information!
93. Step 3: formulate your initial
list of search approaches
⢠What data do you have at this point?
⢠What are potential search approaches?
⢠Non-subject access points
ď§ Inventors?
ď§ Assignee?
ď§ Patent number?
ď§ Classification?
94. Patent search is an iterative
process
⢠âŚcontinuous modification of searches as
more information becomes available
⢠Recheck
⢠Locate
⢠Lead
⢠Tips
95. Step 4: formulate initial list of
subject search terms
⢠List all essential parts or steps
ď§ Think of the parts of the front page
ď§ Descriptive
ď§ Function
ď§ How it works
ď§ How it is structured
ď§ Results produced
ď§ Steps involved
ď§ Problem solved
96. Step 5: choose database(s)
⢠Many to choose from
⢠Each differs somewhat in content and
search capabilities
⢠Crossfile searching can be useful
ď§ Ease of search varies depending upon the extent
to which the files have been âharmonizedâ for
crossfile searching
97. Step 6: choose which
service(s) to use
Spectrum from free to premium services
⢠Multiple access points to the same data
⢠Get to know what value added each service offers
ď§ Content
ď§ Searchability
ď§ Cost
ď§ Employer preferences
ď§ Data output options / analytics
ď§ Time
ď§ Familiarity
99. What is the value added?
Content
Coverage
Scope
Enhancements
Accuracy
Error rate
Ease of use
Search sophistication
Integration
Output flexibility
Customer support
Training
Pricing options
100. Alternative approaches
âThe 1 click that takes 1000 clicks on other servicesâ
⢠Search and examination relied on operator quality
and could not be held to an empirical standard.
⢠Heuristics
⢠Latent semantic analysis
⢠Natural language
102. Step 7: strategize on which
order you use services
⢠Know the content on each
⢠Start with U. S. patents fulltext?
⢠Want to search one database or many
⢠Begin the data harvesting
103. Step 8: approach the service
⢠Scope out each database you plan to search
⢠What data is in the database?
⢠How is the data arranged?
⢠What are the searchable parts (fields,
segments, indexesâŚ)
118. Proximity connectors
⢠specify the relative nearness or adjacency of
search terms
⢠They are used in two-word or multiple-word
phrases or phrases that have punctuation or stop
words.
⢠Proximity connectors on Dialog include (w), (n),
(#w), (#n), and (s)
119. Step 10: FINALLY formulate
some searches
⢠Classification search - controlled approach
⢠Keywords - uncontrolled approach
⢠Hybrid - best of both worlds
⢠Think like a patent examiner
⢠Keep the structure of the document in mind
120. But why canât I just plug in
my search terms?
ď§ Full text search may be OK in small records
ď§ FULL TEXT BIGâŚLONG RECORDSâŚ
MASSIVE DATABASESâŚ
⢠Search fields
⢠Use proximity connectors
ď§ Title, Abstract searches find most relevant
ď§ Do class search and use keywords within the
class
121. What ifâŚ
⢠I get no hits
ď§ Try other searches
ď§ Show me the approached/searches you tried
ď§ Move along to next database
⢠I get tons of hits
122. Broaden & narrow searchâŚ
Unrestricted full text
Try different field
restrictors:
Class
Spec
Abstract
Title
Mix & match with keywords
Bibliographic file
123. Why should I have to review 5,000
patents to find the 200 relevant ones?
⢠Everyone, including searchers, has an instinctive
feeling that searching should be more
straightforward. Was my capture strategy inefficient?
The answer is, "No." Try this. After you have
completed a search and have your list of relevant
patents, try to rewrite your capture strategy to capture
only the relevant patents without capturing thousands
more. You cannot. This frustrating exercise will show
you just how noisy the capture part is. So culling
remains necessary because capture methods are just
too noisy. Searchers work within the noise to deliver
relevance.
ď§ 2004 WL 64936393
124. Step 11: mine the data
⢠Cull the results
⢠Remove duplicates?
⢠View results
ď§ Bib formats
ď§ Full text
ď§ Document delivery
125. Patent search is an iterative
process
⢠âŚcontinuous modification of searches as
more information becomes available
⢠Recheck
⢠Locate
⢠Lead
⢠Tips
126. Step 12: present the data
⢠Again, no standard
⢠How much data to collect
ď§ Pay per view - $ - $$ - $$$ - $$$$
ď§ Minimal citation ~ let lawyer ask for full
text
⢠How to manage data
⢠How to present data
ď§ Patent searchers donât provide legal
opinions
ď§ Do you need pretty pictures?
127.
128. DevelopmentsâŚ
⢠AIA?
⢠Common Patent Classification
⢠Patent offices outsource searches?
⢠Certified Patent Searchers?
⢠Accelerated Patent Examination: applicant
provides an appropriate search of the prior
art and an improved explanation of the
claims and prior art found.