Documentos de Académico
Documentos de Profesional
Documentos de Cultura
e4
Patents are a valuable source of information which, if analyzed, can help to generate
knowledge about the relative positions of the different players or establish the state
2006
of the art in a given field. The report describes how patent professionals exploit and
utilize software packages and it compares to the features of the evaluated software Evaluation of softwares
packages. It also shows the value attached to the characteristics provided by the
producers. We think that this report, unique in his work, offers a framework for
and technological intelligence needs
those working with intellectual property”
Published by:
EMECOM Ediciones in collaboration with
PUZZLE - Revista Hispana de la Inteligencia Competitiva (www.revista-puzzle.com).
Printed in Spain
© Copyright 2006: Juan Carlos Vergara, Alessandro Comai and Joaquín Tena Millán.
No part of this publication, including cover design, may be reproduced, stored in a retrieval system or transmitted,
in any form or by any means, electronically or optically, without the prior written permission of the publisher.
TABLE OF CONTENTS
- vi -
- vii -
TABLE OF CONTENTS
We would like to thank all those individuals and companies who have taken
part in our research, completing the questionnaire and attending to our
requests for information. We would like to express our gratitude also to all
the other individuals and companies we contacted in the course of this study,
for the attention and time they have given us and for the interest they have
shown.
We would also like to thank PUZZLE Magazine for providing us with the
space and resources with which to prepare and carry out the poll.
- xii -
SECTION ONE
Main Findings
- 14 -
- 15 -
1. MAIN FINDINGS
Five companies took part and allowed us access to their patent analysis
software, which we were then able to evaluate.
The model was applied to both supply and demand; in other words, our
research looked, on the one hand, to experts in patent analysis for their
evaluation of the functions and characteristics specified in the model, whilst
on the other hand, we also assessed, according to the same model, the software
created by individuals or companies to which we had access.
Only a few functions within this group, such as “Ability to import patent
records” for instance, adapt to demand.
show that only two of the five software programs cover 50% of these
functions.
“Graphic generation” shows positive global results, although there are several
areas which are not dealt with quite as persistently as, for instance, graphic
and statistical exploitation of the searches carried out. “Space or topographic
representation of a patent collection – text mining analysis” or “Ability to
use local databases to integrate new data and complete the patent analysis”
are hence barely covered by the software studied, when they are in fact used
relatively frequently.
2. INTRODUCTION
A number of studies and articles in this field have shown the importance and
benefits of carrying out more or less sophisticated patent analysis (Vergara,
- 20 - Software for Technological Patent Intelligence
2004; Rodriguez, 2003; Lozano, 2003). Other work has focused on tackling
patent exploitation (Paap, 2002; Adams, 2006) or how to organize a systematic
collection process using patents as a primary source.
As we mentioned earlier, computer applications make the job of statistical
analysis or the preparation of patent maps far simpler, thus giving rise to
Competitive Technological Intelligence (CTI). It has been shown that
computer applications can have a very large number of characteristics and
functions aiding the work of experts in this area. For this reason, it is currently
highly important that we be aware that applications do exist in the market
and that we know which of them can best meet the needs of professional
individuals working in CTI.
- The lack of studies on the demand for software for patent analysis.
That is, the lack of awareness regarding the use of and the value
attached to the characteristics provided by the producers of this type
of computer application.
- Non-existence of any comparison of applications together with a
need expressed by users of said products.
- To obtain an assessment of the magnitude and growth of the supply.
We recorded over 21 applications existing in the market5, which in
our opinion, is an extensive supply for this specialized field. Another
trend which stands out is the increase, if only marginal, in the
number of this type of application.
- We are dealing with a wide range of available computer applications
for PA. An application can have a very large number of functions.
However, applications currently existing in the market include
different groups of functions and it is, as a result, hard to make any
kind of partial comparison of them. For this reason, we believe that
it is necessary to standardize or have a uniform approach to the study
of these applications in order to make the comparison valid.
The answers to these key questions are given in the following chapters.
**********************************
Footnotes
1
See: “PatentCafe’s Patent Software Tools survey” (http://tinyurl.com/96mja)
[Consulted on August 11, 2005].
2
See for instance Paap, J. (2002). Using technical intelligence to drive innovation and
technical decisions. Workshop given at the Annual International SCIP conference in
Cincinati, USA.
3
It can be observed that the difference between CI and CTI does not occur only
- 22 - Software for Technological Patent Intelligence
in patent analysis. CTI also uses primary and secondary sources specializing in
recovering technological information.
4
See Assessing Competitive Intelligence Software by Bouthiller and Shearer (2003),
Software Report 2004-2005 published by Fuld&Company (2004) (http://
www.fuld.com/Products/ISR2004/HomePage.html) or How can We Determine which
Competitive Intelligence Software Is Most Effective? By Nikkel (2003, p.163). Full
references can be obtained at the end of the book (see page 151-152).
5
See summary table 1.
SECTION THREE
Methodology
- 24 -
- 25 -
3. METHODOLOGY
In order to answer the research questions posed above, two separate sections of
the study were developed, using and integrating them in a joint framework.
Both sections used the same framework for studying the application functions.
In other words, the same groups of functions were studied from both a
demand as well as a supply viewpoint (see the following section on this).
The study of supply was made separately - that is, it was “blind” - with
no knowledge of the results of the study on demand. In this way, we tried to
avoid any bias in the judgments made in both sections of the study.
The functions identified in this way (41 in total) were divided into 6 groups,
as shown in the following table1:
- 26 - Software for Technological Patent Intelligence
This section assesses all the characteristics relating to the process of information
collection and its automation.
The environment in which a typical user currently carries out his or her
work might include access to patent databases as well as access to other
bibliographical databases that are normally scientific and technological and
which usually complement one another.
In addition to this, these databases can be located in a local network (in a
private database for instance or in a commercial CDROM-based database) as
well as in a website, which means that changes must be implemented in the
program in order to allow access to each of these options.
Wherever the program included an interface for information searching in
a website, we also considered the option for saving the search strategy and for
programming its periodic implementation, since these are basic tasks for the
Technological Observation function.
Another very basic characteristic is the ability to import the results of the
searches carried out in any information source, normally in csv format (comma
separated values), in text format delimited by fields or in XML format.
Lastly, the ability to integrate other information relating to patents in
order to add to their value was also assessed. This information is normally
in the form of graphics or .pdf documents, but it can also be, for instance:
legal information which could increase or remove the value of a patent, or
economic information relating to a company or a technology.
This section covers a whole list of tasks all of which have in common the
fact that, when they are carried out, the information becomes far cleaner and
better organized and assessed, making subsequent analysis far easier and, in
addition, resulting in much firmer conclusions.
In patent analysis, it is important to define the information “unit” to be
analyzed. Generally speaking, analysts work with “patent families” which
group together in one single record all the documents generated from the
same priority number. The deletion of duplicate patents and the grouping of
patents by families is a task which should be carried out either prior to loading
the information into the software or once it has been loaded.
Automatic generation of indexes, the ability to easily generate new indexes
from elements contained in different information fields (for instance terms
Methodology - 29 -
In this section, we assess the basic abilities of the software to manage the
information accumulated (filters and advanced searches, classification of
results by different criteria, etc.).
Other more advanced abilities are also assessed, such as the generation of
automatic abstracts for each patent, the extraction of the most representative
concepts of each document, the automated “clustering” of the patents into
different categories or semantic searching. These abilities already have some
relation to text mining.
This section includes a list of the different tasks and functions which can be
automated in order to reinforce collaborative work. Firstly, the initial idea is
that each user should have his, or her, own information profile and receive
any alerts corresponding to said profile. From here, the ability of each user to
generate reports using predefined templates in his or her specialty is assessed.
In addition, the aim is to generate new knowledge among different
individuals by means of discussion and joint analysis of different
multidisciplinary issues dealing with said shared information. The possibilities
for group work considered were: the generation of polls, the generation of
forum discussion and the existence of a shared agenda for the work to be
carried out.
Several options for exporting information to standard formats, allowing
their use in other software, were also listed.
Methodology - 31 -
This section cites a group of functions that assess the ease with which
the software adapts to the requirements of different users (for instance:
different languages) with different rights (edition or creation, for instance),
simultaneous work or whether or not the information can be published in an
intranet.
The ability to generate system use statistics is included, so that an
assessment can be made as to how the use of the software is developing and in
short, as to how efficient it is.
The initial software selection was carried out using the following criteria:
These two questions establish a starting point for patent analysis software
demand. In other words, we wish to identify the potential needs of patent
users, without actually specifically defining the real current and future needs
of professional individuals.
Having carefully pondered the possibilities for studying user needs, we
came up with a simple and representative approach. In effect, the approach we
suggest in our study focuses on the “use” of the functions described in Table
1 of our poll.
In addition, we introduced a section in which we assess the degree of
- 34 - Software for Technological Patent Intelligence
The main focus of our study is the situation as regards software currently
available and whether this software lives up to user expectations. The
assessment of user needs and the definition of the future of the software or
the determination of the “value” attached to each function incorporated in
existing software and included in our sample are not the main focus of this
study.
Reality does not always match the wishes of users. Despite the fact that a
user may wish for a function and value it highly, as long as he has no software
program which offers said function, he will have to reply in the questionnaire
that he does not use this function. This gives rise to a simplification of the
questions.
3.3.1 Instrument
The assessment of the use of the 41 functions (items) divided into 6 groups
described in the questionnaire was carried out by requesting information on
the following aspects (see Annex 8.5 or page 26):
1. Function use - The degree of use was assessed using a 7-point Likert
scale.
2. Relative importance of functions - It was suggested to those
responding to the questionnaire that they use an evaluation system in
which they awarded 1 point to the factor which was least important to
them when compared with the other factors. This method allowed us
to define which groups of functions users consider most important6.
3.3.3 Sample
Methodology - 35 -
The invitation to participate in this initial study was sent by electronic mail.
A total of three invitations were sent:
of which was a total of 102 replies. This final call set a deadline at
July 31, 2005.
Despite the fact that a major increase was observed in the number of replies
obtained each time a new invitation was sent, we considered that three
invitations gave a sufficiently satisfactory result.
3.3.5 Results
The end results of the poll were 102 valid questionnaires. The following table
summarizes the results of the PIUG and EPO listing.
*******************************************
Footnotes
1
This characteristic is not a selection criteria but it did significantly reduce the
number of applications assessed.
2
The profile of those polled and of how the sample was chosen is described in later
sections.
3
This part of the poll is relatively simplified. We have taken a simple approach,
however, since an assessment of the relevant functions would have become excessively
complex otherwise.
4
The scale was as follows: “Not at all”, “Very little”, “Little”, “Sometimes”, “Often”,
“Almost every time”, “Always”. “Not Applicable (N/A)” was also added.
5
For further information on the provider used, see Surveymonkey.com (http://
Methodology - 37 -
www.surveymonkey.com).
6
Personal comunication; Juan Manel Batista (ESADE, Barcelona, Spain). To see an
application which uses this method, see Comai, A (2005) “Factores y Contingencias
en la Inteligencia Competitiva: Resultado en un estudio piloto,” PUZZLE - Revista
Hispana de la Inteligencia Competitiva, 4(18):12-15. (see http://www.revista-
puzzle.com/puzzle_sum_18.htm).
7
By way of an example, it can be observed that the total number of confirmations of
having read or rejections of the invitations we sent by electronic mail exceeded 230.
It should also be noted that these readings were made long after the questionnaire
expired.
8
We contacted both PIUG as well as EPO in order to obtain this information. The
replies, however, were negative. In other words, they had no information on the
subject. For PIUG, see “http://piug.org/list.html#Majordomo%20Commands%20-
%20How%20to%20Join%20the%20Discussion%20List” and for EPO see “http://
www.european-patent-office.org/mail.htm”.
9
See Annex 8.1
10
Estimation (see section Sample, page 35).
11
Approximation (see section Sample, page 35).
- 38 -
SECTION FOUR
Results of the study: demand, users
- 40 -
- 41 -
This chapter gives the results in detail of the poll carried out among patent
users.
4.1 Profession
There is some variety in the profession of those polled. The questionnaire
included three clusters or groups of activity: R+D Manager (6.3%), Librarian
(5.3%), and technicians (10.5%). These three activities accounted for less than
23% of the total number of individuals polled but constitute an important
minority. The professions identified as most common were “Patent specialist
or searcher” (18/102) “Patent attorney” (12/102) and “Copyright manager”
(7/102). Nevertheless, lawyers, patent experts and company directors also
completed the questionnaire (see Figure 1).
In the rest of the sample (43.3%) several sectors are represented, including
consumer goods, biotechnology, cosmetics, software design, as well as
consultants, the government and universities. The consultancy group contains
16 companies.1
4.3 Experience
We inferred from the sample, that the patent experts have notably extensive
experience in everything relating to their work. Despite the fact that we did
not specifically ask in the questionnaire whether their experience related
exclusively to patents or whether it was wider-ranging, we are inclined to
think that their experience in the former is very extensive. In fact, the vast
majority of the sample (75.5%) reports having 6 years experience and 53.1%,
over 11 years. Figure 3 gives a breakdown of the sample in terms of experience
in the field of patents.
These results support those obtained for our research, due to the fact
that extensive experience in the field of patent analysis is very relevant to an
adequate response to our questionnaire.
Of these functions, the one which is most used (frequent use)2 or which has
obtained the highest average rating is “Saving search strategies”. This function
stands out from the others due to the fact that around 30% of those polled
maintain that they always use it.
Other functions obtained similar average ratings, such as “Downloading
and linking of pdf documents”, “Launching simultaneous searches in
multiple databases” or “Downloading and integration of patent legal status”,
for instance, which are used with a frequency very close to “often”.
In a second group not far behind the first, we find less frequently used
functions. For instance, “Scheduling repetitive searches”, “Searching in
complementary technical / grey literature online databases”, “Searching in local
(intranet) databases”, “Downloading and integration of patent legal status”,
“Importing patent records from other software” are used “Sometimes”.
It should be observed, however, that the data for each function is somewhat
dispersed. That is, users make quite different use of the functions. A maximum
of 32 cases and a minimum of 5 users per type of use were recorded for all
functions. Figure 4 shows this distribution.
Sometimes
40
Often
Sometimes
Always
% 30
Often
Very little
Not at all
Always
Not at all
Very little
20
Little
Little
10
N/A
N/A
0
% 30
20
10
0
40
% 30
20
10
0
40
% 30
20
10
0
40
% 30
20
10
0
The results of the poll show that there is some difference in the way in which
these functions are used.
Firstly, it should be observed that the most frequently used functions are
“Automatic grouping of patent families” and “Automatic patent duplicate
detection and removal”. The use associated with these two functions is “often”
(4.88 and 4.73 respectively). In addition, 18 and 20 experts in each case stated
that they “always” use these functions.
On the other hand, the function which is least used is “Creation and
edition of taxonomies”. Half of all users of this function report using it “little”
(2.46).
Other functions, such as “Automatic generation of field indexes” (3.76),
“Evaluation of pertinence” (3.70), “Annotation of patents” (3.64) or “Linking
to other related documents” (3.60), for instance, have an average use ranging
between “sometimes” and “often”, whilst use of the remaining functions, such
as “Definition and building of additional indexes” (3.31), “Definition and
edition of patent groups” (3.29), or “Grouping and cleaning of index terms”
(3.26) appears to be closer to “sometimes”.
It should be emphasized, however, that although some functions obtained
a specific average rating, responses were also extremely wide ranging. In other
words, there are as many experts “always” using a specific function as those
“never” using that same function. There are clear differences in the use by
experts of function 3) “Automatic generation of field indexes”, shown by the
fact that a sizeable number of experts in all use groups appeared on the scale
of 7 points we established in the questionnaire. These findings can be seen in
figure 5.
Results of the Study: demand, users - 47 -
The results of the study show that the function which stands out most is “Full
text indexing/searching” of patents, having ascertained that experts use this
function “often”. It should be observed that almost 25% of experts (of the
total 82 who answered this question) always use this function.
A second group includes the remaining functions with relatively similar
ratings, an average use of between “little” and “sometimes). The function
in this group which is most used is the obtaining of “Automatic abstracts”
(3.73) and the least used is “Automatic classification of patents in pre-defined
categories”. The other functions are situated between these two extremes (see
Figure 6).
Not at all
Sometimes
Not at all
% 30
Sometimes
Always
Often
Very little
Very little
Often
Always
20
Little
Little
N/A
N/A
10
0
% 30
20
10
0
40
% 30
20
10
0
40
% 30
20
10
0
The two most used functions are “Rankings – Analysis of one field” and “Cite
analysis”, with (4.03 and 3.97)2 respectively, suggesting that they are used
“sometimes”.
These are followed by “Matrix or bar graphs – Two fields co-occurrence
analysis” which is used from “little” to “sometimes” (3.59) and “Network
relations analysis - Two fields co-occurrence analysis” which is used slightly
more than “little” (3.19).
Lastly, “Space or topographic representation of a patent collection – text
mining analysis” is seldom used, given that the majority of the responses we
received fell into the categories “very little” and “little” (2.70) (see Figure 7).
The functions above refer to the ability of the software to manage the
applications in such a way that user needs are met. It should be emphasized
that the overall results obtained in this section are among the lowest in the
poll.
As is shown in figure 9, user responses are fairly similar for all functions,
with an average use which is close to “sometimes”. Although the responses are
quite varied, the one which appears most frequently (from 24% to 38%) is
“never”. The function which stands out as being least used is “Multilanguage
interface”.
4.10. Importance
The importance attached to each of the functions included in this section is
a relevant estimate of how highly they are rated. This is how we distinguish
frequency of use from the relative value each user considers has been added to
his work by each group of functions.
Results of the Study: demand, users - 53 -
********************************************************
Footnotes
1
For instance: legal, copyright management consultants, analysis and assessment of the
financial risk in copyright, lawyers and consultants or IP consultants.
2
The scale used was: (1) Not at all, (2) Very little, (3) Little, (4) Sometimes (5) Often, (6)
Almost every time, (7) Always y (0) N/A.
- 56 - Software for Technological Patent Intelligence
3
Method suggested by Juan Manel Batista of ESADE Business School (Personal
Communication).
4
The questionnaire contains the following explanation for users interviewed: - NOTE: Assign
“1” to the least important group of methods/techniques and rate ALL the others groups
against it. 6 If you rate “1” it means that the group of methods are equivalent to the one
being compared with. If you rate 1.5 then it means that the group of methods are 50% more
important to the one being compared with. If you rate 2, that means that it is one time more
important or dobble and if you rate 3 it is two time more important and so on... -
5
The correction was carried out as follows: The lowest average rating was identified. In the
case of “Software Management” it is 1.40. Since the system takes 1 as the lowest rating, we
made the lowest rating the reference rating. That is, reducing 1.4 by 0.4. In this case, “Software
Management” became the reference and thus took on the rating 1. The ratings of other items
were reduced by the same amount in order to study the incremental ratings in terms of
importance, in accordance with the evaluation system used in the questionnaire.
SECTION FIVE
Comparison of Software: Supply
- 58 -
- 59 -
In this section the technical specifications for five of the softwares will be
evaluated in depth in this study. However, information will also be added on
a further ten softwares which have not been fully evaluated. The partial or
complete description of the programs is carried out using the format below:
The technical details of each program are presented in the following section
in alphabetical order.
- 60 -
- 61 -
Evaluation Table1
The reference table can either be entered by hand or by selecting a text file
with a list of the terms (one per line).
In both cases, the user can create as many crossed or reference tables (crossed
tables) as they wish.
- Forms: By selecting a field, the index in that field is displayed with all
the terms contained within, and their frequency. This index forms
the basis for carrying out other operations.
- Pairs: By selecting any two fields, an index with all combinations of
pairs of terms and their frequency of co-occurrence is shown.
4) Graphic Generation
Matheo Analyzer 3.0 can create the following graph types:
Histograms:
- Frequency histogram: this analyses the content of a field. The height
of each bar represents the number of patents corresponding to each
term. This is the most commonly used type.
- Range histogram: this analyses the frequency of the terms used in
a particular field. The height of each bar represents the number of
terms in this field with a determined frequency.
- Indexing depth histogram: this analyses the lists with a defined
number of terms in a particular field. It indicates to what extent
this field is wide-ranging (many terms used to define this field) or
extremely concentrated (very few terms used to define this field).
In all three cases a previous condition can be created (based on text, frequency
Comparison of Software: Supply - Matheo Analyzer v3.0 - 69 -
not be repeated. It also permits the selective loading of a specific list of patents
of interest. At any given moment and for any patent, Matheo Patent permits
the download of those fields which were not initially done so.
At any given moment, the user may carry out any of the above search
strategies for the project and download the new patents published concerning
that subject. Matheo Publisher detects the date on which this search was last
run and runs the search only from that date onwards. All new patents are
assigned an icon indicating that this information is pending revision.
4) Graphic Generation
Matheo Patent can create the following types of graphs:
Chart (histogram): Corresponds to the analysis of a field’s content. The
vertical axis. Represents the number of patents which correspond to each
term. The graph can easily be set to limit the minimum frequency of each
term.
Matriice (Matrix): Carries out analysis of matches in two fields. The fields
which can be used for this analysis are: inventor, applicant, year of priority, year
of publication, number of family members, group, 4-diit CPI classification,
complete IPC classification and ECLA classification.
The result is a matrix of cells in which the match number appears for
each one. The cells’ colour becomes more intense the higher the number of
matches.
Comparison of Software: Supply - Matheo Patent v7.1 - 81 -
Every element can have a link to another one in another field. In this case, they
form a “pair”. This pair has a reference number indicating its co-occurrence
(the number of patents in which the two terms appear).
The user can create a network from one field (for instance, to analyse the
relationship between inventors) or from two. In the latter case, the relationship
between companies and the groups of patents they have created are shown.
- 82 - Software for Technological Patent Intelligence
6) Management of Tool
The functions mentioned in this section refer to software accessed by various
users at once. Matheo Patent does not have these functions.
Comparison of Software: Supply - PatentLabII v1.41 - 83 -
PatentLab is a tool used for the statistical analysis and visualisation of registry
details from patents obtained through the Delphion patent search service.
This site offers the following collections of patents: PCT, European (request
& concession), North American (request and concession), German, Japanese
- 86 - Software for Technological Patent Intelligence
and Inpadoc.
5.3.2. Comments on the features studied
Thesaurus: PatentLabII does not allow the creation of lists of key words,
thesauruses, empty words, etc.
Text-mining: PatentLabII does not have text-mining technology. The title,
summary and enquiries can be loaded in PatentLabII but cannot be analysed.
User Classifications: There are four fields into which the user can download
information of his choice. However, there is no tool to ease this task or to
guarantee that the classifications are added without errors.
- 88 - Software for Technological Patent Intelligence
PatentLabII does not have specific fields for making notes, nor for the
evaluation of a patent’s importance. However, the “fields defined by the user”
can be used for this purpose. It is also not possible to include a link for each
patent to its corresponding .pdf document.
Comparison of Software: Supply - PatentLabII v1.41 - 89 -
4) Graphic Generation
PatentLabII can create graphics in two or three dimensions. The graphics in
two dimensions correspond with the statistical analysis of a field’s content
in which the height of each bar is proportional to the number of patents
corresponding to each term. By clicking twice on each element, the list of
patents corresponding to that term appears.
The three-dimensional graphics correspond to analysis of the co-
occurrences between two fields. The height of the bar is proportional to the
number of co-occurrences existing between each pair of terms. By clicking
twice on each cell, the list of patents corresponding to these terms appears.
- 90 - Software for Technological Patent Intelligence
6) Management of Tool
The functions mentioned in this section are of relevance only to those
programs designed for use by several people at once. PatentLabII does not
offer these functions.
Comparison of Software: Supply - PM Manager v1.4.0.3 - 93 -
PM Manager is able to load files obtained from WIPS Global and then,
classify, analyse and process information from patents by applying different
viewpoints. PM Manager’s focus is on giving the user the capacity to modify,
annotate and complete information by adding new personalized fields.
It then allows basic statistical analysis, advanced statistical analysis and
other analysis tailor-made for the personalized fields created by the user.
Thesaurus: PM Manager does not permit the creation of lists of key words,
thesaurus or empty words.
Text mining: PM Manager does not have text mining technology.
User Classifications: The software allows for the creation of a three-level
thematic classification which can be as extensive as required. In order to
facilitate this task, this structure appears when the user wishes to classify the
content of a patent.
Last but not least, with each patent a range of links with documents and
applications can be included.
4) Graphic Generation
MP Manager can create graphs in two or three dimensions. The two-
dimensional graphs correspond to the statistical analysis of a field’s content, in
Comparison of Software: Supply - PM Manager v1.4.0.3 - 101 -
PM Manager allows for the creation of matrices, using three fields in the
analysis including those created by a user. In this case, PM Manager permits
the running of a co-occurrence analysis in two fields. If this co-occurrence
exists, it shows the value of the third field. The usefulness of this analysis
may be very different depending on the type of content within the two fields
defined by the user.
via the web in order for others to visualize information from other terminals.
It can interact with MS Excel and MS Word.
Apart from the report on the “Key Information List”, it can generate files
in MS Excel from practically any analysis which has been carried out.
6) Management of Tool
The functions mentioned in this section are of relevance only to those
programs designed for use by several people at once. PM Manager does not
offer these functions.
- 104 - Software for Technological Patent Intelligence
Comparison of Software: Supply - Vantage Point v4.0 - 105 -
- Advanced users can build and edit import filters for a database
editing the parameters of an “importation engine”. Any register
obtained from any database can be considered as importable from
Vantage Point.
- Elementary users have an assistant at their disposal, which will guide
Comparison of Software: Supply - Vantage Point v4.0 - 109 -
listing the variants in the Priority field and clicking on an entry, all members
of that family can be seen.
Working with fields: Vantage Point is capable of carrying out a multitude
of operations using field content: copying, merging existing fields to create
new ones, or creating new indices from the contents of a group of terms.
Standardization and index-cleaning: Vantage Point has developed
several diffuse logic programs to minimize the work involved in cleaning
certain information fields. In particular, it is possible to indicate the system
which automatically groups inventors and patentees in line with previously
programd criteria. Furthermore, the user can manually check the indices and
group synonymous terms using the “drag and drop” tool.
Thesaurus: Lists of key words can be made, taken from the contents of one
or more fields. They can then be edited or combined. There are also lists of
empty words, used to clean the indices. These lists can be edited.
Text-mining: Vantage Point uses its NLP (Natural Language Processing)
technology to analyse any non-structured text (title, summary, full text) written
Comparison of Software: Supply - Vantage Point v4.0 - 111 -
in English with the purpose of extracting the most significant concepts. Firstly,
it divides the text into sentences and labels each word as noun, adjective or
verb. It then applies linguistic rules to identify “multiterms” composed of
various words. If the user wishes to analyse text in another language, it can
retrieve the text but cannot recognize verbs or adjectives. Thus, it cannot
identify the “multiterms” made up of several words.
- Carry out sequential searches for fields and create a group with the
results.
- Combine groups.
- Export groups to MS Excel.
- Export part of a matrix to MS Excel and create associated 3-D
graphs.
- Export selected fields from a group of registers to MS Word.
- Create a thesaurus from marked elements in a matrix.
- Detect the terms which have appeared for the first time year after
year and Export them to MS Excel.
4) Graphic Generation
Vantage Point can create 2-dimensional matrices or maps. The matrices may
be of various types:
6) Management of Tool
The functions outlined in this section are of relevance for multiple-user
software. Vantage Point does not have these functions.
Comparison of Software: Supply - Programs not evaluated - 115 -
Producer: I2
The Visual Space, Capital Park,
Fulbourne, Cambridge, CB1 5XH, United Kingdom
Tel: 01223-728600
Fax: 01223-728601
info@i2.co.uk
Website: http://www.i2.co.uk/anacubis/
Definition of software
Anacubis Desktop is an application which creates an intuitive visual
representation of information, showing all types of entities (people, companies,
patents, etc.) as icons and indicating their relationships through links.
A differentiating feature of Anacubis is that it permits the users to consult
various sources of information and combine them to instantaneously obtain a
single all-encompassing view.
Anacubis Desktop has a specific complement for the treatment of patents.
Its usefulness lies in that it allows the inexperienced user to carry out advanced
analysis by following the step-by-step instructions of a guide.
Main Features
The main analyses “Anacubis Desktop”, complemented by “Intellectual
Property Analysis Add-in”, can carry out are the following:
Analysis of citations
- A patentee’s citations: Creates an analysis of “who mentions whom”.
It can produce an analysis of all the patentees appearing in a
document or, alternatively, a restricted analysis of several patentees
receiving a minimum number of mentions.
- History of citations: Shows generations of citations, both forward
and backward, based on selected patents.
- Patent citations: Shows who cites or is cited by whom for one or
more key patents.
- 116 - Software for Technological Patent Intelligence
Analysis of Inventors
- Company inventors: Creates a vision about who invents in a
company.
- Common inventors: From a group of selected companies, it shows
those inventors who have or have had a connection with more than
one of that company.
- Research groups: For a company, this shows “who is working with
whom”.
Seasonal analysis
- Patent citations over time: Produces an analysis with all patents that
cite or are cited by one or more key patents in sequential order. Each
patent is linked to the request made for it.
- Citations from patentees over time. This produces a timeline which
illustrates the relationship of citations between one patentee and
others selected by the user.
Definition of software
Aureka is a wide-ranging program focusing on Industrial property management
within a company. It includes functions for searching for patent information,
options for their validation and evaluation, the capability of adding a user’s
own fields, advanced functions for analysing results (including text-mining)
cartographic representation of groups of patents, management of warnings
and functions for working in groups.
Main features
Since Aurigin (original manufacturer of Aureka) was taken over by
Micropatent, Aureka has become the application used by this company for the
management and analysis of information sourced from its patent databases.
At present, Micropatent offers access to the United States Patent Collection
(requests and concessions), European Patents (requests and concessions),
PCT Patents and patents from Germany, France, Great Britain and Japan.
Messaging
The collaboration or working groups functions are based on an e-mail
messaging system among all users of the system. There may be several user
categories (administrator, trainer, expert) and each has its own particular
functions. This system is used for distributing the results of the searches, filters
or analysis among the users.
Comparison of Software: Supply - Programs not evaluated - 119 -
Notes
Aureka allows each user to make notes on each patent and search notes for all
patents.
Alerts
Each user may define weekly or monthly alerts which can be sent to other
people or groups within the organization.
Technological maps
Aureka has software for the textual analysis of information contained within
a group of patents. The result is the generation of a topographical map in
which the highest peaks represent an accumulation of documents sharing a
particular concept.
The user can specify the fields whose text he wishes to analyse (his own
fields may be included) before the generation of the map. Then, one part of the
map can be selected in order to obtain a more detailed overview. Additionally,
a search for key words can be carried out and the program is able to find the
area of the map where the answers are concentrated.
These maps can be shared among selected groups of users so that they
can comment on and discuss specific areas of interest and reach shared
conclusions.
Citation “trees”
Aureka also has an analysis of citations in which the citing and cited patents
are visualized in their time sequence.
These analyses can be limited through certain key words or by demanding
that patents fulfil certain requirements (including those in own fields).
Aureka allows for colour-coding of patents dependent on, for example, the
patentee or the rate of inflation.
- 120 - Software for Technological Patent Intelligence
Definition of software
ClearForest Analytics is text-mining software which converts non-structured
information into structured and clearly related concepts and data. This allows
users to concentrate on their key competencies: analysing data and taking
decisions. The three objectives are:
Main features
ClearForest Analytics is part of the ClearForest platform which includes
ClearForest Tag and ClearForest Industry Modules. Once the labels for
ClearForest Tag are entered, ClearForest Analytics extracts relevant concepts
from the tagged terms, permits navigation by concept and detects trends
empirically.
Patent Analysis is a module specifically for professionals working in
the field of Industrial Property. This module has a “rule book” concerning
patents, to extract specific information from patent databases.
It is designed to shorten the time required to take decisions concerning
technology, to carry out more wide-ranging competitive analysis, to obtain
an improved overview of competitors’ R&D activities and to reduce research
costs. By examining patent databases, this module can discover the names of
- 122 - Software for Technological Patent Intelligence
Definition of software
Derwent Analytics is a software program which is not sold independently but
offered in combination with a subscription to the Derwent World Patents
Index database. This has coverage in 41 countries and patent organizations.
It offers exclusive fields and very high-quality content. The software itself is a
personalized version of Vantage Point (see earlier section) which includes the
programming of several specific macros to fully utilize the database.
Therefore, Derwent Analytics can be defined as text-mining software
specializing in the analysis of registers obtained from the World Patents
Index. Its aim is to analyse these registers to gain new perspectives on the
most important concepts they contain.
Main features
Derwent Analytics does not contain modules to connect to and use databases.
Its starting point is the importation of a list of registers obtained from various
hosts where Derwent WPI is sold by way of previously programd importation
modules.
Index cleaning
Inventors and Patentees can be automatically grouped, even though they may
show small variations.
- 124 - Software for Technological Patent Intelligence
Thesaurus
Lists of key words can be created from the contents of one or more fields.
They may then be edited or combined. The program also has lists of empty
words which can be used to clean the indices. These lists are also editable.
Graphics generation
Derwent Analytics can create 2-dimensional lists, matrices or maps.
Data analysis
Derwent Analytics has 6 specifically programd macros for DWPI, so a variety
of tasks and analyses are automated:
Definition of Software
Goldfire Innovator uses patent analysis within a structured system aimed at
improving the solving of invention problems.
It facilitates the identification of problems for users and capacitates them
to solve the problem and generate solutions. The software:
methodically explores and validates system designs which are more efficient,
cheaper more competitive and of higher quality.
Main features
Goldfire Innovator has three modules:
Definition of software
OmniViz is geared towards “visual intelligence”. That is, the visualization and
analysis of large volumes of structured and unstructured information as an aid
to decision-making. Its purpose is to offer an overall vision which can lead to
rapid identification and interpretation of the most relevant details.
It can analyse numerical, categorical and complete text data (including
patents), all within the same visual layout. It is also able to support the analysis
of chemical structures and genome sequences. This allows the inclusion of all
relevant information when taking decisions.
OmniViz is capable of integrating the analysis of experimental data with
that of scientific literature, patents and marketing data sourced from press
releases. Its fields of application are research, development, testing, control of
processes, marketing, finance or legal information.
Main features
The software is suitable for use as much by beginners or elementary users as
well as experts.
It can import data from any source or format. As well as accepting the
automatic importation of common data types, it also has an interface to create
reusable instructions to analyse other data formats. Additionally, it permits
the creation of tailor-made data importers.
Analysis
The program is flexible with different kinds of analysis. OmniViz can carry
out simple analysis for inexperienced users, those who are not experts in data
processing or in the standardization of data analysis. On the other hand, it
is capable of controlling all analytical parameters and offers a wide range of
Comparison of Software: Supply - Programs not evaluated - 129 -
advanced tools for the expert user to analyse any problem. Furthermore, new
methods of clustering or grouping data can be added.
Integration
XML can be used to define which data to use, how to carry out analysis and
which visualisation to use. OmniViz can be executed from other interfaces
and therefore, it can be used by those who do not necessarily want or require
advanced parameters. Automated analyses allow end users to see and evaluate
their data rapidly, while still providing a platform which can be exploited by
expert users to carry out further analysis.
OmniViz can share its data with other applications. It may both receive
data and send it to other applications via .xml format. Moreover, data can be
sent in personalized format for other applications.
OmniViz Platinum: Contains six interactive visualizations, designed to
respond to fundamental questions concerning the relationships between
registers and how certain attributes are distributed throughout a group of
registers. The opportunity to analyse and visualize multiple data permits
integrated analysis.
It also contains questioning tools, including dynamic questioning, three-
dimensional graphic tools, various sophisticated statistics packages and many
other unique features.
OmniViz Titanium: This package integrates further advanced modules
such as the Barnard Chemical Information (BCI), the Ward clustering
algorhythm or the Stanford SAM analytical algorhythm. These make possible
the direct identification and grouping of chemical compounds from large
volumes of data, as well as the identification of changes in large amounts of
data.
- 130 - Software for Technological Patent Intelligence
Definition of software
PatList is used for the reformatting and statistical analysis of information on
patents extracted from certain databases.
Main features
Importation: Can import data from bibliographic reference lists taken from
Derwent WPI or from IFI/Claims in the following hosts – Dialog, STN or
Questel-Orbit.
Analysis
This program permits different types of statistical analysis:
The table contents can be exported in csv format in order to visualize or work
on them in MS Excel.
- 132 - Software for Technological Patent Intelligence
Definition of software
STN AnaVist is software used for visualization and interactive analysis which
offers a wide variety of ways of analysing the results of searches in scientific
and patent databases. It also visualizes trends in a research field. STN AnaVist
can help to solve complex questions as well as providing information which
can be used to take faster and better business decisions. STN AnaVist allows
the user to:
Main features
STN AnaVist offers a unique combination of functions for gathering,
analysing and interpreting information obtained from scientific and STN
patent databases. Its main functions are:
Definition of software
Tetralogie is software aimed at analysis and graphic representation of large
amounts of registers from databases. Its principal field of application is in
Technological Monitoring and Economic Intelligence. Tetralogie is a tool
for:
Main features
Tetralogie accesses its material from lists of bibliographical registers of
databases available on-line, CD-ROMs or from any other source. These
registers may be factual or may have fields in complete text. Tetralogie is a
powerful tool in statistical, exploratory analysis and interactive cartography
methods.
After a preliminary phase preparing the data, it carries out analysis of the
information provided as a result of the appearance of focuses and knowledge
which cannot be obtained through reading the information sequentially, such
as:
- Identification of the main protagonists and their workplaces.
- Evaluation of their importance, their relationships and mobility in
time.
- Emergence and evolution of ideas and concepts.
- 136 - Software for Technological Patent Intelligence
6. General Conclusion
2. Filtering and Value Adding: In the large majority of cases, this group
of functions fulfils the requirements of users very satisfactorily. In
- 140 - Software for Technological Patent Intelligence
other words, the range of functions available slightly exceeds the use
made of them.
Supply Demand
Relative Importance
(Average)
1.- Searching and 1,51 4,42 2,23
Downloading
Ability to search in a set of 7,04 1,4 1,75 4,46
online patent databases
MA
MP
PL
PM
VP
Adding
Automatic duplicate 7,0 7,0 7,0 7,0 5,6 4,7
detection and removal
Automatic grouping of 7,0 7,0 2,8 4,9
patent families
Automatic generation of 7,0 7,0 7,0 7,0 5,6 3,8
field indexes
Ability to define and build 5,0 7,0 2,8 3,3
new indexes
Wizard for grouping and 2,8 4,2 2,8 7,0 3,4 3,3
cleaning terms of indexes
Patent pertinence (user filled 7,0 7,0 7,0 4,2 3,7
field)
Annotation of patents (user 7,0 7,0 7,0 7,0 5,6 3,6
filled field)
Ability to define and edit 7,0 7,0 7,0 5,6 5,3 3,3
patent groups
Links to other related 1,4 7,0 1,4 2,0 3,6
documents
Taxonomies creation and 7,0 1,4 1,1 2,5
edition
MP
PL
PM
VP
Exploitation
Automatic extraction of 1,4 1,4 0,6 3,7
main keywords from patents
Automatic abstracts 3,7
MA
MP
PL
PM
VP
Cite Analysis (cited and
citing patents in relation to a
known patent) 4,0
Rankings - Analysis of one 7,0 7,0 7,0 7,0 7,0 7,0
field. 4,0
Matrix or Bar graphs – Two 7,0 7,0 7,0 7,0 7,0 7,0
field’s co-occurrence analysis. 3,6
Network relations analysis 7,0 7,0 4,2 3,6
– Two fields co-occurrence
analysis 3,2
Space or topographic 1,4 0,3
representation of a patent
collection – text mining
analysis 2,7
Ability to use local databases 2,8 1,4 1,7
to integrate new data and
complete the patent analysis
MP
PL
PM
VP
Workgroup
MA
MP
PL
PM
VP
Management of users 3,8
access rights
Management of Document 3,6
collections access rights
Simultaneous multi-user 3,8
access and edition
Customization of access 3,4
and search interface
Multilanguage interface 2,6
The table above (table 16) summarizes the conclusions drawn about the
previous groups of functions, with those functions not satisfying the demand
marked in red and those which (to a greater or lesser extent) do marked in
green. We can conclude that only two features fulfil the average expectations
of program users.
The results from table 16 are summarized in table 17. Figure 56 offers
a different visualization of the results obtained in table 17. The groups of
factors have been compared with the average use of those factors. It can be
seen that, in the majority of cases, there is a certain positive relationship
between the frequency of use of those functions and the value given to them.
This relationship is similar to that discovered in other studies using the same
methodology (for example, see Comai, 2005).
Conclusion and discussion - 145 -
Average Value
Average Value
Importance
of Demand
of Supply
Relative
Group of Functions
We believe that the results of this test show that there is still work to be done
by the companies producing these softwares. Manufacturers need to increase
- 146 - Software for Technological Patent Intelligence
1 - SEARCHING AND
DOWNLOADING (4.4/2.2)
Relative Importance
their capacities in the areas underlined above in order to increase the added
value of these programs. The main improvements in the softwares should be
centred on the functions most commonly accessed by users, adopting the
priorities suggested in table 17 as “Searching and Downloading”, “Local
Analysis and Exploitation”, “Management of Tool” and “Dissemination and
Workgroup”.
******************************************************
Footnotes
1
This value was calculated by taking the average of the adjusted values for all the functions in
a class.
2
The value reflects the average of the functions classified in a category.
3
The value comes from table 11 (page 55).
4
The value reflects the individual tables for each program in the study, presented in chapter 5
which have an original scale of 0 (no function in the program) to 5 (the program fully covers
this function). This scale has been adjusted to 0 to 7 in order for supply and demand table
to be directly comparable. See comparison tables and survey used in the indices. If the box
indicates no mark and is not marked this means that the program does not fulfil the function
in question.
5
This value was obtained by taking the average of the associated values of the five programs for
the same function.
6
This value was calculated by taking the average of users’ responses to the programs in the
study, obtained in the study of use of functions (see chapter 4).
- 148 -
SECTION SEVEN
References and Authors
- 150 -
- 151 -
7.1 REFERENCES
Dou, H.; Levillé, V.; Manullang, S. and J. M. Dou, (2005). “Patent Analysis
for Competitive Technicall Intelligence and Innovative Thinking,”
Data Science Journal, 4:209-237.
7.2 AUTHORS
Alessandro Comai
8. ANNEXES
Dear Sir,
This study seeks to understand the usage of patents analysis software and
make a comparison between several softwares. The benchmark evaluation
will be done for each software against a defined list of characteristics which
will be validated from a survey we are doing with practitioners. This survey
will define the relevance and importance of different methods or techniques
within the profession.
If this were not possible, we would appreciate if you could send us the
documentation and other relevant material that would allow us to study the
software. In this specific case the conclusion of the software benchmark will
be classified in a special group of products that have not been tested.
patent analysis software programs, and the results will be published at the
beginning of November 2005.
Sincerely,
Dear Sir,
http://www.surveymonkey.com/s.asp?u=72931084818
The deadline will be tJuly 15, 2005. As a small token of our appreciation, we
will provide a summary of our findings for each stage of the research. The
full report, which provides several patent analysis software programs, and the
results will be published at the beginning of November 2005.
Sincerely,
Alessandro Comai
Director of PUZZLE and Ph.D. candidate (ESADE Business School)
alessandro.comai@revista-puzzle.com
- 160 - Software for Technological Patent Intelligence
Dear Sir,
I am sending a 3rd. and last call for participating in our study. This study
seeks to understand the usage of patent analysis software.
After the 2nd call we have added 19 valid questionnaires to the initial 34.
Therefore, we have achieved 53 valid responses at the moment. We would
very much appreciate it if you could complete the questionnaire online at:
http://www.surveymonkey.com/s.asp?u=72931084818
The deadline will be July 31, 2005. As a small token of our appreciation, we
will provide a summary of our findings for each stage of the research.
Sincerely,
Alessandro Comai
Director of PUZZLE and Ph.D. candidate (ESADE Business School)
alessandro.comai@revista-puzzle.com
Annexes - 161 -
Estimado/a Sr/a,
http://www.surveymonkey.com/s.asp?u=648091192001
Sinceramente,
Alessandro Comai
Director de PUZZLE y doctorando (ESADE)
alessandro.comai@revista-puzzle.com
- 162 - Software for Technological Patent Intelligence
Joaquín Tena
Co-Director de PUZZLE
Joaquin.tena@upf.edu
The questionnaire used in the poll was a form consisting of a total of 10 pages
in HTML format, accessible only through the internet (see example screens
below).
The entire contents of the questionnaire are shown below although the format
of the document is slightly different to the one which appears in the internet.
- 166 - Software for Technological Patent Intelligence
1. Purpose
This study seeks to understand the usage of patents analysis software. We are
particularly interested in defining the relevance and importance of different
methods or techniques with respect to your profession.
The questionnaire is divided in two parts. The first will explore what methods
or techniques are most commonly used. The second one distinguishes the
importance of the methods.
2. Demographic Information
Confidentiality: We assure you that your identity and that of your company
will be treated as strictly confidential. The information you provide will not
be shared with any other person and all references to your company’s data will
be blinded in any report resulting from this research.
If you are interested in receiving an executive copy of the full study, please fill
the following information:
4. Name:
5. Surname:
6. e-mail
Annexes - 167 -
3. Search
7. Please rate the extent to which you use the following methods/techniques
to SEARCH patents?
Sometimes
every time
Very little
Not at all
Almost
Always
Often
Little
N/A
1. Search in
complementary
technical/grey
literature online
databases
2. Search in local
(intranet) databases
3. Import patent
records from other
software
4. Launch simultaneous
searches in multiple
databases
5. Save search
strategies
6. Schedule repetitive
searches
7. Download and
integrate of patent
legal status
8. Download and
integrate of graphics
9. Download and link
of pdf documents
- 168 - Software for Technological Patent Intelligence
8. Please rate the extent to which you use the following methods/techniques
to FILTER & VALUE ADD patents?
Sometimes
every time
Very little
Not at all
Almost
Always
Often
Little
N/A
1. Automatic patent
duplicate detection
and removal
2. Automatic grouping
of patent families
3. Automatic
generation of field
indexes
4. Definition and
building of additional
indexes
5. Group and clean of
index terms
6. Evaluation of
pertinence (user filled
field)
7. Annotation of
patents (user filled
field)
8. Definition and
edition of patent
groups
9. Link to other
related documents
10. Creation and
edition of Taxonomies
Annexes - 169 -
9. Please rate the extent to which you use the following methods/techniques
to ANALYSE & EXPLOIT patents?
Sometimes
every time
Very little
Not at all
Almost
Always
Often
Little
N/A
1. Automatic
extraction of main
keywords from
patents
2. Automatic abstracts
3.Automatic clustering
of patents
4. Automatic
classification of
patents in pre-defined
categories
5. Full text indexing/
searching
6. Semantic indexing/
searching
7. Ability to use local
databases to integrate
new data and
complete the patent
analysis
- 170 - Software for Technological Patent Intelligence
6. Graphic Generation
10. Please rate the extent to which you use the following methods/techniques
to GENERATE GRAPHICS from patents?
Sometimes
every time
Very little
Not at all
Almost
Always
Often
Little
N/A
1. Cite Analysis (cited
and citing patents in
relation to a known
patent)
2. Rankings - Analysis
of one field.
3. Matrix or Bar
graphs – Two fields
co-occurrence
analysis.
4. Network relations
analysis – Two fields
co-occurrence
analysis
5. Space or
topographic
representation of
a patent collection
– text mining analysis
Annexes - 171 -
11. Please rate the extent to which you use the following methods/techniques
to DISSEMINATE patents?
Almost every
Sometimes
Very little
Not at all
Always
Often
Little
time
N/A
1. Publish the contents
in the intranet /
internet
2. Customised alerts
3. Alerts with changes
on the legal status
4. Automatic reports
using templates
5. Export all the fields
(csv, xml, etc)
6. Link a patent to
a poll with a key
question
7. Link a patent to
a forum and begin a
discussion
8. Link a patent to
an event in a shared
agenda
- 172 - Software for Technological Patent Intelligence
8. Management of Tool
12. Please rate the extent to which you use the following characteristics to
MANAGE Patent SOFTWARE?
Sometimes
every time
Very little
Not at all
Almost
Always
Often
Little
N/A
1. Management of
users access rights
2. Management of
Document collections
access rights
3. Simultaneous
multi-user access and
edition
4. Customisation of
Access and search
interface
5. Multilanguage
interface
6. System utilisation
statistics
Annexes - 173 -
9. Please rate how important you think the following group of methods/
techniques for analyzing patents is
Relative Importance:
N /A
1.00
1.25
1.50
1.75
2.00
2.25
2.50
2.75
3.00
1. Search
5. Dissemination and
Workgroup
6. Software
Management
- 174 - Software for Technological Patent Intelligence
10. Conclusion
Thank you for completing this Survey. We appreciate your input. If you have
any questions, please send an e-mail to:
alessandro.comai@revista-puzzle.com
Annexes - 175 -