U.S. patent application number 12/352827 was filed with the patent office on 2010-05-13 for analytic measurement of online social media content.
This patent application is currently assigned to Buzzient, Inc.. Invention is credited to Andreas Goeldi.
Application Number | 20100119053 12/352827 |
Document ID | / |
Family ID | 42165224 |
Filed Date | 2010-05-13 |
United States Patent
Application |
20100119053 |
Kind Code |
A1 |
Goeldi; Andreas |
May 13, 2010 |
ANALYTIC MEASUREMENT OF ONLINE SOCIAL MEDIA CONTENT
Abstract
A social media analytics platform including methods, apparatuses
and computer-readable media for providing analytic measurements of
online social media content by harvesting and aggregating
unstructured qualitative online social media conversations relevant
to subject matter of interest in a category from one or more online
social media sources, quantifying the aggregated online social
media conversations, and providing actionable information based on
the quantified aggregated online social media conversations, the
actionable information including sentiment expressed among online
social media participants concerning subject matter of interest in
the category.
Inventors: |
Goeldi; Andreas; (Cambridge,
MA) |
Correspondence
Address: |
PERKINS COIE LLP
P.O. BOX 1208
SEATTLE
WA
98111-1208
US
|
Assignee: |
Buzzient, Inc.
Cambridge
MA
|
Family ID: |
42165224 |
Appl. No.: |
12/352827 |
Filed: |
January 13, 2009 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
61114445 |
Nov 13, 2008 |
|
|
|
Current U.S.
Class: |
379/265.09 ;
704/9; 705/14.49; 705/14.73; 705/319; 707/E17.108 |
Current CPC
Class: |
G06Q 30/0251 20130101;
G06Q 30/00 20130101; G06Q 30/0263 20130101; G06Q 50/01 20130101;
G06Q 30/0277 20130101; G06Q 10/00 20130101 |
Class at
Publication: |
379/265.09 ;
705/7; 705/10; 704/9; 707/E17.108; 705/14.49; 705/14.73;
705/319 |
International
Class: |
H04M 3/00 20060101
H04M003/00; G06F 17/30 20060101 G06F017/30; G06Q 30/00 20060101
G06Q030/00; G06Q 99/00 20060101 G06Q099/00; G06Q 10/00 20060101
G06Q010/00; G06F 17/27 20060101 G06F017/27 |
Claims
1. A method comprising: harvesting and aggregating unstructured
qualitative online social media conversations relevant to subject
matter of interest in a category from one or more online social
media sources; quantifying the aggregated online social media
conversations to obtain structured analytic measurements of the
online social media conversations including analytic measurements
of sentiment expressed among online social media participants
concerning the subject matter of interest in the category; and
providing actionable information based on the analytic measurements
of the online social media conversations.
2. The method of claim 1, wherein the actionable information is
retrieved based on real-time measurements of the sentiment
expressed by the online social media participants concerning the
subject matter of interest and historical data representing
quantitative measurements of the sentiment concerning the subject
matter of interest in the past.
3. The method of claim 1, wherein the unstructured qualitative
online social media conversations include messages posted to online
social media websites, the messages including one or more of:
Internet messages; social media postings; online dialog; blogging;
interactions between customers and companies; call center logs;
emails; online mail and fax communications; call center records;
online purchasing information; online warranty claims; and other
online traffic.
4. The method of claim 1, wherein the one or more online social
media sources includes one or more of: blogs and sub-blogs; online
discussion forums; social networks; wiki sites; online reviews on
e-commerce sites; video websites; micro-blogging services; call
centers; websites including websites of companies; and other
sources of online social media conversations.
5. The method of claim 1, further comprising displaying the
actionable information in a graphical user interface.
6. The method of claim 1, further comprising: performing sentiment
rating processing on structured analytical measurements of the
online social media conversations referring to the subject matter
of interest to determine a sentiment rating for each of the online
social media conversations; and assigning the sentiment rating to
each of the online social media conversations.
7. The method of claim 6, wherein determining the sentiment rating
includes: identifying terms or phrases of interest associated with
the subject matter of interest in each of the online social media
conversations; searching in a set of closest N words from the terms
or phrases of interest for keywords expressing sentiment about the
terms or phrases of interest; assigning a probability value to each
of the keywords, the probability value indicating the probability
that the keyword suggests something positive or negative about the
terms or phrases of interest; assigning each occurrence of the
terms or phrases of interest with a sentiment score based on the
keywords in the set of closest N words from the terms or phrases of
interest; and adding up the sentiment score assigned to each of the
terms or phrases of interest in each social media conversation to
obtain a sentiment rating concerning the subject matter of
interest.
8. The method of claim 7, wherein the sentiment score is based on
one or more of: how many times each occurrence of the terms or
phrases of interest appears in the social media conversation;
number of keywords expressing sentiment about the terms or phrases
of interest in the set of closest words; whether each keyword
reflects a positive, negative or neutral sentiment about the
subject matter of interest; and relevance of the keywords
expressing sentiment about the terms or phrases of interest.
9. The method of claim 8, wherein the relevance of the keywords is
determined by taking into account one or more of: linguistic
modifiers of the keywords expressing sentiment about the terms or
phrases of interest including one or more of negations,
comparatives, and enumerations; and proximity of the keywords to
the terms or phrases of interest in the online social media
conversation.
10. The method of claim 9, further comprising taking into account
online social media author and website influence in classifying the
sentiment of each online social media conversation.
11. The method of claim 1, further comprising one or more of:
calculating how the sentiment concerning the subject matter of
interest trends over time; calculating how the sentiment concerning
the subject matter of interest varies by online source or group of
sources; and calculating how the sentiment concerning the subject
matter of interest concurrently trends over time and varies by
online source or group of sources.
12. The method of claim 1, further comprising determining an
overall volume of the online social media conversations referring
to the subject matter of interest by adding up a number of
occurrences of the subject matter of interest in online social
media conversations per unit of time.
13. The method of claim 12, further comprising determining how the
overall volume of online social conversations referring to the
subject matter of interest trends over time.
14. The method of claim 1, further comprising determining a share
of online voice acquired by the subject matter of interest with
respect to other online social media subject matter.
15. The method of claim 14, wherein the other online social media
subject matter includes subject matter associated with competitors
in the category.
16. The method of claim 15, further comprising calculating one or
more of: how the share of online voice acquired by the subject
matter of interest trends over time; and how the share of online
voice acquired by the subject matter of interest trends over time
with respect to the subject matter of the competitors in the
category.
17. The method of claim 1, further comprising performing text edge
processing on the online social media conversations to determine
frequency of occurrence of one or more topics in conjunction with
the subject matter of interest and relatedness of the one or more
topics to the subject matter of interest.
18. The method of claim 17, wherein the text edge processing
comprises: splitting up each sentence in the online social media
conversations into individual words and tuples of adjacent words;
identifying words or tuples of interest associated with the subject
matter of interest; identifying relationships between the words or
tuples of interest and each other word and tuple in the sentence as
an instance of an edge; and adding up each of the instances.
19. The method of claim 18, further comprising determining how the
frequency of occurrence of the one or more topics in conjunction
with the subject matter of interest trends over time.
20. The method of claim 14, further comprising determining an
overall advocacy of the subject matter of interest in the online
social media conversations for the category based on: the sentiment
rating of the subject matter of interest; number of advocates of
the subject matter of interest, the advocates of the subject matter
of interest including online social media authors having most
positive sentiment for the subject matter of interest in the
category; and the share of online voice acquired by the subject
matter of interest with respect to the subject matter associated
with the competitors in the category.
21. The method of claim 1, further comprising utilizing the
actionable information in conjunction with traditional research and
measurements including: quantitative and qualitative market
research; paid media tracking; and traditional website analytics;
sales information; public relations information; advertising
information; investor relations; brand management; and product
development information.
22. The method of claim 1, further comprising identifying and
actively engaging influential authors of online social media based
on the actionable information.
23. The method of claim 1, further comprising identifying and
engaging representatives of influential online social media
websites.
24. The method of claim 1, wherein the actionable information
includes real-time alerts providing early warnings about customer
service or quality issues.
25. A method for operating an enhanced customer service call center
comprising: receiving a customer service call at a call center from
a customer concerning a product or service; extracting information
from the customer service call; and retrieving actionable
information stored in a database relating to the customer service
call, the actionable information generated by harvesting,
aggregating and quantifying unstructured qualitative online social
media conversations from one or more online social media sources;
and providing enhanced customer service to the customer based on
the actionable information.
26. The method of claim 25, further comprising taking action based
on the actionable information including one or more: recommending
or performing an action based on the actionable information
relevant to the customer service call; and providing guidance to a
call-center operator on how to respond to the customer service call
based on the actionable information.
27. The method of claim 26, wherein the guidance to the call-center
operator includes how to deal with ongoing issues concerning the
product or service.
28. The method of claim 26, wherein the guidance to the call-center
operator includes identifying new issues concerning the product or
service.
29. The method of claim 28, further comprising providing crisis
prevention and management based on the real-time alerts.
30. The method of claim 29, wherein the real-time alerts are based
on user-configurable events including one or more: abnormally
positive online social media conversations about the subject matter
of interest; abnormally negative online social media conversations
about the subject matter of interest; changes in the sentiment
expressed among online social media participants concerning the
subject matter of interest above or below a pre-selected threshold;
abnormally high volume of online social media posts concerning the
subject matter of interest; abnormally low volume of online social
media posts concerning the subject matter of interest; social media
posts by certain authors; social media posts to certain websites;
and social media posts containing certain keywords.
31. A method of enhancing targeted advertising campaigns
comprising: retrieving actionable information stored in a database
relating to a product or service to be advertised, the actionable
information generated by harvesting, aggregating and quantifying
unstructured qualitative online social media conversations from one
or more online social media sources; and targeting advertising
campaigns based on the actionable information.
32. The method of claim 31, wherein the actionable information
includes one or more of: industry or sector corresponding to the
product or service to be advertised; sentiment expressed by online
social media participants about the product or service in the
online social media conversations; trends in the sentiment
expressed by the online social media participants about the product
or service; and alerts concerning the product or service.
33. The method of claim 31, wherein the actionable information
includes one or more of: most relevant websites, wherein the most
relevant websites include one or more of websites most likely to be
reached in online searches for information relating to the product
or service and websites where high-affinity social media
participants are exchanging opinions and making purchasing
decisions; and most relevant advertisement networks for advertising
the product or service.
34. The method of claim 31, wherein the actionable information
provides one or more of: perception of the product or service among
the online social media participants to be advertised; trend
recognition of sentiment expressed by the online social media
participants concerning the product or service; opportunity
identification; and competitor monitoring.
35. The method of claim 31, further comprising observing results in
the marketplace for the product or service based on the targeted
advertising campaign.
36. A method of providing enhanced marketing research comprising:
retrieving quantitative actionable information stored in a database
relating to a product or service to be marketed, the actionable
information generated by harvesting, aggregating and quantifying
unstructured qualitative online social media conversations from one
or more online social media sources; and providing enhanced
marketing research based on the actionable information.
37. The method of claim 36, wherein the actionable information is
made widely available inside an organization using an interface to
push the actionable information to everyone inside the
organization.
38. The method of claim 36, wherein the actionable information
enables marketing staff to establish an overall sense of the voice
of their customers and to make informed decisions at a customer
level.
39. The method of claim 36, further comprising providing
statistical analysis of the online social media conversations to
gain insight into social behavior of potential customers to inform
enhanced marketing decisions.
40. A method for enhanced product development comprising:
retrieving quantitative actionable information stored in a database
relating to one or more products, the actionable information
generated by harvesting, aggregating and quantifying unstructured
qualitative online social media conversations from one or more
online social media sources; and developing new products based on
the actionable information.
41. The method of claim 40, wherein the actionable information
includes sentiment expressed among online social media participants
about the one or more products.
42. The method of claim 41, wherein the actionable information
further includes sentiment expressed among the online social media
participants about one or more features of the one or more
products.
43. The method of claim 40, wherein the retrieved actionable
information includes information relating to one or more of
competitors' products, competitors' pricing, competitors'
advertising and competitors' sales strategies.
44. The method of claim 40, further comprising providing
statistical analysis on the online social media conversations to
gain increased knowledge of competitors based on the actionable
information and develop new products in response to the actionable
information.
45. A method for enhancing the gathering of opinion polls
comprising: retrieving quantitative structured actionable
information stored in a database relating to the subject matter of
an opinion poll, the actionable information generated by
harvesting, aggregating and quantifying unstructured qualitative
online social media conversations from one or more online social
media sources; and ascertaining opinion among online social media
participants concerning the subject matter of the opinion poll
based on the actionable information.
46. An method for providing enhanced national defense intelligence
comprising: retrieving quantitative structured actionable
information stored in a database related to potential violent
terrorist activity, the actionable information based on harvested,
aggregated and quantified unstructured qualitative online social
media conversations from one or more online social media sources;
and identifying potential violent terrorist activity in the online
social media conversations based on the actionable information.
47. The method of claim 46, further comprising identifying
previously unknown online social media sources used by groups to
engage in violent terrorist activity based on the actionable
information.
48. The method of claim 46, further comprising: monitoring keywords
associated with violent terrorist activity in the unstructured
qualitative online social media conversations; performing text edge
processing on the keywords associated with violent terrorist
activity; and identifying previously unknown keywords as associated
with violent terrorist activity based on performing the text edge
processing.
49. The method of claim 48, further comprising identifying keywords
and phrases that appear with greatest frequency in the online
social media conversations in conjunction with keywords and phrases
identified as associated with violent terrorist activity.
50. The method of claim 46, further comprising predicting
preparation, planning, and execution of violent activities based on
actionable information.
51. The method of claim 46, wherein the violent terrorist
activities include manufacture or distributing improvised explosive
devices (IEDs).
52. An article of manufacture comprising: a computer-readable
storage medium providing instructions which, when executed by a
computer, cause the computer to perform a method of providing
actionable information, the instructions comprising: instructions
to harvest and aggregating unstructured qualitative online social
media conversations relevant to subject matter of interest in a
category from one or more online social media sources; instructions
to quantify the aggregated online social media conversations to
obtain structured analytic measurements of the online social media
conversations including analytic measurements of sentiment
expressed among online social media participants concerning the
subject matter of interest in the category; and instructions to
provide actionable information based on the analytic measurements
of the online social media conversations.
53. The article of manufacture of claim 52, further comprising
instructions to display the actionable information in a graphical
user interface.
54. The article of manufacture of claim 52, further comprising:
instructions to perform sentiment rating processing on the
structured analytical measurements of the online social media
conversations referring to the subject matter of interest; and
instructions to assign a sentiment rating to each of the online
social media conversations.
55. The article of manufacture of claim 54, further comprising
instructions to determine the sentiment rating including:
instructions to identify terms or phrases of interest associated
with the subject matter of interest in each of the online social
media conversations; instructions to search in a set of closest N
words from the terms or phrases of interest for keywords expressing
sentiment about the terms or phrases of interest; instructions to
assign a probability value to each of the keywords, the probability
value indicating the probability that the keyword suggests
something positive or negative about the terms or phrases of
interest; instructions to assign each occurrence of the terms or
phrases of interest with a sentiment score based on the keywords in
the set of closest N words from the terms or phrases of interest;
and instructions to add up the sentiment score assigned to each of
the terms or phrases of interest in each social media conversation
to obtain a sentiment rating concerning the subject matter of
interest.
56. The article of manufacture of claim 52, further comprising one
or more of: instructions to calculate how the sentiment concerning
the subject matter of interest trends over time; instructions to
calculate how the sentiment concerning the subject matter of
interest varies by online source or group of sources; and
instructions to calculate how the sentiment concerning the subject
matter of interest concurrently trends over time and varies by
online source or group of sources.
57. The article of manufacture of claim 52, further comprising
instructions to determine an overall volume of the online social
media conversations referring to the subject matter of interest by
adding up a number of occurrences of the subject matter of interest
in online social media conversations per unit of time.
58. The article of manufacture of claim 57, further comprising
instructions to determine how the overall volume of online social
conversations referring to the subject matter of interest trends
over time.
59. The article of manufacture of claim 52, further comprising
instructions to determine a share of online voice acquired by the
subject matter of interest with respect to other online social
media subject matter.
60. The article of manufacture of claim 59, wherein the other
online social media subject matter includes subject matter
associated with competitors in the category.
61. The article of manufacture of claim 60, further comprising
instructions to calculate one or more of: how the share of online
voice acquired by the subject matter of interest trends over time;
and how the share of online voice acquired by the subject matter of
interest trends over time with respect to the subject matter of the
competitors in the category.
62. The article of manufacture of claim 52, further comprising
instructions to perform text edge processing on the online social
media conversations to determine frequency of occurrence of one or
more topics in conjunction with the subject matter of interest and
relatedness of the one or more topics to the subject matter of
interest.
63. An article of manufacture comprising: a computer-readable
storage medium providing instructions which, when executed by a
computer, cause the computer to perform for operating an enhanced
customer service call center, the instructions comprising:
instructions to receive a customer service call at a call center
from a customer concerning a product or service; instructions to
extract information from the customer service call; instructions to
retrieve actionable information stored in a database relating to
the customer service call, the actionable information generated by
harvesting, aggregating and quantifying unstructured qualitative
online social media conversations from one or more online social
media sources; and instructions to provide enhanced customer
service to the customer based on the actionable information.
64. The article of manufacture of claim 63, further comprising
instructions to take action based on the actionable information
including one or more: instructions to recommend or perform an
action based on the actionable information relevant to the customer
service call; and instructions to provide guidance to a call-center
operator on how to respond to the customer service call based on
the actionable information.
65. The article of manufacture of claim 64, wherein the guidance to
the call-center operator includes how to deal with ongoing issues
concerning the product or service and how to identify new issues
concerning the product or service.
66. The article of manufacture of claim 63, further comprising
instructions to identify and engage representatives of influential
online social media websites.
67. The article of manufacture of claim 63, wherein the actionable
information includes real-time alerts providing early warnings
about customer service or quality issues.
68. The article of manufacture of claim 67, further comprising
instructions to provide crisis prevention and management based on
the real-time alerts.
69. An article of manufacture comprising: a computer-readable
storage medium providing instructions which, when executed by a
computer, cause the computer to perform a method of enhancing
targeted advertising campaigns, the instructions comprising:
instructions to retrieve actionable information stored in a
database relating to a product or service to be advertised, the
actionable information generated by harvesting, aggregating and
quantifying unstructured qualitative online social media
conversations from one or more online social media sources; and
instructions to target advertising campaigns based on the
actionable information.
70. The article of manufacture of claim 69, wherein the actionable
information includes one or more: industry or sector corresponding
to the product or service to be advertised; sentiment expressed by
online social media participants about the product or service in
the online social media conversations; trends in the sentiment
expressed by the online social media participants about the product
or service; and alerts concerning the product or service.
71. The article of manufacture of claim 69, wherein the actionable
information includes one or more of: most relevant websites,
wherein the most relevant websites include one or more of websites
most likely to be reached in online searches for information
relating to the product or service and websites where high-affinity
social media participants are exchanging opinions and making
purchasing decisions; and most relevant advertisement networks for
advertising the product or service.
72. The article of manufacture of claim 69, wherein the actionable
information provides one or more of: perception of the product or
service among the online social media participants to be
advertised; trend recognition of sentiment expressed by the online
social media participants concerning the product or service;
opportunity identification; and competitor or monitoring.
73. The article of manufacture of claim 69, further comprising
instructions to observe results in the marketplace for the product
or service based on the targeted advertising campaign.
74. An article of manufacture comprising: a computer-readable
storage medium providing instructions which, when executed by a
computer, cause the computer to perform a method of enhancing
marketing research, the instructions comprising: instructions to
retrieve quantitative actionable information stored in a database
relating to a product or service to be marketed, the actionable
information generated by harvesting, aggregating and quantifying
unstructured qualitative online social media conversations from one
or more online social media sources; and instruction to provide
enhanced marketing research based on the actionable
information.
75. The article of manufacture of claim 74, wherein the actionable
information is made widely available inside an organization using
an interface to push the actionable information to everyone inside
the organization.
76. The article of manufacture of claim 74, wherein the actionable
information enables marketing staff to establish an overall sense
of the voice of their customers and to make informed decisions at a
customer level.
77. The article of manufacture of claim 74, further comprising
instructions to provide statistical analysis of the online social
media conversations to gain insight into social behavior of
potential customers to inform enhanced marketing decisions.
78. An article of manufacture comprising: a computer-readable
storage medium providing instructions which, when executed by a
computer, cause the computer to perform a method of improved
product development, the instructions comprising: instructions to
retrieve quantitative actionable stored in a database information
relating to one or more products, the actionable information
generated by harvesting, aggregating and quantifying unstructured
qualitative online social media conversations from one or more
online social media sources; and instructions to develop new
products based on the actionable information.
79. The article of manufacture of claim 78, wherein the actionable
information includes sentiment expressed among online social media
participants about the one or more products.
80. The article of manufacture of claim 79, wherein the actionable
information further includes sentiment expressed among the online
social media participants about one or more features of the one or
more products of the user of the social media analytics
platform.
81. The article of manufacture of claim 79, wherein the retrieved
actionable information includes information relating to one or more
of competitors' products, competitors' pricing, competitors'
advertising and competitors' sales strategies.
82. The article of manufacture of claim 79, further comprising
instructions to provide statistical analysis on the online social
media conversations to gain increased knowledge of competitors
based on the actionable information and develop new products in
response to the actionable information.
83. An article of manufacture comprising: a computer-readable
storage medium providing instructions which, when executed by a
computer, cause the computer to perform a method for enhanced
opinion polling, the instructions comprising: instructions to
retrieve quantitative structured actionable information stored in a
database relating to the subject matter of an opinion poll, the
actionable information generated by harvesting, aggregating and
quantifying unstructured qualitative online social media
conversations from one or more online social media sources; and
instructions to ascertain opinion among online social media
participants concerning the subject matter of the opinion poll
based on the actionable information.
84. An article of manufacture comprising: a computer-readable
storage medium providing instructions which, when executed by a
computer, cause the computer to perform a method for providing
enhanced national defense intelligence, the instructions
comprising: instructions to retrieve quantitative structured
actionable information stored in a database related to potential
violent terrorist activity, the actionable information based on
harvested, aggregated and quantified unstructured qualitative
online social media conversations from one or more online social
media sources; and instructions to identify potential violent
terrorist activity in the online social media conversations based
on the actionable information.
85. The article of manufacture of claim 84, further comprising
instructions to identify previously unknown online social media
sources used by groups to engage in violent terrorist activity
based on the actionable information.
86. The article of manufacture of claim 84, further comprising:
instructions to monitor keywords associated with violent terrorist
activity in the unstructured qualitative online social media
conversations; instructions to perform text edge processing on the
keywords associated with violent terrorist activity; and
instructions to identify previously unknown keywords as associated
with violent terrorist activity based on performing the text edge
processing.
87. The article of manufacture of claim 86, further comprising
instructions to identify keywords that appear with greatest
frequency in conjunction with keywords and phrases in the online
social media conversations identified as associated with violent
terrorist activity.
88. The article of manufacture of claim 84, further comprising
instructions to predict preparation, planning, and execution of
violent activities based on actionable information.
89. The article of manufacture of claim 84, wherein the violent
activities include manufacture or distributing improvised explosive
devices (IEDs).
90. An apparatus comprising: a harvesting module configured to
harvest online social media conversations relevant to subject
matter of interest in a category from one or more online social
media sources; and a vertical module configured to: aggregate the
harvested online social media conversations; quantify the
aggregated online social media conversations to obtain structured
analytic measurements of the online social media conversations
including analytic measurements of sentiment expressed among online
social media participants concerning the subject matter of interest
in the category; and provide actionable information based on the
analytic measurements of the online social media conversations.
91. The apparatus of claim 90, wherein the vertical module
comprises a sentiment rating processing module configured to rate
the sentiment expressed among online social media participants
concerning subject matter of interest.
92. The apparatus of claim 90, wherein the vertical module
comprises a text edge processing module configured to determine:
frequency of occurrence of one or more topics in conjunction with
subject matter of interest; and relatedness of the one or more
topics to the subject matter of interest.
93. The apparatus of claim 92, wherein the vertical module further
includes a text trend processing module configured to quantify how
the frequency of occurrence of the one or more topics in
conjunction with the subject matter of interest trends over
time.
94. The apparatus of claim 90, wherein the vertical module further
includes a volume trend processing module to perform one or more
of: calculate an overall volume of harvested online social media
conversations referring to the subject matter of interest; and
calculate how the overall volume of harvested online social media
conversations referring to the subject matter of interest trends
over time.
95. The apparatus of claim 90, wherein the vertical module further
includes a sentiment aggregation processing module configured to
perform one or more of: calculate how the sentiment among online
social media participants concerning the subject matter of interest
trends over time; and calculate how the sentiment concerning the
subject matter of interest varies by online source or group of
sources.
Description
CROSS-REFERENCE TO RELATED APPLICATIONS
[0001] This application claims the benefit of U.S. provisional
application No. 61/114,445, entitled "Aggregating and Presenting
Quantitative Online Social Media Content," filed on Nov. 13,
2008.
FIELD OF THE INVENTION
[0002] At least certain embodiments of the invention relate
generally to information management, and more particularly to
providing analytic measurement of online social media content.
BACKGROUND OF THE INVENTION
[0003] Traditional methods of collecting, managing and providing
real-time or near real-time relevant information have been enhanced
through the use of the Internet and online research and information
collection tools. One such set of tools is known as web analytics.
Web analytics focus on a company's own website for collection of
online information, particularly traffic data. Web analytics are
limited because they only consider a subset of the relevant online
universe, specifically the behavior of users of a given website.
They do not discover other information about the users such as
interests and opinions expressed in interactive systems. Behavioral
analytics are another set of information collection and management
tools that attempts to analyze the "click stream" of users and show
advertisements based on this information. However, this method has
many technical limitations since it tends to provide only a very
limited picture of a user's overall interests. Also there is a lack
of consolidation between a user's work and home PCs.
[0004] Online social media is a new source of valuable information
on the Internet that may be harvested to generate information and
other data about products or services, branding, competition, and
industries. Online social media encompasses online media such as
blogs and sub-blogs, online discussion forums, social networks,
wiki sites such as Wikipedia, online reviews on e-commerce sites
such as Amazon.com.RTM., video sites such as YouTube.RTM.,
micro-blogging services such as Twitter.RTM., and so on. There are
currently over 106 million blogs growing at a rate of 11% per year.
There are several million forums with active contributions by more
than 33% of Internet users. There are 483 million users of social
networks worldwide growing at a rate of 47% annually. As a result,
social media is becoming a crucial and rapidly growing source of
consumer opinion. This information may allow users to quantify
opinion on social media sites to gain useful insights into current
consumer sentiment and trends relating to their products or
services, brands, and/or technologies, and those of their
competitors. Collecting and presenting this information can help
users in a variety of ways such as, for example, target advertising
revenues and expenditures, marketing, sales, customer service,
brand management, product development, investor relations, and so
on. Social networking sites are currently trying to leverage their
own user profiles to target advertising based on their users'
behavior and declared interests. However, most users today
participate in several different online social media sites. Online
content analytics are another set of information collection tools
that attempts to analyze content in social media sites such as
online forums, blogs, and so on. However, these techniques require
a high degree of manual human intervention by analysts.
Additionally, the reports generated by these analysts can be very
expensive and can't be updated very frequently due to the necessity
of human intervention in the data gathering and analysis
process.
SUMMARY OF THE DESCRIPTION
[0005] At least certain embodiments disclose a social media
analytics platform for providing analytic measurements of online
social media content by harvesting and aggregating unstructured
qualitative online social media conversations relevant to subject
matter of interest in a category from one or more online social
media sources, quantifying the aggregated online social media
conversations, and providing actionable information based on the
quantified aggregated online social media conversations, the
actionable information including sentiment expressed among online
social media participants concerning subject matter of interest in
the category.
BRIEF DESCRIPTION OF THE DRAWINGS
[0006] A better understanding of at least certain embodiments of
the invention can be obtained from the following detailed
description in conjunction with the following drawings.
[0007] FIG. 1A illustrates a block diagram of a social media
analytics platform according to an exemplary embodiment of the
invention.
[0008] FIG. 1B illustrates a block diagram of the harvesting layer
according to an exemplary embodiment of the invention.
[0009] FIG. 2 illustrates harvesting layer processing according to
an exemplary embodiment of the invention.
[0010] FIG. 3 illustrates a block diagram of the vertical layer
according to an exemplary embodiment of the invention.
[0011] FIG. 4A illustrates vertical layer processing according to
an exemplary embodiment of the invention.
[0012] FIG. 4B illustrates additional vertical layer processing
according to an exemplary embodiment of the invention.
[0013] FIG. 4C illustrates text edge parsing for an individual
sentence according to an exemplary embodiment of the invention.
[0014] FIG. 4D illustrates an excerpt from an aggregated term graph
according to an exemplary embodiment of the invention.
[0015] FIG. 5 illustrates a block diagram of the top websites
filtering subsystem according to an exemplary embodiment of the
invention.
[0016] FIG. 6A illustrates performing top websites filtering
according to an exemplary embodiment of the invention.
[0017] FIG. 6B illustrates an exemplary website link network
according to one embodiment of the invention.
[0018] FIG. 6C illustrates an excerpt from a website graph
according to an exemplary embodiment of the invention.
[0019] FIG. 7 illustrates a block diagram of the presentation layer
according to an exemplary embodiment of the invention.
[0020] FIG. 8 illustrates presenting the aggregated and quantified
online social media content to users of the social media analytics
platform according to an exemplary embodiment of the invention.
[0021] FIG. 9 illustrates a dashboard display in a graphical user
interface according to an exemplary embodiment of the
invention.
[0022] FIG. 10 illustrates a newest posts display in a graphical
user interface according to an exemplary embodiment of the
invention.
[0023] FIG. 11 illustrates an online social media post as it
appears in its originating site according to an exemplary
embodiment of the invention.
[0024] FIG. 12 illustrates a search results display in a graphical
user interface according to an exemplary embodiment of the
invention.
[0025] FIG. 13 illustrates an overall brand sentiment menu display
in a graphical user interface according to an exemplary embodiment
of the invention.
[0026] FIG. 14 illustrates a products or services sentiment display
in a graphical user interface according to an exemplary embodiment
of the invention.
[0027] FIG. 15 illustrates a smoothed view of a brand trend lines
display in a graphical user interface according to an exemplary
embodiment of the invention.
[0028] FIG. 16 illustrates a detailed view of a brand trend lines
display in a graphical user interface according to an exemplary
embodiment of the invention.
[0029] FIG. 17 illustrates a brand sentiment by source menu display
in a graphical user interface according to an exemplary embodiment
of the invention.
[0030] FIG. 18 illustrates a display of sentiment indices for a
brand's products or services for a particular source in a graphical
user interface according to an exemplary embodiment of the
invention.
[0031] FIG. 19 illustrates a brand source trends for a particular
source group display in a graphical user interface according to an
exemplary embodiment of the invention.
[0032] FIG. 20 illustrates a positive/negative posts display in a
graphical user interface according to an exemplary embodiment of
the invention.
[0033] FIG. 21 illustrates an example ad hoc sentiment trend chart
in a custom query display in a graphical user interface according
to an exemplary embodiment of the invention.
[0034] FIG. 22 illustrates a custom query for sentiment display in
a graphical user interface according to an exemplary embodiment of
the invention.
[0035] FIG. 23 illustrates a products or services trend lines
display in a graphical user interface according to an exemplary
embodiment of the invention.
[0036] FIG. 24 illustrates a products or services sentiment by
source display in a graphical user interface according to an
exemplary embodiment of the invention.
[0037] FIG. 25 illustrates a products or services source trends
display in a graphical user interface according to an exemplary
embodiment of the invention.
[0038] FIG. 26 illustrates a share of voice display in a graphical
user interface according to an exemplary embodiment of the
invention.
[0039] FIG. 27 illustrates a share of voice trends display in a
graphical user interface according to an exemplary embodiment of
the invention.
[0040] FIG. 28 illustrates a volume trends display in a graphical
user interface according to an exemplary embodiment of the
invention.
[0041] FIG. 29 illustrates a topic radar display in a graphical
user interface according to an exemplary embodiment of the
invention.
[0042] FIG. 30 illustrates a tag cloud display in a graphical user
interface according to an exemplary embodiment of the
invention.
[0043] FIG. 31 illustrates a products or services share of voice
trends display in a graphical user interface according to an
exemplary embodiment of the invention.
[0044] FIG. 32 illustrates a custom query for topics display in a
graphical user interface according to an exemplary embodiment of
the invention.
[0045] FIG. 33 illustrates a forum opinion leader list display in a
graphical user interface according to an exemplary embodiment of
the invention.
[0046] FIG. 34 illustrates an overall brand advocacy display in a
graphical user interface according to an exemplary embodiment of
the invention.
[0047] FIG. 35 illustrates an exemplary data processing system upon
which the methods and apparatuses of the invention may be
implemented.
DETAILED DESCRIPTION
[0048] Throughout the description, for the purposes of explanation,
numerous specific details are set forth in order to provide a
thorough understanding of the present invention. It will be
apparent to one skilled in the art, however, that the present
invention may be practiced without some of these specific details.
In other instances, well-known structures and devices are shown in
block diagram form to avoid obscuring the underlying principles of
embodiments of the invention.
[0049] At least certain embodiments disclose a social media
analytics platform including methods, apparatuses, and
computer-readable media for providing analytic measurements of
online social media content by harvesting and aggregating
unstructured qualitative online social media conversations relevant
to subject matter of interest in a category from one or more online
social media sources, quantifying the aggregated online social
media conversations including performing statistical analysis on
the conversations to obtain the analytical measurements of the
online social media content, and providing actionable information
based on the quantified aggregated online social media
conversations, the actionable information including sentiment
expressed among online social media participants concerning subject
matter of interest in the category. The actionable information may
be retrieved based on real-time measurements of the online social
media sentiment expressed among the online social media
participants. The actionable information may also be retrieved
based on historical data representing quantitative measurements of
the sentiment expressed among online social media participants in
the past. Embodiments also assign a sentiment rating to each of the
online social media conversations referring to the subject matter
of interest. Embodiments then present the analytic measurements of
the online social media content in a user interface in an intuitive
and user-friendly manner at different levels of granularity that
enables the quantified online social media content to be grouped
and filtered in a variety of default and/or customizable ways. In
one embodiment, the user interface is a web-based service. In
alternative embodiments, the user interface may be provided as a
browser-based interface hosted by a behind-the-firewall hardware or
software appliance.
[0050] Embodiments provide analytic measurement of online social
media content for users such as global enterprises, advertising
agencies, sales and marketing departments, media companies,
government agencies, and virtually any entity requiring real-time
or near real-time access to such information. This online social
media content is quantified and provided in a relevant and
user-friendly manner to these entities using an interface such as a
graphical user interface (GUI). These embodiments provide both
historical and current measurements to enable analysis of past and
present information. Online social media content is harvested,
sorted, and provided to relevant groups or entities. Certain
embodiments describe a social media analytics platform for
collecting and converting raw online social media conversations
into actionable information that can be used to increase the
top-line growth and margins of its recipients. Additionally, this
aggregation of social media information can be analyzed to
determine trends in each of the above discussed categories.
[0051] Monitoring and aggregating this new information source may
be used on its own or in conjunction with traditional research and
measurements such as, for example, quantitative and qualitative
market research, paid media tracking, and traditional web site
analytics. This process is automated so that qualitative
measurements can be aggregated, quantified, and presented with
minimal human intervention. At least certain embodiments
contemplate a harvesting process referred to herein as "scraping"
where social media sources are discovered or located and exploited
for relevant information. The content is then aggregated and
quantified in a manner relevant to the industry or other category.
The aggregated and quantified online social media content is then
provided to the user of the social media analytics (SMA) platform
in an efficient, timely and user-friendly manner using the
interface. In one embodiment, the interface is user-specific.
[0052] Examples of the quantitative online social media content
data that can be provided by embodiments include: brand and
product/service sentiment for users and their competition; the
share of voice of the brand (e.g., volume of discussion about the
brand, product or service) over the social media versus the
competition; topics and keywords used by online discussion
participants for the brand and the competition; information on the
opinion leaders for the category (e.g., online social content
authors with the most influential voices); top websites resulting
from the brand search; automated alerts for changes in sentiment;
keywords, terms or phrases in posts to the online social media
websites; and much more. This information is aggregated,
quantified, and provided to users in real-time or near-real-time
for the purpose of, for example, marketing, public relations,
advertising, sales, customer service, brand management, product
development, investor relations, and so on. The result of this
process is to provide highly relevant and timely actionable
information to users of the SMA platform.
[0053] This information may be advantageous for several reasons
including brand and product/service perception or sentiment
analysis, trend recognition and opportunity identification, early
warnings about customer service or quality issues, opinion leader
identification and engagement, competitor monitoring, and optimized
online advertising to name a few. This information allows users to
quantify opinion on social media sites to gain insights into
current consumer sentiment about the users' products or services,
brands, and technologies and those of their competitors. This
information also enables users of the SMA platform to recognize
trends in consumer buzz about new technologies, product or service
types, and attributes. In addition, users may receive early-warning
signs to identify dissatisfied customers. Users also may identify
and target opinion leaders for a given product/service or category
using this information. Embodiments of the SMA platform can also
supplies users with a list of highly relevant websites where
high-affinity users are exchanging opinions and making purchasing
decisions. This information can also be made widely available
inside users' organizations using an interface to push analytics to
potentially everyone inside the organization instead of just the
top-level marketing staff enabling entire organizations to
establish an overall better sense of the voice of their customers
and to make informed decisions at the customer level because
embodiments focus on the social behavior of potential customers
using online social media sources and provide far better insight
into commercially relevant interests.
[0054] FIG. 1A illustrates a block diagram of a social media
analytics platform according to an exemplary embodiment of the
invention. In the illustrated embodiment, the SMA platform 199 is
separated into three layers or phases--the harvesting layer 100,
vertical layer 300 and presentation layer 700. The harvesting layer
100 includes locating or discovering social media sources (e.g.,
websites) from the Internet related to a particular industry or
other category, and harvesting the relevant content from those
sources. The harvesting layer may process the relevant content from
these Internet sources at any frequency such as daily, hourly,
weekly, and minute-by-minute. The vertical layer includes
aggregating and quantifying the harvested social media content, and
the presentation layer includes a user interface to display the
quantified online social media content and an alerter to alert
users of the SMA platform 199 in a real-time or near real-time
manner when changes occur in sentiment. The basic structure
includes data collection and storage of online social media content
for specific industries or other categories. The data collection
and storage of online social media content may be performed for any
type of category or product line.
[0055] The harvesting layer 100 of FIG. 1A includes online social
media sources discovered or located on the Internet 101 including
social media source 1_107, social media source 2_109, social media
source 3_111, and so on through social media source N_113. Vertical
layer 300 of SMA platform 199 is where the online social media
content relevant to each industry is aggregated, quantified, and
stored in a database. In the illustrated embodiment,
Industry1-specific data aggregation and quantification 115 receives
content from social media source 1_107 and social media source
2_109 of harvesting layer 109, industry2-specific data aggregation
and quantification 117 receives content from social media source
1_107, social media source 3_1 11, and social media source N_113,
and industryN-specific data aggregation and quantification 119
receives content from social media source N_113. For every
identified source, relevant social media content is retrieved and
processed.
[0056] The vertical layer 300 stores the aggregated and quantified
online social media content in a database and supplies the content
to the presentation layer 700 for display. Presentation layer 700
of FIG. 1 A includes user-specific web user interface 121 for
display of the aggregated and quantified online social media
content received from vertical layer 300. Presentation layer 700
also includes a web service application programming interface (API)
to provide fully automated data integration into third-party
analytics or data presentation systems, and a user-specific alerter
123 to provide alerts relating to changes in online social media
sentiment. The user-specific alerter 123 may be tailored for each
user of the SMA platform 199.
[0057] FIG. 1B illustrates a block diagram of the harvesting layer
according to an exemplary embodiment of the invention. As discussed
above, the harvesting layer 100 locates online social media content
sources on the Internet and harvests relevant content from them.
The block diagram components of the harvesting layer 100 will be
discussed in conjunction with process 200 of FIG. 2, which
illustrates harvesting layer processing according to an exemplary
embodiment of the invention. Process 200 begins with performing
forum analysis using forum analyzer 127 (operation 201). The
function of the forum analyzer 127 is to scour the Internet 101
searching for online social media conversations (threads) relevant
to a particular industry, product/service or other category. In at
least certain embodiments, the forum analyzer 127 accomplishes this
using automated tools for identifying industry-specific social
media data sources from which to harvest information and provide to
the users of the SMA platform. This includes a forum analysis to
locate or discover which forums and/or sub-forums are relevant to a
specific user's industry or other category from which the online
social media content should be harvested. To accomplish this,
search results from publicly available online search engines are
processed to determine relevant websites based on the relevance
score of each site for the keywords of interest. Each website found
through this process is then accessed by the system to determine
structural properties such as the technical nature of the source
(e.g. RSS feeds, certain discussion forum software packages) and to
identify the entry page locations later used in the content
scraping module 131. The online social media content sources that
are identified in this operation are then staged in the scraping
queue 129 to feed the content scraping module 131 for the scraping
process (operation 203).
[0058] At operation 203 the scraping process is performed including
scouring the identified online social media sources for
conversations relevant to a particular sector or other category and
breaking down the content into pieces to be stored for later
processing. The scraping process starts at an overview page
typically provided by each social media source and identifies
hyperlinks to potentially relevant subpages and content pages based
on the structural properties of these hyperlinks. The process then
iteratively drills down multiple levels of subpages in the same
manner until a specific relevant discussion thread is found. Each
discussion thread is then analyzed in order to isolate its atomic
content components for further processing. For example, a
particular relevant social media source (e.g., website) may have a
web page with a thread containing 20 different posts relating to
the Audi A6 automobile. In such a case, the web page would be
retrieved and broken apart into 20 pieces, with each piece stored
individually along with the user-profile information of the authors
who posted the content.
[0059] The results of the scraping process include: the raw
conversations of each social media post referred to as the raw post
content data; the metadata of the raw post content; and information
relating to the author of each post, as well as relationships
between authors, referred to as the raw social graph data. The raw
post content retrieved from the online social media sources is
stored in raw content storage 133 (operation 205). This includes
the actual text of the relevant social media post. The raw content
metadata is also stored in raw content metadata storage 135
(operation 207). The raw content metadata includes information such
as the URL of the social media website, and the length, context,
and time of the post. Additionally, the raw social graph data is
stored in raw social graph data storage 137 (operation 209). This
data may include the social media post's author profile data such
as the author's username, demographic information, number of posts
to the social media website, those responding to the author's
posts, and the author's contacts.
[0060] In the illustrated embodiment, the social network analysis
(SNA) processing is then performed on the raw social graph data
stored in raw social graph data storage 137 (operation 21 1). Here,
information on each author of a social media post and on those
responding to the author's post is retrieved from the raw social
graph storage 137 and used to generate a social graph which
includes an aggregation of social network information that can be
useful in several contexts. For example, the social graph data may
be analyzed to determine information about the author's social
network including which authors are communicating about what
topics, who is responding to which posts, what the related content
is, and so on. The SNA processing is used to develop this
information on networks of related authors and posts and to
determine which authors are the most influential within these
networks based on the social graph. The SNA processing first
calculates a so-called centrality value for each author that
expresses the author's degree of influence in a given social
network. Authors that are connected to a large number of other
authors and also connected to distinct sub-groups of authors are
assumed to have higher influence than less well-connected authors.
In order to calculate the centrality value, a version of Brandes'
Betweenness Centrality algorithm is applied to the raw social graph
for each website. The resulting raw centrality value is then
modified with the activity level of the author, i.e. the number of
posts written by this person, and an importance score for the
website where that author is active. Within graph theory and
network analysis, there are various measures of the centrality of a
vertex within a graph that determine the relative importance of a
vertex within the graph. Betweenness is a centrality measure of a
vertex within a graph. Vertices that occur on many shortest paths
between other vertices have higher betweenness than those that do
not. For instance, an influential author on a large website such as
MySpace.RTM. will receive a higher influence score than the author
of a little known blog. In at least one embodiment, the influence
score for each author is calculated by the following formula:
Influence score=bc*(c.sub.a+a/p.sub.a)*(c.sub.p+p), where [0061] bc
is the raw betweenness centrality value for the author; [0062] a is
the number of active authors on the website where the author is
active; [0063] p is the number of posts that the author has
contributed; [0064] c.sub.a, p.sub.a, c.sub.p are correction
parameters that are fine-tuned for the purposes of a specific
vertical (i.e., a specific category of interest).
[0065] The SNA processing also provides information including: the
websites on which each of the social media authors have
contributed; registrations in social networks; the status of
influence of the authors; the author's sentiment towards a given
brand, product or service; known demographic and geographic
information about the authors; and trends in all of the above.
[0066] The social graph is then stored in social graph storage 141
(operation 213). An additional input into the social graph storage
141 is from user-profile scraping data accumulated from the
Internet 101 using user-profile scraping module 143. At operation
215, the user profile scraping module 143 scours the Internet 101
to find any other information about the authors of the online
social media conversations. Whatever information associated with
the author that can be harvested from the Internet 101 is collected
and stored along with the social graph in social graph storage 141
(operation 217). This completes the harvesting layer process 200
according to an exemplary embodiment.
[0067] FIG. 3 illustrates a block diagram of the vertical layer
according to an exemplary embodiment of the invention. As discussed
previously, the data collected using the scraping process 100 is
fed into the vertical layer 300. The vertical layer 300 is a
grouping based on sector, industry, or other category. A vertical
layer may be generated for every conceivable category such as
industry, topic of interest, type of website, geographic region,
and so on. There is essentially no limit to the types of categories
that can be harvested, aggregated and quantified to provide
relevant, timely and actionable information to users of the SMA
platform. The block diagram components of the vertical layer 300
will be discussed in conjunction with process 400A of FIG. 4A and
process 400B of FIG. 4B. FIG. 4A illustrates vertical layer
processing according to an exemplary embodiment of the invention
and FIG. 4B illustrates additional vertical layer processing
according to an exemplary embodiment of the invention.
[0068] Process 400A begins with receiving data 145 at processing
module 301 from storage (operation 401). The data 145 received from
storage is the output data 145 from FIG. 1B including the raw
content data from raw content data storage 133, the raw content
metadata from raw content metadata storage 135, and the social
graph data from social graph storage 141. Process 400A continues
with performing text edge processing on the raw content data from
raw content data storage 133 and the raw content metadata from raw
content metadata storage 135 (operation 403). Text edge processing
is performed using text edge processing module 303 of processing
module 301. Text edge processing, in one embodiment, utilizes graph
theory to analyze the terms and concepts contained within the
online social media conversations to determine the frequency of
occurrence of these terms and concepts in conjunction with the
relevant brand, product or service and the relatedness of the
concepts and/or terms in the post to that brand, product or
service. Relationships between these terms are analyzed to
determine graph edges which indicate the strength of these
relationships. In a first step, a relevant sentence is parsed and
split up into individual words and tuples of adjacent words. Stop
words with little informational value such as "of," "it," "is" and
so on are excluded in this step. Next, the relationship between the
main term of interest (e.g. a brand, service or product name) and
each found word or tuple is stored. FIG. 4C illustrates text edge
parsing for an individual sentence according to an exemplary
embodiment of the invention. In the illustrated embodiment, the
sentence, "[t]he Audi A6 is a very fast car with a great engine,"
is parsed to determine relationships between the main terms of
interest (Audi 431 and Audi A6 433) and each found word or tuple in
the sentence (A6 435, fast 437, car 439, fast car 441, great 443,
engine 445, and great engine 447). In FIG. 4C, the lines between
the main terms of interest and each found word or tuple indicate
that such a relationship exists. Each relationship is then counted
as one instance of an "edge" between these connected objects. In
the following aggregation step, the number of edges between objects
is added up. The resulting frequency of edge occurrences is an
indication of how closely two terms are connected. For instance, if
the tuple "fast car" is used significantly more frequently in
connection with car brand A than with car brand B (corrected by the
total number of posts about each brand), we can assume that social
media users perceive car brand A as a stronger producer of fast
cars. FIG. 4D illustrates an excerpt from an aggregated term graph
according to an exemplary embodiment of the invention. The number
of edges between brands (brand A 451, brand B 453, and brand C 455)
and each found word or tuple (fast 457, car 459, fast car 461,
great 463, engine 465, and great engine 467) are added up to
determine the frequency of edge occurrences. For example, in the
illustrated embodiment, brand A 451 has a total of n=3,983 edge
occurrences with respect to the tuple "fast car 461." In contrast,
brand B 453 only has n=2664 edge occurrences with respect to the
tuple "fast car 461." Thus, from the fact that the tuple fast car
461 is used significantly more frequently in connection with brand
A 452 than with brand B 453 (corrected by the total number of posts
for each brand), we can assume that social media users perceive
brand A 451 as a stronger producer of fast cars than brand B 453.
The data resulting from text edge processing module 303 of
processing module 301 is then stored in text edge storage 307
(operation 405).
[0069] Sentiment rating processing is then performed using
sentiment rating processing module 305 on the raw content data
stored in raw content data storage 133, the raw content metadata
stored in raw content metadata storage 135, and the social graph
information stored in social graph storage 141 (operation 402).
Sentiment rating processing includes analyzing the actual text of
online social media conversations to find keywords, terms or
phrases to determine if a particular post refers to the particular
brand, product or service of interest. This helps to determine the
sentiment about the brand, product or service. The input to
sentiment rating processing module 305 includes the actual text of
the social media post, lists of keywords, and so on.
Industry-specific keywords are identified and a value or sentiment
rating is assigned to each of these keywords. In at least certain
embodiments, this processing includes natural language and sentence
structure analysis to determine which parts of the text of a social
media post apply to the particular brand, product or service. Once
the keywords are identified, they are processed using a number of
factors including how many times the keyword appears in the social
media post, the closeness and linguistic context of the keyword in
relation to the brand, product or service, and whether the keyword
reflects a positive, negative, or neutral sentiment about the
brand, product or service. This processing may also require
balancing opposing keywords (e.g., both positive and negative
keywords in the same post) to determine an overall sentiment rating
of how positive, negative, or neutral the social media post is in
relation to a brand, product or service.
[0070] Keywords are assigned with a positive and negative
probability value each that express the probability that the
keyword means something positive or negative in the context of the
specific vertical. Since the same word can have different meanings
per industry or topic, these probabilities can be specifically set
per vertical. Also, some embodiments include a training or feedback
loop where keywords may be re-rated over time based on experience.
During the processing, terms of interest (brands, products, service
names) and their synonyms are identified in the text of the social
media post. In a next step, the environment (the closest n words)
of this occurrence is searched for relevant sentiment keywords that
might refer to the term of interest. Linguistic elements such as
negations, comparatives, or enumerations are taken into account
when determining the relevance of a sentiment keyword for the term
of interest. Each occurrence of the term of interest is assigned
with a sentiment score depending on the keywords in the
environment, the linguistic modifiers present, the proximity of the
keyword to the term of interest, and potentially reduced confidence
due to ambiguities. Finally these atomic scores are added up for
the whole post and corrected by the relevance of the post for the
term of interest, i.e. the percentage of the post that actually
refers to the term of interest.
[0071] This information is then combined with the social graph data
from social graph storage 141 to determine a weighting factor of
the social media post. That is, the sentiment rating processing of
operation 402 takes into consideration the level of influence the
author of the social media post has in determining the sentiment
rating. A weighting factor is determined based on the influence of
the author of the social media post. The resulting data from
sentiment rating processing module 305 is then stored in the
sentiment rating storage 309 (operation 404). Additionally, the
sentiment rating data stored in sentiment rating storage 309 is
aggregated over time in the sentiment aggregation queue 311 for
sentiment trend processing to be discussed infra. This completes
process 400A according to an exemplary embodiment and control flows
to process 400B of FIG. 4B. In short, the sentiment rating is
generated using a combination of natural language processing,
statistical processing, positive/negative keyword modifiers and
author and site influence.
[0072] Process 400B begins at operation 409 where data from storage
is received at processing module 302 from storage. The data
received from storage includes the social graph data 149 output
from social graph storage 141 of FIG. 1 B, the data from text edge
storage 307, the data from sentiment rating storage 309, and the
data from sentiment aggregation queue 311. At operation 41 1,
volume trend processing is performed on the data from storage using
volume trend processing module 313 of FIG. 3. The overall volume of
opinions about users' brands, products or services is calculated
and trends over time can be determined based on volume trend
processing. Additionally, volume trends about competing brands and
products or services can be provided in this operation. Basic
volume is calculated using the number of occurrences of a brand,
product or service name and its synonyms per unit of time (e.g.,
day, month, or year). The content authored in each unit of time is
searched for the terms of interest, and the number of occurrences
is added up per unit of time and per term. When plotted in a time
series, these volume data points describe the volume trend for the
brand, product or service. At operation 413, text trend processing
is performed on the data. The text trend processing analyzes the
text edge information stored in text edge storage 307 in
conjunction with time information to determine text trends over
time. This processing is used to determine how sentiment changes
over time. At operation 415, sentiment aggregation processing is
performed on the sentiment rating and aggregation data from storage
using sentiment aggregation processing module 317 of FIG. 3. The
sentiment aggregation processing module 317 determines the
aggregation of sentiment over time for various sources (or groups
of sources) such as relevant websites, blogs, My Space.RTM. pages,
and et cetera. This information may then be used to compare online
social media sources to determine which sources are more favorable
for advertising a user's brands, products, or services. For
example, this processing may determine a particular user's products
or services are better advertised on My Space.RTM. instead of
topic-specific blogs. Additionally, information can be gathered
regarding which websites are initially more relevant for product
releases, for example, and which websites are more relevant over
time. This allows users of the SMA platform to follow these trends
and to roll-out or switch advertising campaigns based on this
information. Process 400B continues with opinion leader aggregation
processing using opinion leader aggregation processing module 319
of FIG. 3 on the data from storage (operation 417). The opinion
leader aggregation processing module 319 determines the aggregation
of opinion leader data over time to determine trends in opinion
leader data. This information may be valuable to users by enabling
them to identify and target social media authors with the most
influence to enter into conversations with these lead authors and
influence their opinion to influence the opinions of many
others.
[0073] At least certain embodiments include additional external
data processing (operation 419). For example, sales data may be
included in the trend processing using sales data processing module
321, traffic data may be included in the trend processing using
traffic data processing module 323, and demographics data may be
included in the trend processing using demographics processing 324.
Sales data processing module 321 allows users to correlate the
sales data with sentiment data over time. This can lead to
predictions in sales volume data and pricing. Traffic data
processing module 323 allows users to correlate the traffic data
with sentiment data over time. Likewise, demographics processing
325 allows users to correlate demographics data with sentiment data
over time. Other external data from users' database sources may
also be included in the processing and correlated with sentiment
data over time.
[0074] Process 400B continues with storing the results of the above
processing in a database referred to herein as the vertical
database (operation 421), and sending this data as output data to
the user interface 705 of the presentation layer 700 for display
(operation 423). Additionally, the results of the above processing
are also output to the alert queue 425 for user alerts when
sentiment trends change above or below a certain threshold, for
example (operation 425). This allows for constant, real-time
monitoring of emerging trends and consumer sentiment. This
completes the vertical layer processing according to an exemplary
embodiment. Control flows to FIG. 8 where the output of the
vertical layer 300 processing is fed into the presentation layer
700 for display to users of the SMA platform.
[0075] FIG. 5 illustrates a block diagram of the top websites
filtering subsystem according to an exemplary embodiment of the
invention. The top websites filtering subsystem 500 is considered a
part of the vertical layer 300 and determines websites that are the
most relevant to a particular user. Subsystem 500 performs one or
more searches using a search engine API (such as Google, Yahoo or
Technorati), pulls out search results from the search engine, and
assembles the search results data to model search behaviors of
search engine users so that a list of the most relevant websites
for a users' brands, products or services can be compiled and
provided to users of the SMA platform. This can provide users with
a list of websites having a high affinity for the users' industry
or products/services so that targeted advertising campaigns can be
launched, for example. Interestingly, this may not always be the
websites with the highest traffic volume. This information is also
fed into the user interface 705 of the presentation layer 700. The
block diagram components of the top websites filtering subsystem
500 will be discussed in conjunction with process 600 of FIG. 6A,
which illustrates performing top websites filtering according to an
exemplary embodiment of the invention.
[0076] Process 600 begins with staging one or more search run
definitions 503 for processing in search queue 501 (operation 601).
Search run definitions contain one or more brand or product names
in combination with any number of other relevant keywords that a
consumer might be searching for. One or more searches of the
Internet 101 corresponding to the one or more search run
definitions 503 staged in search queue 501 are then performed using
one or more search engine APIs 505 (operation 603). The results of
these searches are fed into website and link scraping module 507.
Website and link scraping is then performed (operation 605) using
the website and link scraping module 507. During this operation,
the top websites filtering subsystem 500 actually goes into the
websites found in the one or more searches and follows the website
links within each of these websites. The websites found in the
searches and the links within these websites is assembled for the
purpose of attempting to model search engine users' behavior by
determining which websites search engine users will likely visit
when they run each of the one or more searches. In at least one
embodiment, this information can provide users of the SMA platform
with a list of websites with a high affinity for the users'
industry or products/services. This information may be useful in a
variety of circumstances including allowing users to launch
targeted advertising campaigns. For example, the top websites
filtering subsystem 500 may run a search in Google for digital
cameras and determine that a typical search engine user will only
look at the first 3 web pages listed in the search results. The top
websites filtering subsystem 500 will then follow the links in
these 3 web pages to find more web pages and then follow the links
in those web pages, and so on. The top websites filtering subsystem
500 will assemble this information and use it to build up a website
and link network graph discussed below. The raw search result data
resulting from website and link scraping module 507 is then stored
in search result raw data storage 509 and the metadata is stored in
search result raw metadata storage 511 (operation 607) to be
provided to processing module 502.
[0077] Process 600 continues with performing website graph
processing (operation 609). In at least one embodiment, the website
graph processing includes using graph theory to analyze the website
network to determine the frequency of occurrence of each website in
the website network in connection with the relevant brand, product
or service and to determine the relatedness of each website in the
website network to that brand, product or service. Relationships
between these websites and the relevant brand, product or service
are analyzed to determine graph edges which indicate the strength
of these relationships. First, links between websites that contain
content relevant to the brand, product or service are counted. The
number of links between two websites provides an indication of how
strongly the two websites are interconnected. FIG. 6B illustrates
an exemplary website link network according to one embodiment of
the invention. In the illustrated embodiment, website link network
620 includes three websites with links connecting to one another.
In the example, there are two (2) connections between the websites
Yahoo.com 621 and Edmunds.com 623 including a link from subpage 1
of Yahoo.com 621 to subpage 1 of Edmunds.com 623 and a link from
subpage 3 of Yahoo.com 621 to subpage 3 of Edmunds.com 623.
Likewise, there are four (4) connections between the websites
Edmunds.com 623 and Autoblog.com 625 and two (2) connections
between the websites Autoblog.com 625 and Yahoo.com 621 in the
exemplary website link network 620. Once the number of links
between each pair of websites is counted, a version of Brandes'
Betweenness Centrality algorithm is applied to the resulting graph.
This algorithm calculates centrality values that indicate how
strongly connected a given website is to other relevant websites,
either directly or indirectly. This is depicted in FIG. 6C which
illustrates an excerpt from a website graph according to an
exemplary embodiment of the invention. In the illustrated
embodiment, website graph excerpt 640 includes lines representing
"edges" where each "edge" is a connection between each pair of
websites in the graph. Website A 641 is connected to website B 643,
website D 647, website F 651, website G 653 and website I 647
within one (1) edge. Website A 641 is further connected to website
C 645, website E 649 and website H 655 within two (2) edges.
Therefore, website A 641 is connected to each other website within
one or two edges, so it will receive a high centrality value in
comparison to the other websites. Internet users that find any of
the other websites in the graph when looking for information are
very likely to end up on website A 641; therefore, it is assumed
that website A 641 is highly relevant to this graph. In this manner
websites that are the most relevant to a particular user of the SMA
platform are located.
[0078] The resulting website network graph generated by the website
graph processing module 513 is then stored in website graph storage
517 (operation 611) and the data 519 from the website graph storage
517 is output to the user interface 705 of the presentation layer
700 of FIG. 7 (operation 613). Process 600 continues at operation
608 where website advertisement network processing is performed
using website ad network processing module 151. The website
advertisement network processing, in at least certain embodiments,
uses typical link patterns to identify advertisement networks that
put advertisements on the analyzed websites. Since each
advertisement network uses a particular type of software to provide
advertisement banners, sponsored text links or other forms of
online advertising, the resulting link patterns identify each
advertisement network. Each website might carry advertisements from
one or multiple networks, or no advertising at all. The website
advertisement network processing is performed to provide users of
the SMA platform with information as to which advertisement
networks are the most relevant for advertising their brands,
products, or services. The resulting website advertisement network
information generated by the website ad network processing module
515 is also stored in website graph storage 517 (operation 610) and
output to the user interface 705 of the presentation layer 700 in
FIG. 7 (operation 613). This completes the top websites filtering
process 600 according to an exemplary embodiment. In short, the top
websites filtering subsystem 500 is used to locate websites users
of the SMA platform are most likely to reach when searching online
for information about a particular brand, product or service.
[0079] FIG. 7 illustrates a block diagram of the presentation layer
according to an exemplary embodiment of the invention. The results
of the vertical layer 300 processing and the top websites filtering
subsystem 500 processing are fed into the presentation layer 700.
In the illustrated embodiment, data 147 of FIG. 1B, data 329 of
FIG. 3, and data 519 of FIG. 5 are each fed into user interface
705. That is, the raw social media content stored in raw content
data storage 133, the social graph stored in social graph storage
141, the data stored in vertical data base 327, and the website
graph and website ad network data stored in website graph storage
517 are fed into the user interface 705. Likewise, the data 331
including the results of the processing performed within processing
module 302 of FIG. 3 is fed into the alert queue 703. The user
interface 705 may be a GUI, some embodiments of which are discussed
infra. The block diagram components of the presentation layer 700
will be discussed in conjunction with process 800 of FIG. 8, which
illustrates presenting the aggregated and quantified online social
media content to users of the SMA platform according to an
exemplary embodiment of the invention.
[0080] Process 800 begins by receiving the data stored in the
vertical database 327 of the vertical layer 300 in FIG. 3,
receiving the data stored in the social graph storage 141, and
receiving the data stored in the raw content data storage 133 of
the harvesting layer 100 in FIG. 1B (operation 801). This data is
received and displayed in the user interface 705 (operation 803).
Process 800 also includes receiving data directly from the results
of the processing performed in processing module 302 of FIG. 3
(operation 802). This data is received and staged in the alert
queue 703 (operation 804) to be output to the alerter 701 and the
user interface 705. Among other things, the alerter 701 is used for
alerting users of the SMA platform of real-time or near real-time
changes in user sentiment regarding their brands, products, or
services. This completes process 800 according to an exemplary
embodiment.
[0081] Some of the advantages of the social media analytics
platform are that embodiments provide: brand/product/service-level
analytics including websites frequently talking about the relevant
brand, product or service; social media authors frequently talking
about the brand/product/service; overall volume of opinions about
the brand, product or service; overall sentiment towards the brand,
product or service; volume and sentiment of opinions about
competing brands, products or services; competing brands, products
or services most frequently mentioned in connection with the users'
own brand, product or service; terms used most frequently in
connection with a brand, product or service; and trends and
early-warning alerts for all of the above. Embodiments also provide
site-level analytics including site traffic (unique visitors and
pages viewed), topic distribution of site, overall sentiment
towards a given brand, product, service or technology, number of
active or contributing users, relevance of the active users,
relationships to other relevant sites, and trends in all of the
above. Finally, embodiments provide user-level analytics (users
referred to here are participants in social media sites) including:
sites on which users contributed content; known identities of
users, users' registrations in social networks; influence of users;
users' known ownership and/or use of a given product, service or
technology; users' sentiment toward a given brand, product, service
or technology; users' known demographic and geographical
attributes; and trends in all of the above.
[0082] In at least certain embodiments, a GUI is utilized to
present the quantified and analyzed online social media content in
a manner relevant to the user. The GUI may be fully customizable
giving users the ability to select which charts and graphs should
appear on the login page of the interface. The GUI provides an
intuitive display to visualize brand, product or service sentiment
over time. This display is a quantitative measure of opinion or
sentiment for a brand, product, services, or its competitors and is
derived from an automated aggregation of sentiment ratings on each
individual post to online social media about a brand, product,
services and/or those of their competitors. The GUI includes
various knobs or switches to manipulate the above information in a
variety of ways. Among many other things, inside the GUI users can
filter information by product/service or competitor, groups of
websites, data ranges, or drill down to the lowest level of
granularity of the information to see the actual text of online
social media posts as it appears on the originating source website.
The GUI provides a visualization that allows users to give context
to each social media post and gain familiarity with the posting
website. The GUI is designed to be used by non-expert users without
help from consultants. The GUI not only provides standard
spreadsheet-style visualization such as bar and pie charts, but
also highly innovative approaches including: radar screen;
heatmaps; geographical visualization; 3D clustering, tag clouds,
and timelines. Content may be harvested from as far back as sources
make available. For example, discussion boards can have posts from
many years ago. The start date on the GUI is configurable and is
designed for ease-of-use allowing for a visualization of the
underlying data calculations and aggregations instead of simply raw
data.
[0083] FIG. 9 illustrates a dashboard display in a graphical user
interface according to an exemplary embodiment of the invention.
The GUI display includes top-level menus and submenus. Top-level
menus take users to main measurement categories. Submenus take
users to more detailed information about the main measurement
category. In the illustrated embodiment, the "overview" category is
selected from top-level menu 903 and the "dashboard" category is
selected in the submenu 901. The dashboard display provides a quick
view into key measures of social media participation in users'
particular brands, products, or services. It displays four (4)
small reporting charts on one screen as a way for users to quickly
see key measurements about their brand, product or service.
[0084] The dashboard may be customized according to the users'
needs. The dashboard display in FIG. 9 includes: a brand sentiment
index gauge 907 in the upper left corner; a brand trend line graph
911 in the upper right corner; a share of voice chart 909 in the
lower left corner; and a brand sentiment chart 913 in the lower
right corner. The brand sentiment index gauge 907 tells how
positively or negatively social media participants are talking
about users' brands, products, or services. The brand sentiment
index gauge 907 reflects this online activity for the current
month. They value of zero (0) means neutral sentiment. Positive
values of 20 or above are typically very good. The brand trend line
graph 911 shows how social media participant attitudes and opinions
for a user's brand, product or service have changed over time. This
enables users to see how sentiment has responded to various events
such as advertising campaigns, programs and product launches. The
share of voice chart 909 indicates the percentage of social media
posts referring to the users' brands in comparison with their
competitors. This allows users to gain important insight into the
relative activity the users' brands are generating in online social
media. The brand sentiment chart 913 displays users' annualized
sentiment index in comparison with the indices of users'
competitors for the current year. Clicking on a chart in the
dashboard display takes users to the full-screen version (except
for the sentiment index gauge 907). In one embodiment, each user
can customize the dashboard by selecting the charts the user wishes
to see by default.
[0085] FIG. 10 illustrates a newest posts display in a graphical
user interface according to an exemplary embodiment of the
invention. In the illustrated embodiment, the "overview" category
is selected in top-level menu 1003 and the "newest posts" category
is selected in the submenu 1001. The newest posts display is a view
of user posts 1007 filtered to show the newest posts. Different
filters may be selected such as positive, negative or neutral posts
1015, product/service-level posts, different date ranges 1013, or
to see posts for competitive brands, products, or services. Users
can select the latest content and/or to see posts according to
other parameters. Additionally, the newest posts menu includes a
"link to original post" 1009 capability that allows users to see
content as it appears in the originating site. This can help give
context to the post and let users gain familiarity with the website
containing the post. Linking to the original post takes users to
the content in the originating site. For example, clicking on the
link to original post 1009 takes users to the post as it appears on
the website such as that shown in FIG. 11, which illustrates an
online social media post as it appears in its originating site
according to an exemplary embodiment of the invention.
[0086] The GUI also enables users to perform keyword searches and
displays a listing of the keyword search results. FIG. 12
illustrates a search results display in a graphical user interface
according to an exemplary embodiment of the invention. In the
illustrated embodiment, the "overview" category is selected in
top-level menu 1203 and the "search" category is selected in the
submenu 1201. The search feature allows users to execute ad hoc
searches for posts to online social media using keywords 1211 and
clicking on the search 1207 button. The search may be constrained
by date range 1209 if desired. The results of the search are shown
in the summaries 1205 of the list of matches. The full post content
can be seen by clicking on the summary 1205 in the list of
matches.
[0087] FIG. 13 illustrates an overall brand sentiment menu display
in a graphical user interface according to an exemplary embodiment
of the invention. In the illustrated embodiment, the "brand
sentiment" category is selected in top-level menu 1303 and the
"overall brand sentiment" category is selected in the submenu 1301.
As discussed above, the brand sentiment index for brands, products
or services and competitors is a quantitative measure of opinion.
This index is an aggregation of automated sentiment ratings on each
individual post to online social media about the brand, products,
services or those of a competitor. A combination of natural
language processing, statistical processing, positive/negative
keyword modifiers and author and site influences may be used to
rate each post to online social media. In at least certain
embodiments, the index is based on a scale from -150 to +150 where
zero (0) equals neutral opinion, +150 reflects extreme positive
sentiment, and -150 reflects extreme negative sentiment. Values
above +20 are typically good. This bar chart is a comparative
display of the brand's sentiment index with respect to the
competition (based on year-to-date sentiment). The overall brand
sentiment chart provides a quick assessment of opinion about the
brand relative to the opinion about the brands of competitors. The
x-axis 1305 reflects the brand sentiment index values and the
y-axis 1307 reflects a list of brands, products or services. The
chart displays year-to-date sentiment by default, but users can
select a narrower date range 1309 if desired. Holding a mouse over
a bar in the graph causes the display of the year-to-date sentiment
index 1311. Clicking on a bar in the graph drills down to show
sentiment for the brand's products or services as shown in FIG. 14
which illustrates a products or services sentiment display in a
graphical user interface according to an exemplary embodiment of
the invention. In the illustrated embodiment, the "brand sentiment"
category is still selected in top-level menu 1403 and the "overall
brand sentiment" category is still selected in the submenu 1401
(even though the products or services sentiment for the particular
brand is displayed). The x-axis 1405 reflects the brand sentiment
index values and the y-axis 1407 reflects the brands, products or
services. Holding the mouse over a bar in the graph displays the
year-to-date sentiment index 1409. Clicking on the bar in the graph
drills down to show the actual posts to online social media for the
brands, products or services. The chart displays year-to-date
sentiment by default, but users can select a narrower date range
1404 if desired.
[0088] FIG. 15 illustrates a smoothed view of a brand trend lines
display in a graphical user interface according to an exemplary
embodiment of the invention. In the illustrated embodiment, the
"brand sentiment" category is selected in the top-level menu 1503
and the "brand trend lines" category is selected in the submenu
1501. The x-axis 1505 reflects the brand sentiment index values and
y-axis 1507 reflects the months in selected date range 1509. This
is a graph that shows how sentiment for the brand and competition
has trended over time. Users can reference historical changes in
opinion to external events, campaigns, and et cetera. This can
enable back-testing on how campaigns have affected sentiment of
social media participants. Users may select a different date range
1509 to assess a narrower or different period of time. Selecting
the "trend line/detailed data" button 1521 toggles between the
trend line or "smoothed" view that enables easier viewing with no
jagged lines and the "detailed data" view which shows all the peaks
and valleys rather than smoothing the graph. The detailed data view
is shown in FIG. 16 which illustrates a detailed view of a brand
trend lines display according to an exemplary embodiment of the
invention. Mousing over lines at month intersections displays the
sentiment index 1511 for that month. Clicking on lines at month
intersections allows users to view the actual posts for that month.
Users may view the positive or negative post content for that month
depending upon whether sentiment was positive or negative for that
month. This capability allows users to assess opinions at a
particular point in time and ascertain why sentiment was trending
in a particular way.
[0089] FIG. 17 illustrates a brand sentiment by source menu display
in a graphical user interface according to an exemplary embodiment
of the invention. In the illustrated embodiment, the "brand
sentiment" category is selected in the top-level menu 1703 and the
"brand sentiment by source" category is selected in the submenu
1701. The x-axis 1705 reflects the brand opinion value by source
and the y-axis 1707 reflects the sources. This is a bar chart
showing sentiment indices for the brand by source grouping so users
can see how sentiment various by online social media sites. Source
groupings may be selected using drop-down menu 1711. By default the
drop-down menu includes most active, most positive, and most
negative source groups for the brand and competitors. In at least
certain embodiments, source groups are user-configurable to give
flexibility to create appropriate groupings so users can select a
different source group and/or brand to view how sentiment differs.
For example, it might be valuable to define source groups such as
"mainstream media blogs," "industry forums," "fan sites," and et
cetera. Mousing over a bar displays a sentiment index value 1713
for that source. Clicking on a bar takes the user a level deeper to
display sentiment indices for the brand's products or services for
that particular source as depicted in FIG. 18.
[0090] FIG. 18 illustrates a display of sentiment indices for a
brand's products or services for a particular source in a graphical
user interface according to an exemplary embodiment of the
invention. In the illustrated embodiment, the "brand sentiment"
category is selected from the top-level menu 1803 and the "brand
sentiment by source" category is selected from the submenu 1801
(even though the sentiment indices for the brand's products or
services for a particular source are displayed). The x-axis 1805
reflects the product or service sentiment for that particular
source and the y-axis 1807 reflects the products or services. Users
may select a different brand 1813 to view how sentiment differs
depending on the source. Mousing over a bar in the display shows
the numeric sentiment index value 1811 for that product or service
for the particular source in the associated date range 1809.
Clicking on a bar in the display drills down to a listing of the
online social media posts specific to the product or service and to
the source. That is, only the posts from the particular source
relating to that particular product or service are listed.
[0091] FIG. 19 illustrates a brand source trends for a particular
source group display in a graphical user interface according to an
exemplary embodiment of the invention. In the illustrated
embodiment, the "brand sentiment" category is selected in the
top-level menu 1903 and the "brand source trends" category is
selected in the submenu 1901. The x-axis 1905 reflects the brand
sentiment index by source and the y-axis 1907 reflects the months
in the selected date range 1909. This line chart shows how
sentiment has trended over time based on the selected source group
1911. Users are able to analyze whether opinion has changed for a
particular source group and research the online social media
conversations to try and determine the causes. Users can also view
the chart for competitors and selected a different date range 1909
for the chart. Mousing over lines at month intersections displays
the sentiment index 1913 for that month for that source. Clicking
on lines at month intersections drills down to the actual text of
the online social media posts for the brand for the month from the
particular source (drills down to positive or negative post content
for that month depending on whether sentiment was mostly positive
or negative for that month). This capability allows users to assess
opinions at a particular point in time and ascertain why sentiment
was trending a particular way for a particular source.
[0092] FIG. 20 illustrates a positive/negative posts display in a
graphical user interface according to an exemplary embodiment of
the invention. In the illustrated embodiment, the "brand sentiment"
category is selected in the top-level menu 2003 and the
"positive/negative posts" category is selected in the submenu 2001.
The x-axis 2005 reflects the number of posts per month and the
y-axis 2007 reflects the months in the selected date range 2009 for
the selected product or service 2011. This is a bar chart that
shows the distribution of positive, negative and neutral posts per
month. Users can see very quickly if there have been changes in the
distribution of opinion from month-to-month for the users' products
or services and those of their competitors. Mousing over the
different sections of the bar in the display shows the number of
positive, negative or neutral posts for that month along with the
percentage representing the monthly total 2013. Clicking on the
positive, negative or neutral section of a bar drills down to the
positive, negative or neutral posts post content for that month so
that users can assess what people are saying about the particular
product or service at that time.
[0093] FIG. 21 illustrates an example ad hoc sentiment trend chart
in a custom query display in a graphical user interface according
to an exemplary embodiment of the invention. In the illustrated
embodiment, the "brand sentiment" category is selected in the
top-level menu 2103 and the "custom query" category is selected in
the submenu 2101. The x-axis 2105 reflects the brand sentiment
index value and the y-axis 2107 reflects the months in the selected
date range. Custom query allows users to generate an ad hoc
sentiment trend chart for a specific set of brands, products and/or
services 2109 over a particular time period. This gives users the
flexibility to report the trends of fewer, more or different
brands, products or services.
[0094] FIG. 22 illustrates a products or services sentiment display
in a graphical user interface according to an exemplary embodiment
of the invention. In the illustrated embodiment, the "product
sentiment" category is selected in the top-level menu 2203 and the
"product sentiment" category is selected in the submenu 2201. The
x-axis 2205 reflects the brand sentiment index value and the y-axis
2207 reflects the brands, products or services for the selected
brand 2211 in the selected date range 2209. This bar chart compares
sentiment indices for a brand's products or services. Providing
measurements for products or services gives users a more
granular-level of sentiment analysis so that users can easily see
whether there are differing opinions about the brand's products or
services. Users can also view the chart for competitors to see how
their products/services sentiment compares. Mousing over a bar in
the display allows users to see the numeric sentiment index values
2213 and clicking on a bar in the display drills down to positive,
negative or neutral post content about the product or service.
[0095] FIG. 23 illustrates a products or services trend lines
display in a graphical user interface according to an exemplary
embodiment of the invention. In the illustrated embodiment, the
"product sentiment" category is selected in the top-level menu 2303
and the "product trend lines" category is selected in the submenu
2301. The x-axis 2305 reflects the brand sentiment index value and
the y-axis 2307 reflects the months in the selected date range 2309
for the selected brand 2311. This is a line chart that shows how
sentiment for the brand's products or services has trended over
time. Users can quickly analyze how events, campaigns, and et
cetera have impacted opinions about their products or services.
Users can also view the chart for competitors and select a
different date range 2309 for viewing. Mousing over lines at month
intersections displays a sentiment index for that month 2313 and
clicking on lines at month intersections drills down to positive,
negative or neutral post content about that product or service for
that month. This capability allows users to assess opinions at a
particular point in time and ascertain why sentiment was trending
in a particular direction.
[0096] FIG. 24 illustrates a products or services sentiment by
source display in a graphical user interface according to an
exemplary embodiment of the invention. In the illustrated
embodiment, the "product sentiment" category is selected in the
top-level menu 2403 and the "product sentiment by source" category
is selected in the submenu 2401. The x-axis 2405 reflects the brand
sentiment index value by source and the y-axis 2407 reflects the
selected sources. The brand sentiment index value by source may be
displayed for a selected date range 2409 for a selected product or
service 2415 and a selected group of sources 2411. This is a bar
chart showing sentiment indices for the brand's products or
services by source group so users can see how sentiment varies by
online sites. By default the source groups 2411 include most
active, most positive and most negative source groups for the brand
and its competitors. In one embodiment, source groups may be
configurable to give flexibility to create appropriate groupings.
For example, it might be valuable to create source groups such as
"main stream media blogs," "industry forums," "fan sites," and et
cetera. Users can also view the chart for competitors and select
different date ranges 2409 for viewing. Mousing over a bar displays
a sentiment index for that source for that particular product or
service 2413 for the associated date range. Clicking on a bar
drills down for a closer look at the sentiment indices for the
brand's products or services for that particular source.
[0097] FIG. 25 illustrates a products or services source trends
display in a graphical user interface according to an exemplary
embodiment of the invention. In the illustrated embodiment, the
"product sentiment" category is selected in the top-level menu 2503
and the "product source trends" category is selected in the submenu
2501. The x-axis 2505 reflects the brand sentiment index value by
source and the y-axis 2507 reflects the months in the selected date
range 2509. The brand sentiment index value by source may be for a
selected brand 2511, product/service 2513 and a selected group of
sources 2515. This line chart report shows how sentiment has
trended over time for a brand's products or services based on
source group. Users can also view the chart for competitors, select
particular product or service 2513 and selected different date
range 2509 for the trend report. Mousing over lines at month
intersections displays a sentiment index 2517 for that month for
that source for the selected product or service. Clicking on lines
at month intersections takes the user to positive or negative post
content for that month for that source for the selected product or
service.
[0098] FIG. 26 illustrates a share of voice display in a graphical
user interface according to an exemplary embodiment of the
invention. In the illustrated embodiment, the "share of voice"
category is selected in the top-level menu 2603 and the
"percentages" category is selected in the submenu 2601. This is a
pie chart showing how much conversations in the online social media
are talking about this set of brands relative to each other for the
date range 2605. For example, section 2607 of the pie chart in FIG.
26 indicates that 54.77% of the volume of online social media
conversations about the brands shown for the month of October 2008
refers to Audi. Users can quickly see if their volume of mentions
in online social media is high or low in comparison to the
competition and can view the chart for a different month for
comparison. Clicking on a section of the chart takes users to the
newest posts about that brand.
[0099] FIG. 27 illustrates a share of voice trends display in a
graphical user interface according to an exemplary embodiment of
the invention. In the illustrated embodiment, the "share of voice"
category is selected in the top-level menu 2703 and the "share
trends" category is selected in the submenu 2701. The x-axis 2705
reflects the volume of voice value and the y-axis 2707 reflects the
months in the selected date range 2709. This line chart report
shows how share of voice for the brand and competitors have trended
over time. Users are able to quickly see if they are gaining or
losing online share of voice. Clicking on lines at month
intersections drills down to the actual text of the online social
media post content for that month so users can assess opinions at
the particular point in time they had a particular share of
voice.
[0100] FIG. 28 illustrates a volume trends display in a graphical
user interface according to an exemplary embodiment of the
invention. In the illustrated embodiment, the "share of voice"
category is selected in the top-level menu 2803 and the "volume
trends" category is selected in the submenu 2801. The x-axis 2805
reflects the number of posts per month and the y-axis 2707 reflects
the months in the selected date range 2709. This line chart report
shows how volume of postings for the brand and competitors has
trended over time. Users can see how post volume has reacted to
events, programs, and et cetera over time. Clicking on lines at
month intersections takes users to post content for that month so
they can assess opinions at the particular point in time they had a
particular post volume.
[0101] FIG. 29 illustrates a topic radar plot display in a
graphical user interface according to an exemplary embodiment of
the invention. In the illustrated embodiment, the "topics" category
is selected in the top-level menu 2903 and the "tag radar" category
is selected in the submenu 2901. This is a visualization of terms,
concepts and competitors most frequently mentioned in online posts
in conjunction with the users' brand. The closer words appear
(e.g., BMW 2909 and Mercedes 2907) to the center where the users'
brand is located (e.g., Audi 2905), the more frequently they are
mentioned in conjunction with the brand. These are the words online
authors are employing in their actual posts. Brands can leverage
these words in creating messaging and communications and in search
engine keyword purchases, for example. Brand, product or service
managers can utilize these to see which competitors are most often
mentioned along with the brand. Additionally, users' customer
service departments can monitor whether terms such as "problem,"
"issue," and et cetera are appearing frequently in conjunction with
the users' brand. Clicking on a term in the topic radar plot
display takes the user to post content containing the brand and
words so users can see how they are used in context. Users can also
view topic radar for different months in the past by changing the
year and month selection to the desired date range. This can enable
users to see how terms used online have changed over time and
correlated those changes to events such as new advertising
campaigns or other external forces.
[0102] FIG. 30 illustrates a tag cloud display in a graphical user
interface according to an exemplary embodiment of the invention. In
the illustrated embodiment, the "topics" category is selected in
the top-level menu 3003 and the "tag cloud" category is selected in
the submenu 3001. This is a visualization that displays the same
data as topic radar in tag cloud format for a selected product or
service 3005. The larger the words are in the tag cloud (e.g., BMW
3009 and Mercedes 3007), the more frequently they are mentioned in
conjunction with the selected brand (e.g., Audi 3005). As with the
topic radar, users can view the chart for competitors and click on
terms to see the post content for the brand and the term(s).
[0103] FIG. 31 illustrates a products or services share of voice
trends display in a graphical user interface according to an
exemplary embodiment of the invention. In the illustrated
embodiment, the "topics" category is selected in the top-level menu
3103 and the "product trends" category is selected in the submenu
3101. The x-axis 3105 reflects the percentage of posts per month
for selected products or services 3121 and the y-axis 3107 reflects
the months in the selected date range. This is a bar chart
comparing frequency of mention of a brand's products or services
relative to each other over time (e.g., 3109, 3111, 3113, 3115,
3117, and 3119). Users can quickly see how participation of their
product/service in online social media conversations and that of
their competitors change and compare from month-to-month. This
provides users with insight into how campaigns and programs
promoting particular products or services are affecting online
posts. Clicking on the chart takes users to a list of post content
for the selected product or service for that particular month. The
same information can be obtained with regard to various selected
features using the "feature trends" category in the submenu 3101.
This is likewise a bar chart comparing the frequency of mention of
the features of a product or service relative to each other over
time so that users can quickly see how feature mentions and those
of their competitors change and compare from month-to-month.
[0104] FIG. 32 illustrates a custom query for topics display in a
graphical user interface according to an exemplary embodiment of
the invention. In the illustrated embodiment, the "topics" category
is selected in the top-level menu 3203 and the "custom query"
category is selected in the submenu 3201. The x-axis 3205 reflects
the percentage of posts per month for various custom selected
topics (e.g., products, services, and/or features) and the y-axis
3207 reflects the months in the selected date range. This is a bar
chart comparing frequency of mention of a brand's products,
services and/or features relative to each other over time (e.g.,
3207, 3209, and 3211). Custom query allows users to generate an ad
hoc trend bar chart report for a specific set of terms, concepts or
brands over a particular time period. This is the same type of
report generated in the product/service and feature trends in that
it compares frequency of mention of terms, concepts or brands in
the query relative to each other over time. Users can generate ad
hoc trend charts by entering terms and selecting a date range for
the report. In custom query, users can enter any terms that they
are interested in for a closer analysis.
[0105] FIG. 33 illustrates a forum opinion leader list display in a
graphical user interface according to an exemplary embodiment of
the invention. In the illustrated embodiment, the "opinion leaders"
category is selected in the top-level menu 3303 and the "forum
opinion leaders" category is selected in the submenu 3301. This
report is a list of most influential forum users or other online
social media authors for the category (e.g., automotive, computers,
financial services, etc.) sorted by importance. The importance is
donated by the centrality values generated during the social
network analysis processing discussed previously, which leverages
the social graph to determine the influence of online users. The
online social media users' preferred brands, home websites,
demographics, and that number of posts are also displayed. Users
can drill down into the posts and brand list for the influencer.
These drill-downs provide users with the capability to assess what
these influencers are saying online. Also, the opinion leaders list
can be filtered to show only opinion leaders who post about the
users' brand more than others.
[0106] Additionally, a listing of the top 10 most positive and top
10 most negative users for the brand can be displayed using the
"positive/negative users" category of submenu 3301. This enables
users to see who has the highest opinion of the brand and who has
the lowest. As with the opinion leaders list, users can drill down
into posts and brand information for these authors of online social
media posts. This list can show users who are the most positive
online social media authors that could be a potential source of
feedback and who are the most negative online social media authors
that might need extra customer service attention. Likewise, a list
of blogs with posts about the category sorted by ranking can be
displayed using the "influential blogs" category of submenu 3301.
Here, users of the GUI can see which blogs have the highest
influence with respect to the user's brands.
[0107] FIG. 34 illustrates an overall brand advocacy display in a
graphical user interface according to an exemplary embodiment of
the invention. In the illustrated embodiment, the "opinion leaders"
category is selected in the top-level menu 3403 and the "brand
advocacy" category is selected in the submenu 3401. The x-axis 3405
reflects the brand sentiment index value and the y-axis 3407
reflects the number of brand advocates. Also the share of voice is
represented by the size of the plots in the chart (e.g., 3411,
3413, and 3415). This is a chart showing how the brand and
competitors compare based on sentiment, number brand advocates and
share of voice. Thus, brand advocacy is essentially a
representation of the activity and focus of the brand's "fans."
This chart shows users whether their brand sentiment is higher or
lower than the competition, whether there are larger or smaller
numbers of brand advocates than the competition, and whether the
brand has a larger or smaller share of voice. For example, a brand
could have a good sentiment index, but lower number a brand
advocates and share of voice indicating that their fans are
positive, but not extremely active.
[0108] In addition, users may select the "top websites" category in
the top-level menu. This will display a list of the websites users
are most likely to reach when searching online for information
about a user's brand, product or service. This feature allows users
to sort top websites by importance, site name or sites without
advertising. As with the opinion leader list, the centrality metric
for top websites reflects importance. In this case, the centrality
represents the likelihood of users reaching the site when searching
for information about the users' brands, products or services.
Users can then click on the URL to launch the site for reference
and examination. This list can be used to confirm the best sites
for messaging, advertisement and engagement, which can illuminate
sites toward the top of the list (important) that have not been
utilized and those toward the bottom of the list (unimportant)
where valuable dollars are being expended. The list shows: the
advertising vehicle on the site (if any); the number of unique
users; if there is any social media on the site; and the centrality
metric (importance) of the site. Users may also select the
"reports" category in the top-level menu. This list shows alerts
that have been triggered based on user-configuration. For example,
alerts can be sent for: extremely positive or negative posts;
sentiment index changes; high volume of issues mentioned in posts;
posts for particular authors users wish to track; posts for
specific sites; and posts containing specific keywords. In one
embodiment, users can receive these alerts via e-mail or SMS
notifications.
[0109] Embodiments provide methods, apparatuses, and
computer-readable medium for harvesting, aggregating, and providing
analytic measurements of unstructured qualitative online social
media conversations including the sentiment expressed among online
social media participants about a particular subject matter. The
type of subject matter that can be harvested, aggregated and
provided as analytic measurements is virtually limitless as any
subject matter contained in social media postings is envisioned to
be within the scope of this description. Likewise, the applications
of the SMA platform is virtually limitless does any use of
aggregated and quantified social media conversations is envisioned
to be within the scope of this description. Some of the
applications of the SMA platform include: providing enhanced target
advertising campaigns; providing enhanced customer service at a
call-center; providing enhanced market research; providing a method
of improved product development; providing an enhanced method for
generating opinion polls; and providing enhanced methods for
National Defense intelligence to name a few.
[0110] FIG. 35 illustrates an exemplary data processing system upon
which the methods and apparatuses of the invention may be
implemented. Note that while FIG. 35 illustrates various components
of a data processing system, it is not intended to represent any
particular architecture or manner of interconnecting the components
as such details are not germane to the present invention. It will
also be appreciated that network computers and other data
processing systems which have fewer components or perhaps more
components may also be used. The data processing system of FIG. 35
may, for example, be a workstation, or a personal computer (PC)
running a Windows operating system, or an Apple Macintosh
computer.
[0111] As shown in FIG. 35, the data processing system 3501
includes a system bus 3502 which is coupled to a microprocessor
3503, a ROM 3507, a volatile RAM 3505, and a non-volatile memory
3506. The microprocessor 3503, which may be a processor designed to
execute any instruction set, is coupled to cache memory 3504 as
shown in the example of FIG. 35. The system bus 3502 interconnects
these various components together and also interconnects components
3503, 3507, 3505, and 3506 to a display controller and display
device 3508, and to peripheral devices such as input/output (I/O)
devices 3510, such as keyboards, modems, network interfaces,
printers, scanners, video cameras and other devices which are well
known in the art. Typically, the I/O devices 3510 are coupled to
the system bus 3502 through input/output controllers 3509. The
volatile RAM 3505 is typically implemented as dynamic RAM (DRAM)
which requires power continually in order to refresh or maintain
the data in the memory. The non-volatile memory 3506 is typically a
magnetic hard drive or a magnetic optical drive or an optical drive
or a DVD RAM or other type of memory systems which maintain data
even after power is removed from the system. Typically, the
non-volatile memory 3506 will also be a random access memory
although this is not required. While FIG. 35 shows that the
non-volatile memory 3506 is a local device coupled directly to the
rest of the components in the data processing system, it will be
appreciated that the present invention may utilize a non-volatile
memory which is remote from the system, such as a network storage
device which is coupled to the data processing system through a
network interface such as a modem or Ethernet interface (not
shown). The system bus 3502 may include one or more buses connected
to each other through various bridges, controllers and/or adapters
(not shown) as is well known in the art. In one embodiment the I/O
controller 3509 includes a USB (Universal Serial Bus) adapter for
controlling USB peripherals, and/or an IEEE-1394 bus adapter for
controlling IEEE-1394 peripherals.
[0112] It will be apparent from this description that aspects of
the present invention may be embodied, at least in part, in
software, hardware, firmware, or in combination thereof. That is,
the techniques may be carried out in a computer system or other
data processing system in response to its processor, such as a
microprocessor, executing sequences of instructions contained in a
memory, such as ROM 3507, volatile RAM 3505, non-volatile memory
3506, cache 3504, or a remote storage device (not shown). In
various embodiments, hardwired circuitry may be used in combination
with software instructions to implement the present invention.
Thus, the techniques are not limited to any specific combination of
hardware circuitry and software or to any particular source for the
instructions executed by the data processing system 3500. In
addition, throughout this description, various functions and
operations are described as being performed by or caused by
software code to simplify description. However, those skilled in
the art will recognize that what is meant by such expressions is
that the functions result from execution of code by a processor,
such as the microprocessor 3503.
[0113] The invention also relates to apparatus for performing the
operations herein. This apparatus may be specially constructed for
the required purposes, or it may comprise a general purpose
computer selectively activated or reconfigured by a computer
program stored in the computer. Such a computer program may be
stored or transmitted in a computer-readable medium. A
computer-readable medium can be used to store software and data
which when executed by a data processing system, such as data
processing system 3500, causes the system to perform various
methods of the present invention. This executable software and data
may be stored in various places including for example ROM 3507,
volatile RAM 3505, non-volatile memory 3506, and/or cache 3504 as
shown in FIG. 35. Portions of this software and/or data may be
stored in any one of these storage devices. A computer-readable
medium may include any mechanism that provides (i.e., stores and/or
transmits) information in a form accessible by a machine (e.g., a
computer, network device, personal digital assistant, manufacturing
tool, any device with a set of one or more processors, etc.). For
example, a machine readable medium includes
recordable/non-recordable media such as, but not limited to, a
computer-readable storage medium (e.g., any type of disk including
floppy disks, optical disks, CD-ROMs, and magnetic-optical disks,
read-only memories (ROMs), random access memories (RAMs), EPROMs,
EEPROMs, flash memory, magnetic or optical cards, or any type of
media suitable for storing electronic instructions), or a
computer-readable transmission medium such as, but not limited to,
any type of electrical, optical, acoustical or other form of
propagated signals (e.g., carrier waves, infrared signals, digital
signals, etc.).
[0114] Additionally, it will be understood that the various
embodiments described herein may be implemented with data
processing systems which have more or fewer components than system
3500. For example, such data processing systems may be a cellular
telephone or a personal digital assistant (PDA) or an entertainment
system or a media player or a consumer electronic device, and et
cetera, each of which can be used to implement one or more of the
embodiments of the invention. The algorithms and displays presented
herein are not inherently related to any particular computer system
or other apparatus. Various general purpose systems may be used
with programs in accordance with the teachings herein, or it may
prove convenient to construct more specialized apparatuses to
perform the method operations. The structure for a variety of these
systems appears from the description above. In addition, the
invention is not described with reference to any particular
programming language. It will be appreciated that a variety of
programming languages may be used to implement the teachings of the
invention as described herein.
[0115] Throughout the foregoing specification, references to "one
embodiment," "an embodiment," "an example embodiment," and et
cetera, indicate that the embodiment described may include a
particular feature, structure, or characteristic, but every
embodiment may not necessarily include the particular feature,
structure, or characteristic. Moreover, such phrases are not
necessarily referring to the same embodiment. When a particular
feature, structure, or characteristic is described in connection
with an embodiment, it is submitted that it is within the knowledge
of one skilled in the art to bring about such a feature, structure,
or characteristic in connection with other embodiments whether or
not explicitly described. Various changes may be made in the
structure and embodiments shown herein without departing from the
principles of the invention. Further, features of the embodiments
shown in various figures may be employed in combination with
embodiments shown in other figures.
[0116] In the description as set forth above and claims, the terms
"coupled" and "connected," along with their derivatives, may be
used. It should be understood that these terms are not intended to
be synonymous with each other. Rather, in particular embodiments,
"connected" is used to indicate that two or more elements are in
direct physical or electrical contact with each other. "Coupled"
may mean that two or more elements are in direct physical or
electrical contact. However, "coupled" may also mean that two or
more elements are not in direct contact with each other, but yet
still co-operate or interact with each other.
[0117] Some portions of the detailed description as set forth above
are presented in terms of algorithms and symbolic representations
of operations on data bits within a computer memory. These
algorithmic descriptions and representations are the means used by
those skilled in the data processing arts to most effectively
convey the substance of their work to others skilled in the art. An
algorithm is here, and generally, conceived to be a self-consistent
sequence of operations leading to a desired result. The operations
are those requiring physical manipulations of physical quantities.
Usually, though not necessarily, these quantities take the form of
electrical or magnetic signals capable of being stored,
transferred, combined, compared, and otherwise manipulated. It has
proven convenient at times, principally for reasons of common
usage, to refer to these signals as bits, values, elements,
symbols, characters, terms, numbers, or the like.
[0118] It should be borne in mind, however, that all of these and
similar terms are to be associated with the appropriate physical
quantities and are merely convenient labels applied to these
quantities. Unless specifically stated otherwise as apparent from
the discussion as set forth above, it is appreciated that
throughout the description, discussions utilizing terms such as
"processing" or "computing" or "calculating" or "determining" or
"displaying" or the like, refer to the action and processes of a
computer system, or similar electronic computing device, that
manipulates and transforms data represented as physical
(electronic) quantities within the computer system's registers and
memories into other data similarly represented as physical
quantities within the computer system memories or registers or
other such information storage, transmission or display
devices.
[0119] Embodiments of the invention may include various operations
as set forth above or fewer operations or more operations or
operations in an order which is different from the order described
herein. The operations may be embodied in machine-executable
instructions which cause a general-purpose or special-purpose
processor to perform certain operations. Alternatively, these
operations may be performed by specific hardware components that
contain hardwired logic for performing the operations, or by any
combination of programmed computer components and custom hardware
components.
[0120] Throughout the foregoing description, for the purposes of
explanation, numerous specific details were set forth in order to
provide a thorough understanding of the invention. It will be
apparent, however, to one skilled in the art that the invention may
be practiced without some of these specific details. Accordingly,
the scope and spirit of the invention should be judged in terms of
the claims which follow as well as the legal equivalents
thereof.
* * * * *