Final Report Summary - PR2.0 (Integrated reputation and social media management hub to help enterprises manage dispersed online public information)
Executive summary:
Reputation management has always been a concern for companies and has traditionally been handled by public relations and communications firms, which monitored mentions of their clients and their respective brands in print and audiovisual media. As the importance of the Internet has become more apparent, most of these firms have turned to monitoring digital media as well, scouring the Internet daily for clips related to the client or the clients brand, with Google as their most powerful weapon. What most firms - particularly Small and medium-sized enterprise (SME)s - were not prepared for was the momentum that would be gained by user-generated content (such as blogs) and social media (such as Facebook, YouTube, etc.).
In addition to content being created by users, companies are increasingly using social media applications to market to their customers and potential customers, creating their own content on sites like Facebook and Twitter, in an effort to build and maintain brand loyalty, create affinity among consumers and keep their marketing message top of mind. Currently, most businesses carry out their social media campaigns manually; that is, with an internal communications professional or an outside public relations and communications firm handling the creation of content, posting it to the various social media sites and following up on the progress of the campaign. Managing a campaign across several digital media channels is cumbersome and time consuming, and SMEs often do not have the resources to assign a designated person to such tasks. As such, the use of social media - a medium that can really work for small businesses - proves to be elusive to SMEs strapped for resources.
The main objective of the PR2.0 has been to develop a software tool that allows users not only to visualise and analyse data but also adds intelligent analysis modules related to natural language processing techniques. These intelligent modules are necessary since the huge amount of data generated nowadays makes unviable the human analysis of raw data.
The project has finalised with a prototype that has proven to be useful in the analysis of Twitter and Facebook posts in English. However, the high number of competing tools that have appeared during the project executing has led the consortium SMEs to consider alternative paths for the exploitation of the results involving new languages and sectors.
Project context and objectives:
Project objectives per Work package (WP) are:
WP1: System definition:
- to detect and define the end-users requirements that the system will cover;
- to define the system specifications in order to satisfy the user requirements previously detected;
- to define the use cases that will be used in the validation stage of the project to ensure that the results are successful.
WP2: Sources of information and access methods:
- to establish which are the sources from which the system will extract the information needed for the monitoring of digital reputation. The sites and applications that the system will access to allow users to manage profile pages, user accounts and marketing campaigns on social media will be also identified in this WP;
- to detect existing libraries and Application programming interface (API)s that may be useful for the project development;
- to identify potential problems related to the legal aspects that surround the automatic access to the content of social media;
- to develop a library that will contain the methods that will be used by the information extraction system and the management system in order to access the different social media covered by the PR2.0 system.
WP3: Development of a highly usable Graphical user interface (GUI):
- to detect the main recommendations that should be taken into account for the development of the GUI used by the PR2.0 system;
- to develop a GUI that, fed by static data, allows the evaluation of the usability of the GUI;
- to test the usability of the developed GUI using different techniques.
WP4: Information processing system:
The main objective of this WP is to develop the different software modules that in conjunction will allow the monitoring of the digital reputation of companies, brands and public persons.
The various modules to be developed are:
- a search engine that will consist in a crawler that will search in the sources of information allowed by the access library developed in task 2.3 and an information extraction algorithm that will filter this information for user access or for the opinion recognition module;
- an automatic opinion recognition module will analyse the information retrieved by the Information extraction module (IEM) for early detection of trends in the public opinion of brands or public persons;
- a Blackboard system that will allow distributing the search and analysis tasks among different instances (and servers) that will cooperate in monitoring the analysing the data required by the end user.
WP5: Management system:
The main objective of this task is to develop all the business logic that, together with the results of WP4 will be accessed by the end-user by the GUI developed in WP3.
- the module that will allow the end-user to manage and modify its information (or its customer's information) in different social media will be developed under this WP;
- all the communication features needed for real-time warnings will be also developed under this WP;
- as a last objective, the results of this module will be integrated with the previously developed GUI.
WP6: Integration and deployment:
- verify that the communication among modules is efficient;
- verify that the Blackboard system guarantees that the system is scalable in terms of system requirements;
- identify and automate the installation procedures.
WP7: Validation:
The objective of this WP is to proceed to validate the system using the use cases defined in WP1 to measure the achievements of the project objectives. All the partners will be implicated in this WP. The Research and technological development (RTD) will be solving all the technical problems that may be detected during the WP as a result of the validation, the - SMEs SOCIAL will be using the application following the use cases and will report any problem that they detect. The technical SMEs ASSIST and GNOME will be also managing the application and, also according to the use cases, will implement some minor modifications of the application in order to validate that the system developed is fully understood by them and guarantee that in an exploitation stage they are able to accomplish all the needed technical tasks.
WP8: Training and dissemination:
The general objective of this WP will be to facilitate the take-up of results by the SMEs, in particular by training and dissemination, in order to:
- ensure that the SME participants will be able to assimilate the results of the project;
- disseminate activities beyond the consortium to a wider audience.
WP9: Preparing for exploitation:
- support the participating SMEs in protecting and using the research results to their best advantage;
- exploit the results and increase the competitiveness of the SME participants.
WP10: Consortium management:
- to review and assess compliance by the partners and subcontracted RTD performers with their contractual obligations under the project;
- to ensure that all aspects of the European Commission (EC) requirements for communication and reporting are met and to process any recommendations made by EC reviewers;
- to prepare and execute consortium meetings and to process the outcome of the meetings;
- to update the consortium agreement, if and where necessary.
Project results:
The project results by WP can be summarised as follows:
In the scope of WP1 the user requirements were identified. In line with these user requirements, the system specifications and the use cases used in the final validation of the system were defined. The output of this WP was used in all subsequent software developments in the project. The decisions made in this process in relation to the system architecture were especially important due to the amount of different technologies and modules involved. The use of web services as a main communications channel between the Information Processing Module instances and the Management Module and the use of a common database as a shared repository for search results provided by all modules were the most relevant decisions.
WP2 was focused in the sources of information targeted by the project. The main outputs were a study on available methods and conditions under which you can access them (D2.1) and the software layer allowing this access, the access library (D2.2 and D2.3). The main conclusions of this WP were that the access methods available (APIs) and even the sources of information themselves change rapidly. The external reviewer recommended during the midterm review to foster the creation of an open source community around this module in order to keep the module updated. The consortium has identified existing open source communities that are maintaining similar modules. These modules have been used in the development of the Access Library.
WP3 focused in the development of the system GUI. Several studies to evaluate the acceptability and usefulness of the different prototypes were executed during the execution of this WP. This fact allowed the implementation of constant new release that were approaching the user requirements and preferences. The final version of the user interface was implemented over the early integrated prototype.
The information processing system was developed in WP4. This system is formed by different modules (crawler, text pre-processing, sentiment analysis, topic segmentation and black board system) that in cooperation search for keyword combinations in the targeted social networks (using the Access Library) and analyse (providing topic and sentiment) the retrieved result coordinating, if necessary, the execution of different instances of the information processing system. The execution of this WP provided as main conclusions that the use of linguistic rules for extracting sentiment from opinions provides better results statistical methods and potential future improvements. In the case of the Blackboard system, a simple system for coordinating the execution of different instances was developed based on sharing a common database among all implied modules. Although this approach has been successful during the project execution, future improvements focused on dynamic and automatic balance of tasks among all IPS instances are suggested as future improvements.
WP5 focused in the development of the management system. This management system, integrated with the GUI developed in WP3 formed a web application that is the interface between the PR2.0 system and the end user. Every process of the PR2.0 application was programmed and integrated into the management module. The alerts module was the finale development of the management system. This module checks periodically the activity found by the IPS and if necessary send alerts to the end user by the specified channels.
WP6 consisted in the integration and deployment of all developed modules. To that end, the different IPS modules were encapsulated as web services or scheduled jobs. The web services are used for direct communication with the management module. This WP was executed in parallel with the development of the different modules allowing the release of different versions during the execution of the project.
The validation process was executed in WP7. A first functional verification by the RTD partners was followed by a deep validation process in which all partners and two external partners took part. In parallel with this end user validation process ITT focused on the evaluation of the performance of the final version sentiment analysis engine and together with MIIS in the correction of error found for the production of the system final version.
Training and dissemination activities took place in the scope of WP8. Main dissemination activities included the maintenance of the project website and the dissemination of the project in several event in which the partners assisted. These events were not only used to disseminate the project results but also to prepare the future exploitation of the system. In terms of training, several training events took place during the project. In these events (supported with training material specially created) the system architecture, technologies, use and deployment were exposed.
The main aim of WP9 was to prepare the future exploitation of the system. To this end a constant monitoring of competing systems and potential sectors of application was executed during the whole project. The main output of this WP is a plan for use and dissemination for the foreground in which the main conclusions are specified.
Potential impact:
The main objective of the PR2.0 project is to develop an innovative tool to facilitate managing the digital reputation of brands and public people in social media by:
(i) adapting the newest techniques in semantic analysis to the social media environment,
(ii) using a multi-agent architecture to make the system able to process the huge amount of information created on these sources, and
(iii) building a Graphical User Interface (GUI) based on the user needs and skills able to handle and display the a mount of information in an intuitive way.
The reputation management tool comprising the main result of the project features:
A graphical user interface by means of which the user will be able to monitor mentions of their company, brand or client, as well as administer social media campaigns: The interface will be simple and user friendly and provide the window that allows the user to view a large number of social media activities at the same time. The GUI will be modular, allowing the user to choose which modules he or she wants to see and in what way.
An information processing system: This system will be built using three different modules: a crawler, an information extraction system and an automatic opinion extraction system. The Crawler module will search for new information from different sources based on two parameters: the information should be related to the company, brand or public person whose reputation is being monitored and the information should be new, not previously processed. The crawler system will search the social media sources through an access library, which utilises the search tools provided by the social media sites (if available), such as Twitter API and Facebook API.
The IEM creates structured information by using semantic processing to extract the relevant information from each item (an item can be a blog entry, a comment on a video or any other written information). The semantic analysis will provide relevant sentences or words.
The Automatic opinion extraction module (AOEM) will analyse the records provided by the IEM as an output. The main objective of this module is to interpret the records produced by the IEM and classify them as being positive or negative.
A management system that will allow the user to manage all the information about its company, brand or customers present in social media and manage marketing campaigns on social media sites. This system will be the main tool that will allow the user to execute actions on the various social media services. This system will result from the integration of all the previous modules and the development of the business logic that will control the activity of the system and allow the user to:
- be updated on mentions of the brand or company in social media or the Internet at large through the information retrieved by the information processing module;
- easily update the company's presence on social media (YouTube, blogs, Facebook fan pages, etc.) from a unique and centralised interface using the access library;
- execute corrective actions and respond to customer concerns, including publishing responses in blog comments, YouTube videos or other media;
- provide statistics useful for painting a general picture of the user (or brand) image in the social network.
Besides the technical results, the project will develop training courses, content guidelines, and learning materials for promoting Web 2.0 tools among European SMEs.
The impact of the project will be foremost on the way SMEs will be empowered to manage their reputation in social media:
Impact on SME competitiveness: The PR 2.0 resulting systems ensures that SMEs take full advantage of the low-cost tools offered by social media applications by providing them with a comprehensive tool that allows them to effectively launch, maintain and monitor a social media campaign and monitor and manage their company's reputation online in order to preserve their company's good name or the good name of its clients.
Empowering SMEs to manage their own reputation: PR 2.0 allows SMEs to constantly monitor activity related to their business on social media sites and be instantly updated as to potential issues as soon as possible, which results in increased competitiveness through the early identification of customer-related problems, leading to a higher level of customer satisfaction. Competitiveness will also be increased through a more direct and personal communication with customers, which will differentiate the user's businesses or clients from the competitors who market themselves by talking 'at' the consumer rather than talking 'with' them.
Economic impact: The progress of the PR2.0 project has led to some interesting business opportunities for the partners. Consortium SMEs are not only planning the exploitation of the PR2.0 as a whole solution, but are also considering the creation of new tools for other sectors by reusing some of the modules produced. In addition, partners are also seriously exploring the possibility of exploitation via a PR2.0 API, which could power other applications, providing the competitive monitoring and analysis tools other companies may be looking to integrate into their own platforms. The SMEs are also looking at the possibility of adapting the whole tool or certain modules to each of their own languages, as well as other attractive European and international market languages in order to strengthen the possibility of competitiveness in Europe and beyond.
E-mail: Albert Rodríguez albert.rodriguez@cric.cat
Project website: http://pr20-project.eu
Reputation management has always been a concern for companies and has traditionally been handled by public relations and communications firms, which monitored mentions of their clients and their respective brands in print and audiovisual media. As the importance of the Internet has become more apparent, most of these firms have turned to monitoring digital media as well, scouring the Internet daily for clips related to the client or the clients brand, with Google as their most powerful weapon. What most firms - particularly Small and medium-sized enterprise (SME)s - were not prepared for was the momentum that would be gained by user-generated content (such as blogs) and social media (such as Facebook, YouTube, etc.).
In addition to content being created by users, companies are increasingly using social media applications to market to their customers and potential customers, creating their own content on sites like Facebook and Twitter, in an effort to build and maintain brand loyalty, create affinity among consumers and keep their marketing message top of mind. Currently, most businesses carry out their social media campaigns manually; that is, with an internal communications professional or an outside public relations and communications firm handling the creation of content, posting it to the various social media sites and following up on the progress of the campaign. Managing a campaign across several digital media channels is cumbersome and time consuming, and SMEs often do not have the resources to assign a designated person to such tasks. As such, the use of social media - a medium that can really work for small businesses - proves to be elusive to SMEs strapped for resources.
The main objective of the PR2.0 has been to develop a software tool that allows users not only to visualise and analyse data but also adds intelligent analysis modules related to natural language processing techniques. These intelligent modules are necessary since the huge amount of data generated nowadays makes unviable the human analysis of raw data.
The project has finalised with a prototype that has proven to be useful in the analysis of Twitter and Facebook posts in English. However, the high number of competing tools that have appeared during the project executing has led the consortium SMEs to consider alternative paths for the exploitation of the results involving new languages and sectors.
Project context and objectives:
Project objectives per Work package (WP) are:
WP1: System definition:
- to detect and define the end-users requirements that the system will cover;
- to define the system specifications in order to satisfy the user requirements previously detected;
- to define the use cases that will be used in the validation stage of the project to ensure that the results are successful.
WP2: Sources of information and access methods:
- to establish which are the sources from which the system will extract the information needed for the monitoring of digital reputation. The sites and applications that the system will access to allow users to manage profile pages, user accounts and marketing campaigns on social media will be also identified in this WP;
- to detect existing libraries and Application programming interface (API)s that may be useful for the project development;
- to identify potential problems related to the legal aspects that surround the automatic access to the content of social media;
- to develop a library that will contain the methods that will be used by the information extraction system and the management system in order to access the different social media covered by the PR2.0 system.
WP3: Development of a highly usable Graphical user interface (GUI):
- to detect the main recommendations that should be taken into account for the development of the GUI used by the PR2.0 system;
- to develop a GUI that, fed by static data, allows the evaluation of the usability of the GUI;
- to test the usability of the developed GUI using different techniques.
WP4: Information processing system:
The main objective of this WP is to develop the different software modules that in conjunction will allow the monitoring of the digital reputation of companies, brands and public persons.
The various modules to be developed are:
- a search engine that will consist in a crawler that will search in the sources of information allowed by the access library developed in task 2.3 and an information extraction algorithm that will filter this information for user access or for the opinion recognition module;
- an automatic opinion recognition module will analyse the information retrieved by the Information extraction module (IEM) for early detection of trends in the public opinion of brands or public persons;
- a Blackboard system that will allow distributing the search and analysis tasks among different instances (and servers) that will cooperate in monitoring the analysing the data required by the end user.
WP5: Management system:
The main objective of this task is to develop all the business logic that, together with the results of WP4 will be accessed by the end-user by the GUI developed in WP3.
- the module that will allow the end-user to manage and modify its information (or its customer's information) in different social media will be developed under this WP;
- all the communication features needed for real-time warnings will be also developed under this WP;
- as a last objective, the results of this module will be integrated with the previously developed GUI.
WP6: Integration and deployment:
- verify that the communication among modules is efficient;
- verify that the Blackboard system guarantees that the system is scalable in terms of system requirements;
- identify and automate the installation procedures.
WP7: Validation:
The objective of this WP is to proceed to validate the system using the use cases defined in WP1 to measure the achievements of the project objectives. All the partners will be implicated in this WP. The Research and technological development (RTD) will be solving all the technical problems that may be detected during the WP as a result of the validation, the - SMEs SOCIAL will be using the application following the use cases and will report any problem that they detect. The technical SMEs ASSIST and GNOME will be also managing the application and, also according to the use cases, will implement some minor modifications of the application in order to validate that the system developed is fully understood by them and guarantee that in an exploitation stage they are able to accomplish all the needed technical tasks.
WP8: Training and dissemination:
The general objective of this WP will be to facilitate the take-up of results by the SMEs, in particular by training and dissemination, in order to:
- ensure that the SME participants will be able to assimilate the results of the project;
- disseminate activities beyond the consortium to a wider audience.
WP9: Preparing for exploitation:
- support the participating SMEs in protecting and using the research results to their best advantage;
- exploit the results and increase the competitiveness of the SME participants.
WP10: Consortium management:
- to review and assess compliance by the partners and subcontracted RTD performers with their contractual obligations under the project;
- to ensure that all aspects of the European Commission (EC) requirements for communication and reporting are met and to process any recommendations made by EC reviewers;
- to prepare and execute consortium meetings and to process the outcome of the meetings;
- to update the consortium agreement, if and where necessary.
Project results:
The project results by WP can be summarised as follows:
In the scope of WP1 the user requirements were identified. In line with these user requirements, the system specifications and the use cases used in the final validation of the system were defined. The output of this WP was used in all subsequent software developments in the project. The decisions made in this process in relation to the system architecture were especially important due to the amount of different technologies and modules involved. The use of web services as a main communications channel between the Information Processing Module instances and the Management Module and the use of a common database as a shared repository for search results provided by all modules were the most relevant decisions.
WP2 was focused in the sources of information targeted by the project. The main outputs were a study on available methods and conditions under which you can access them (D2.1) and the software layer allowing this access, the access library (D2.2 and D2.3). The main conclusions of this WP were that the access methods available (APIs) and even the sources of information themselves change rapidly. The external reviewer recommended during the midterm review to foster the creation of an open source community around this module in order to keep the module updated. The consortium has identified existing open source communities that are maintaining similar modules. These modules have been used in the development of the Access Library.
WP3 focused in the development of the system GUI. Several studies to evaluate the acceptability and usefulness of the different prototypes were executed during the execution of this WP. This fact allowed the implementation of constant new release that were approaching the user requirements and preferences. The final version of the user interface was implemented over the early integrated prototype.
The information processing system was developed in WP4. This system is formed by different modules (crawler, text pre-processing, sentiment analysis, topic segmentation and black board system) that in cooperation search for keyword combinations in the targeted social networks (using the Access Library) and analyse (providing topic and sentiment) the retrieved result coordinating, if necessary, the execution of different instances of the information processing system. The execution of this WP provided as main conclusions that the use of linguistic rules for extracting sentiment from opinions provides better results statistical methods and potential future improvements. In the case of the Blackboard system, a simple system for coordinating the execution of different instances was developed based on sharing a common database among all implied modules. Although this approach has been successful during the project execution, future improvements focused on dynamic and automatic balance of tasks among all IPS instances are suggested as future improvements.
WP5 focused in the development of the management system. This management system, integrated with the GUI developed in WP3 formed a web application that is the interface between the PR2.0 system and the end user. Every process of the PR2.0 application was programmed and integrated into the management module. The alerts module was the finale development of the management system. This module checks periodically the activity found by the IPS and if necessary send alerts to the end user by the specified channels.
WP6 consisted in the integration and deployment of all developed modules. To that end, the different IPS modules were encapsulated as web services or scheduled jobs. The web services are used for direct communication with the management module. This WP was executed in parallel with the development of the different modules allowing the release of different versions during the execution of the project.
The validation process was executed in WP7. A first functional verification by the RTD partners was followed by a deep validation process in which all partners and two external partners took part. In parallel with this end user validation process ITT focused on the evaluation of the performance of the final version sentiment analysis engine and together with MIIS in the correction of error found for the production of the system final version.
Training and dissemination activities took place in the scope of WP8. Main dissemination activities included the maintenance of the project website and the dissemination of the project in several event in which the partners assisted. These events were not only used to disseminate the project results but also to prepare the future exploitation of the system. In terms of training, several training events took place during the project. In these events (supported with training material specially created) the system architecture, technologies, use and deployment were exposed.
The main aim of WP9 was to prepare the future exploitation of the system. To this end a constant monitoring of competing systems and potential sectors of application was executed during the whole project. The main output of this WP is a plan for use and dissemination for the foreground in which the main conclusions are specified.
Potential impact:
The main objective of the PR2.0 project is to develop an innovative tool to facilitate managing the digital reputation of brands and public people in social media by:
(i) adapting the newest techniques in semantic analysis to the social media environment,
(ii) using a multi-agent architecture to make the system able to process the huge amount of information created on these sources, and
(iii) building a Graphical User Interface (GUI) based on the user needs and skills able to handle and display the a mount of information in an intuitive way.
The reputation management tool comprising the main result of the project features:
A graphical user interface by means of which the user will be able to monitor mentions of their company, brand or client, as well as administer social media campaigns: The interface will be simple and user friendly and provide the window that allows the user to view a large number of social media activities at the same time. The GUI will be modular, allowing the user to choose which modules he or she wants to see and in what way.
An information processing system: This system will be built using three different modules: a crawler, an information extraction system and an automatic opinion extraction system. The Crawler module will search for new information from different sources based on two parameters: the information should be related to the company, brand or public person whose reputation is being monitored and the information should be new, not previously processed. The crawler system will search the social media sources through an access library, which utilises the search tools provided by the social media sites (if available), such as Twitter API and Facebook API.
The IEM creates structured information by using semantic processing to extract the relevant information from each item (an item can be a blog entry, a comment on a video or any other written information). The semantic analysis will provide relevant sentences or words.
The Automatic opinion extraction module (AOEM) will analyse the records provided by the IEM as an output. The main objective of this module is to interpret the records produced by the IEM and classify them as being positive or negative.
A management system that will allow the user to manage all the information about its company, brand or customers present in social media and manage marketing campaigns on social media sites. This system will be the main tool that will allow the user to execute actions on the various social media services. This system will result from the integration of all the previous modules and the development of the business logic that will control the activity of the system and allow the user to:
- be updated on mentions of the brand or company in social media or the Internet at large through the information retrieved by the information processing module;
- easily update the company's presence on social media (YouTube, blogs, Facebook fan pages, etc.) from a unique and centralised interface using the access library;
- execute corrective actions and respond to customer concerns, including publishing responses in blog comments, YouTube videos or other media;
- provide statistics useful for painting a general picture of the user (or brand) image in the social network.
Besides the technical results, the project will develop training courses, content guidelines, and learning materials for promoting Web 2.0 tools among European SMEs.
The impact of the project will be foremost on the way SMEs will be empowered to manage their reputation in social media:
Impact on SME competitiveness: The PR 2.0 resulting systems ensures that SMEs take full advantage of the low-cost tools offered by social media applications by providing them with a comprehensive tool that allows them to effectively launch, maintain and monitor a social media campaign and monitor and manage their company's reputation online in order to preserve their company's good name or the good name of its clients.
Empowering SMEs to manage their own reputation: PR 2.0 allows SMEs to constantly monitor activity related to their business on social media sites and be instantly updated as to potential issues as soon as possible, which results in increased competitiveness through the early identification of customer-related problems, leading to a higher level of customer satisfaction. Competitiveness will also be increased through a more direct and personal communication with customers, which will differentiate the user's businesses or clients from the competitors who market themselves by talking 'at' the consumer rather than talking 'with' them.
Economic impact: The progress of the PR2.0 project has led to some interesting business opportunities for the partners. Consortium SMEs are not only planning the exploitation of the PR2.0 as a whole solution, but are also considering the creation of new tools for other sectors by reusing some of the modules produced. In addition, partners are also seriously exploring the possibility of exploitation via a PR2.0 API, which could power other applications, providing the competitive monitoring and analysis tools other companies may be looking to integrate into their own platforms. The SMEs are also looking at the possibility of adapting the whole tool or certain modules to each of their own languages, as well as other attractive European and international market languages in order to strengthen the possibility of competitiveness in Europe and beyond.
E-mail: Albert Rodríguez albert.rodriguez@cric.cat
Project website: http://pr20-project.eu