Periodic Reporting for period 1 - BigDataScore (Improving loan quality and acceptance rates by predicting credit behavior through social mediadata.)
Periodo di rendicontazione: 2015-03-01 al 2015-06-30
Credit referencing is one of the pillars for financial systems and it is rapidly growing in importance. The primary purpose of credit referencing is to operate as a trusted third party holder of data which can be used by lenders to make better decisions about whether to give, or continue to give credit to consumers and businesses.
Traditional credit scoring methods nowadays prove inaccurate as they substantially ignore the vast amount of data available online about clients, focusing on just a minor set of details on the customers requiring loans (full name, address, age, gender, bankruptcy/insolvency data whenever available, court judgements, taxpayer or other unique identification number, assets and properties, etc.) and ignoring lots of valuable inputs. On the contrary, an average Facebook profile contains 5.000-10.000 pieces of information, out of which up to 100 is relevant in financial credit scoring and can be used to predict actual payment behaviour.
Traditional methods also fail in serving and supporting entire generations or groups of people, having no banking history and thus not being identifiable by most of these parameters will actually receive weak accreditation for receiving loans or any type of credit. Among these, youngsters (between 18-30) are especially important due to high youth unemployment in EU and a fear of an entire “lost generation” as a result of the debt crisis. In addition, about the 70% of people has no record in traditional credit bureaus, thus having no banking history and making it difficult for a lender to predict his/her payment behavior. Immigrants are part of the EU open services/people/workforce agenda. For youngsters and immigrants, as well as middle-aged specialists or representatives of the working class, local credit bureaus coverage is low and these clients cannot have access to financial products (underbanked segment). However, for such clients there is still a lot of online data available which makes it possible to assess their creditworthiness and allow banks issuing loans acquiring new clients and getting higher incomes (less credit losses).
We will solve the problems mentioned above by providing the innovative BIG DATA SCORE solution that assesses the credit quality of people and overcomes the limitations of currently available scoring methods. After crunching more than 50 million lines of data, Big Data Scoring has discovered that social media own very strong predictive powers when it comes to credit behaviour. The model we are proposing is therefore innovative and owns incredible competitive advantage, as it allows to collect up to 15.000 pieces of information by integrating social media and Internet browser data to exactly determine the customer’s payment behaviour and their default probability. Because of the use of innovative data sources, we can help banks find the data and make sense of it. Hence, young people and immigrants can be more easily served by banks.
In order to expand the current Big Data Score product to all European countries, and according to the feasibility study carried out in Phase 1 (Big Data Score – GA No. 622652), the proposers intend to set and achieve the following objectives:
1. Bring the present Technology Readiness from level 7 to 9.
2. Provide the system with complete access to real life data and Open Data made available from governments.
3. Development of a marketing and sales strategy based on two key principle: vertical approach and distribution approach, and development of an IP protection strategy.
Traditional credit scoring methods nowadays prove inaccurate as they substantially ignore the vast amount of data available online about clients, focusing on just a minor set of details on the customers requiring loans (full name, address, age, gender, bankruptcy/insolvency data whenever available, court judgements, taxpayer or other unique identification number, assets and properties, etc.) and ignoring lots of valuable inputs. On the contrary, an average Facebook profile contains 5.000-10.000 pieces of information, out of which up to 100 is relevant in financial credit scoring and can be used to predict actual payment behaviour.
Traditional methods also fail in serving and supporting entire generations or groups of people, having no banking history and thus not being identifiable by most of these parameters will actually receive weak accreditation for receiving loans or any type of credit. Among these, youngsters (between 18-30) are especially important due to high youth unemployment in EU and a fear of an entire “lost generation” as a result of the debt crisis. In addition, about the 70% of people has no record in traditional credit bureaus, thus having no banking history and making it difficult for a lender to predict his/her payment behavior. Immigrants are part of the EU open services/people/workforce agenda. For youngsters and immigrants, as well as middle-aged specialists or representatives of the working class, local credit bureaus coverage is low and these clients cannot have access to financial products (underbanked segment). However, for such clients there is still a lot of online data available which makes it possible to assess their creditworthiness and allow banks issuing loans acquiring new clients and getting higher incomes (less credit losses).
We will solve the problems mentioned above by providing the innovative BIG DATA SCORE solution that assesses the credit quality of people and overcomes the limitations of currently available scoring methods. After crunching more than 50 million lines of data, Big Data Scoring has discovered that social media own very strong predictive powers when it comes to credit behaviour. The model we are proposing is therefore innovative and owns incredible competitive advantage, as it allows to collect up to 15.000 pieces of information by integrating social media and Internet browser data to exactly determine the customer’s payment behaviour and their default probability. Because of the use of innovative data sources, we can help banks find the data and make sense of it. Hence, young people and immigrants can be more easily served by banks.
In order to expand the current Big Data Score product to all European countries, and according to the feasibility study carried out in Phase 1 (Big Data Score – GA No. 622652), the proposers intend to set and achieve the following objectives:
1. Bring the present Technology Readiness from level 7 to 9.
2. Provide the system with complete access to real life data and Open Data made available from governments.
3. Development of a marketing and sales strategy based on two key principle: vertical approach and distribution approach, and development of an IP protection strategy.
Phase 1 project allowed us to focus on the creation of a sound business plan and an accurate feasibility assessment. In the course of the action’s duration, the following activities have been performed:
1. The creation of Deliverable D1.1 implemented through the active contribution of the project members, the document being drafted, iteratively revised and modified according to the most recent available data and information;
2. The assessment of whether the Big Data Score technology, tested and validated in the European northern countries, can be actually extended to the whole EU and customized for each single country, that has been confirmed in the course of the feasibility analysis;
3. The evaluation of the project’s contribution to societal objectives setup at a European level, and the identification of the necessary privacy-related path to be followed to comply with current EU regulations on the use of personal data;
4. The description of the existing technology and of the improvements that need to be made to the model for the achievement of a new and more efficient scoring method for any type of credit institution;
5. The drafting of a financial plan forecasting revenues and sales volumes for the Big Data Score introduction in the B2B financial market and future commercialization and distribution;
6. The selection of an appropriate framework and funding scheme supporting the Big Data Score new business opportunity and the achievement of a higher maturity level of the proposed technology, identified in a Phase 2 Project of the H2020 Programme.
The main results achieved in the project mostly relate to the creation of the Deliverable reporting:
- The identification of the target market for the Big Data Score method and of the customers and stakeholders groups;
- The description of the regulative framework for the project, and an analysis of the main barriers to entry of the target market (strategies to overcome them);
- An analysis of the main competing products/companies in the reference sector and the creation of a strategy to reach the market and avoid competition issues potentially preventing the new business from having the expected success;
- A study showing the value proposition of the project, the current state of development, the technical feasibility of the proposed solution and the description of the minimum viable product expected at the end of the project;
- The creation of an executive plan (scheme of the Work Breakdown Structure with WPs and Tasks) for Phase 2 project, including an estimate of the project budget and financial projections for the product sales confirming the beneficial aspects of the project and its economic feasibility.
1. The creation of Deliverable D1.1 implemented through the active contribution of the project members, the document being drafted, iteratively revised and modified according to the most recent available data and information;
2. The assessment of whether the Big Data Score technology, tested and validated in the European northern countries, can be actually extended to the whole EU and customized for each single country, that has been confirmed in the course of the feasibility analysis;
3. The evaluation of the project’s contribution to societal objectives setup at a European level, and the identification of the necessary privacy-related path to be followed to comply with current EU regulations on the use of personal data;
4. The description of the existing technology and of the improvements that need to be made to the model for the achievement of a new and more efficient scoring method for any type of credit institution;
5. The drafting of a financial plan forecasting revenues and sales volumes for the Big Data Score introduction in the B2B financial market and future commercialization and distribution;
6. The selection of an appropriate framework and funding scheme supporting the Big Data Score new business opportunity and the achievement of a higher maturity level of the proposed technology, identified in a Phase 2 Project of the H2020 Programme.
The main results achieved in the project mostly relate to the creation of the Deliverable reporting:
- The identification of the target market for the Big Data Score method and of the customers and stakeholders groups;
- The description of the regulative framework for the project, and an analysis of the main barriers to entry of the target market (strategies to overcome them);
- An analysis of the main competing products/companies in the reference sector and the creation of a strategy to reach the market and avoid competition issues potentially preventing the new business from having the expected success;
- A study showing the value proposition of the project, the current state of development, the technical feasibility of the proposed solution and the description of the minimum viable product expected at the end of the project;
- The creation of an executive plan (scheme of the Work Breakdown Structure with WPs and Tasks) for Phase 2 project, including an estimate of the project budget and financial projections for the product sales confirming the beneficial aspects of the project and its economic feasibility.
The new scoring method to be developed by BDS has a huge advantage over all its competitors since the models developed up to now have already been matched with actual payment data, verified and tested as for accuracy, repeatability and stability. Should any competing company come up with similar models to the ones currently in use in Northern and Eastern Europe, it would take them at least 12-18 months to obtain the same maturity level of Big Data Score models.
The work performed by BDS in the past years allowed to have a continuously increasing database resulting in more and more efficient and accurate model development, hence making it very difficult to catch up with this product for anyone. BDS collects up to 15.000 pieces of information for each customer, that is approximately 500 times more than the data collected for traditional loan applications.
The following benefits are expected to result from the use of our innovative scoring method:
1. All data collected is available across the lenders, making integrations seamless and the entire process fully automated; the price for calculating each single is low if compared to traditional scoring technologies (hence, the business is a easily scalable and very profitable);
2. Ease of use – the scoring algorithm runs extremely fast and allows for the reception of a factsheet (including score) on the customer in just a few seconds;
3. Proven 20% improvement in scoring accuracy compared to today’s models and 30% reduction in credit losses, demonstrating that Big Data Score has stronger predictive power than positive credit data from traditional bureaus;
4. Ease of integration of the tool for new clients – integration of Big Data Score requires no initial investment by the client companies and takes only 4-6 hours of IT work by BDS. Scores are calculated for all clients (either on Facebook or publicly available online data). This helps decrease credit losses, increase acceptance rates and prevent fraud;
5. 70% of people globally having no record in traditional credit bureaus or banking history can get served by banks and lending institutions and receive financial services, among which relevant segments are represented by young people and recent immigrants.
The work performed by BDS in the past years allowed to have a continuously increasing database resulting in more and more efficient and accurate model development, hence making it very difficult to catch up with this product for anyone. BDS collects up to 15.000 pieces of information for each customer, that is approximately 500 times more than the data collected for traditional loan applications.
The following benefits are expected to result from the use of our innovative scoring method:
1. All data collected is available across the lenders, making integrations seamless and the entire process fully automated; the price for calculating each single is low if compared to traditional scoring technologies (hence, the business is a easily scalable and very profitable);
2. Ease of use – the scoring algorithm runs extremely fast and allows for the reception of a factsheet (including score) on the customer in just a few seconds;
3. Proven 20% improvement in scoring accuracy compared to today’s models and 30% reduction in credit losses, demonstrating that Big Data Score has stronger predictive power than positive credit data from traditional bureaus;
4. Ease of integration of the tool for new clients – integration of Big Data Score requires no initial investment by the client companies and takes only 4-6 hours of IT work by BDS. Scores are calculated for all clients (either on Facebook or publicly available online data). This helps decrease credit losses, increase acceptance rates and prevent fraud;
5. 70% of people globally having no record in traditional credit bureaus or banking history can get served by banks and lending institutions and receive financial services, among which relevant segments are represented by young people and recent immigrants.