Sep 8, 2022
Learn how it is always better to use as many data sources as possible, such as a mix of psychometric data and behavioural data, to add value to the predictive power of your models.
It is no secret that access to data has increased exponentially since the introduction of digital devices. The world is becoming increasingly interconnected, and the abundance of information from smartphones, websites and sensors provide useful data. In this hyperconnected world, every action users take, whether a financial transaction, search history, geolocation or social media interactions, leaves a digital footprint. Companies are quickly leveraging this availability to offer improved products and services.
Credolab, for example, collects and analyses this data, identifying insights that help companies improve their onboarding strategies, reduce marketing costs, mitigate fraud, and make the most informed decisions for credit scoring. This way, credit scoring has been transformed by big data and is thus helping to democratise the financial system. There is a growing trend of lenders tapping into the credit potential of those who have been excluded due to a lack of credit data, whether they are unbanked, self-employed, or still too young to have a strong credit history. Thanks to the predictability of alternative data, creditworthiness can be assessed securely and accurately.
Several alternative sources of data already exist. For example, some companies use gamified questionnaires to obtain psychometric data. Others use telecommunications data (call duration, SIM, frequent contacts) or information derived from bank accounts, public services or the web. For example, telco data can reduce defaults, increase approval rates, and prevent fraud by leveraging data and voice usage, location, SIM card age, average top-up amount, and similar insights.
With this information, companies can access and extract data to understand user behaviour and thus better manage risks. Moreover, combining alternative data and machine learning can lead to more predictive underwriting models that can help close the financial gap, making products more accessible to all - credit visible and invisible alike.
In recent years, psychometric testing has grown in popularity - a method used to measure an individual's psychological abilities and behavioural styles. It is commonly used to measure a person's suitability for a job based on personality characteristics and required cognitive abilities. Through psychometrics, it is possible to understand at a deeper level an individual’s latent skills, qualities, self-esteem, capacities and capabilities and emotional intelligence, which would otherwise be difficult to assess face-to-face.
The new method is attractive to fintech firms because it allows lenders to determine borrowers' creditworthiness independently of their credit scores. This is done by assessing characteristics and personality traits, such as responsibility, trustworthiness, and dependability. It has also become an effective method for determining SMEs' solvency.
A renowned Israeli company, for instance, uses psychometric data collected through an interactive game to assess a player's creditworthiness. Then, through focus groups and external testing, high-performing variables were selected to form the basis of an expert credit risk model. This method enables applicants referred to as “thin files” to be assessed in real-time with surprising results since 60% of them fall within the lender's risk appetite and can be approved merely based on psychometric data. Furthermore, where a bureau score existed, this score correlates more than 83% of the time, indicating the efficacy of psychometrics for credit risk.
Moreover, the same company has demonstrated that certain psychometric traits correlate with financial conscientiousness by at least 80%. Furthermore, certain behavioural characteristics, such as intelligence and planning, were shown to have a 75% correlation or greater with the aforementioned characteristics.
Another method to understand a person’s creditworthiness beyond traditional tools is through behavioural data. Behavioural data is data generated by, or in response to, a customer’s engagement. In the case of credolab, this privacy-consented and permissioned data is extracted from mobile and web metadata, analysing more than 70,000 anonymous data points for each user. Alternative credit scoring goes beyond the traditional risk assessment methodologies and uses machine learning algorithms applied to various data sources such as:
These innovative data enrichment systems can complete a person’s profile in real-time, allowing companies to understand their prospects better and make substantially more informed decisions.
It is always better to use as many data sources as possible, but only if they add value to the predictive power of the models, regardless of whether alternative telco, psychometric or behavioural data are used. Credolab, for example, is a great complementary solution to alternative psychometric data since behavioural data is layered on and has a very low correlation. The ultimate goal is to improve predictions and lower risk, which is achievable through combining these variables with technology. Nevertheless, it is imperative to note that not all providers extract data similarly.
The collection of data and how it is used are considered top priorities for many companies. To determine if a provider is right for you, consider how data is collected and understand how to find reliable data sources. Check out the infographic guide below:
Credolab is the leading behavioural data analytics platform based on device and web metadata and the only one that uses first-party non-personally identifiable information (non-PII) data with the appropriate permissions and respects user privacy. As a pioneer in the industry, credolab has over 10 million engineered behavioural features and the leading proprietary data modelling pipeline rooted in Machine Learning algorithms.
Credolab collects anonymised and depersonalised device and browser information, language and Operating System (OS), and user interface (UI) interactions. This includes total time spent applying for a loan, time spent in the same position, latency and keystroke patterns. Through this alternative data, credolab detects behavioural patterns and develops digital scorecards for banks, lenders, insurers, BNPL and crypto-lending companies, improving existing models by up to 39.9%. The best news? Integration is fast and easy! The technology seamlessly integrates into any iOS and Android app through the credoSDK and any website through credoweb.
Credolab provides digital credit scoring and solutions that can help you maximise businesses using alternative data insights and complement psychometrics and any other form of alternative data. Through the analysis of the customer's digital footprints, credolab provides:
As a result, lenders can improve their businesses' performance by using multiple sources of alternative data, such as psychometric data and credolab's behavioural data.
Access data insights solutions that deliver growth - Fraud detection | Credit scoring | Marketing segmentation. Helps you say "YES" more confidently to more customers!
Learn more about credolab's products and possibilities with our features through our Blog section, and feel free to share our content with your team!
September 22, 2022
2022: How alternative data sources can open up financial opportunities for US credit invisibles