In a data-rich world, it is increasingly important to know how to select the best data source for each process.
What is the difference between first-party and third-party data?
And why are some lenders moving to first-party data for risk assessment?
First-party data is data that you create or collect directly from / with your customer, always with their consent. It is data that you own. This data may be collected from forms, surveys or through interactions with websites and devices.
Second-party data is data that was collected by another party and passed on to you. You are using it second-hand. In this case you can verify that the company supplying the data is the company that collected it directly from the consumer, with their consent.
Third-party data comes from any company that aggregates customer data. These companies often sell sets of customer data as a product for marketing or segmentation purposes. When dealing with third-party data it is important to consider if the data has been collected responsibly.
How do lenders know which data to include in their risk assessment processes. All data types can be useful in credit assessment and carry their own pros and cons – from traditional bureau scores to alternative data credit scores – understanding the difference in the underlying data is becoming more and more important.
Third-party data is widely available however the challenge for lenders is the lack of standardisation, which leads to volatility. Lenders considering third-party data in their risk assessment processes need to consider the data quality can vary between sources even within the same data type. For example, while it is reported that several BNPL companies will start sharing credit info, there will be differences in what one BNPL provider makes public, as opposed to another.
The same issues with standardisation have been noted with Open Banking. Lack of API standardisations between Banks has slowed down widespread adoption.
Another consideration with third-party data is the ever-changing privacy regulations. Due diligence is required to ensure the data you are purchasing or subscribing to is collected in an honest way and in line with data privacy regulations in your market (noting that the regulations you operate in ay be different from those of the data aggregator).
Building critical business processes around a data supply that could be ‘switched off’ is also a consideration for lenders. The continuous availability of third-party data sources is ultimately out of your control. For example, Google has announced fading out the use of third-party cookies by 2023, citing user demand for more privacy, transparency and choice over how their data is used.
We are coming out of a cycle of distrust of data from consumer’s perspective. Around the globe there is a shift if consumers’ understanding of the power in their data. Increasingly users are moving towards services that enable them to be in control of how, when and what their data is used for.
Third-party data is often unstructured, non-specific data which is ‘reverse-engineered’ to fit a need. First-party data has the advantage of being created with a specific purpose in mind. When considering applications for credit, lenders can build processes which collect and create data which is specific to credit worthiness. From a data science perspective, at Begini we believe that the single best way to create data is with the person.
Users should be informed, empowered and in control of the data they share. With the right tools, lenders should be able to work with their borrowers to structure a process that allows them to share specific, structured and predictive data related to willingness to repay a loan.
Lenders can future-proof their credit assessment with first party data created within the application process. Through psychometric assessments, we can create data with the inherent purpose of willingness to repay to generate an alternative credit score.
Psychometrics for credit assessment is a way to measure psychological traits that have a demonstrated relationship with credit worthiness. Our assessment captures multiple personality traits. The most relevant are locus of control, fluid intelligence, impulsiveness, confidence, delayed gratification and conscientiousness. These traits let us identify applicants who are likely to repay their loans.
Not only are these insights predictive of risk, they have been shown to be very stable. While an individual’s financial situation might change, their personalities tend to remain stable.
First-party data character data can be used to assess a user for which there is no (or little data), such as thin-file applicants. Or it can be combined with traditional and third-party data. Character insights have very low correlation with traditional credit scores, so they can be used in combination to provide greater predictability to existing scores and a more profitable portfolio.
Photo by ThisisEngineering RAEng on Unsplash