3 Proven Ways to Put Alternative Data Insights to Work

What Are Alternative Data Insights? 

Technological advances have increased alternative data’s importance in recent years. This is mainly due to the build-up in the amount of data available, stored online daily. This data proliferation, coming from different non-traditional sources such as mobile devices or the web, is now known as alternative data.

‍

Traditionally, the applicant’s credit history and their ability to pay were the main data used to determine credit risk. However, with the rapid development of AI, alternative data has become increasingly meaningful, attracting new customers and promoting financial inclusion.‍

‍

Although significant steps are being taken to increase financial inclusion, an estimated 3 billion adults are invisible to credit, lack a credit history, and are thus unknown to credit bureaus. Within this number, a large portion is attributed to the unbanked population, which, according to the World Bank’s latest Findex report, encompasses 1,7 billion people.‍

‍

By combining alternative data with traditional data, new and powerful insights can be gained to improve business projections and financial inclusion. For example, according to McKinsey, retailers that benefit from data analytics in their organisations could increase their operating margins by more than 60%, with the most interesting insights coming from the combination of transactional (billing over time), navigation (on mobile devices) and customer service (returns) data.‍

‍

‍

‍

How alternative data can predict behaviour

AI and ML allow analysts to evaluate data from different sources, integrate them and identify patterns related to marketing, fraud and credit scoring. In this way, it is possible to understand and anticipate opportunities and threats that may arise in the organisation and react quickly to any events.‍

‍

A good example of this is email, through which it is possible to view transaction receipts. This allows us to better understand each customer segment’s purchasing behaviour and predict future sales. This information can help detect fraud, establish strategic campaigns, and assess creditworthiness.‍

‍

1. Alternative data to prevent fraud‍

‍With the COVID-19 pandemic, companies have developed their digital services in record time, but unfortunately, cybercrime and fraud (1st and 3rd party fraud) have advanced at a similar pace. Combining traditional data with alternative data makes it possible to identify legitimate customers and suspicious activity in real-time, minimising fraud through behavioural data.‍

‍

With telecommunications data, such as calls, SMS, or a user’s mobile device contacts, it can be established whether a person is engaging in fraudulent behaviour. For example, if low interactions or few contacts are detected, it is likely that the phone is virtual, disposable, or not the person’s actual one. In addition, through device data (model, geolocation, or SIM card), it is possible to detect repeated loan registrations with the same device or IP address.‍

‍

2. Alternative data in marketing and UX‍

While alternative data technology was initially designed to reduce credit risk, today, brands using alternative data strategic insights can also help marketing and UX teams identify patterns that can be used to improve products or services and strategies and enhance user experience. Also, by obtaining real-time information, alternative data allows faster and more accurate decisions, improves messaging and campaign segmentation, and delivers personalised offers based on behaviours.

‍

An example of this is the case of a neobank, which offers instant non-collateral loans, among other features. The company partnered with Credolab to find the ideal audience for marketing campaigns by analysing previously approved customers (those with >536 credit scores, based on the information on their devices and the analysis of app ownership. The analysis of alternative data relating to app installations, updates via mobile data or Wi-Fi, uninstallations, category, and similar, was then used to identify a lookalike audience that helped improve the target selection at Google and Facebook.‍

‍

From a UX perspective, it is possible to identify friction, and understand the competition and the user brand sentiment better through social networks. Furthermore, with device information, designers can know the consumer behaviour towards technology and adapt the design to improve the onboarding experience.‍

‍

3. Alternative data in credit scoring‍‍

A person’s credit rating is identified with the credit score, a tool that determines how reliable a person is in obtaining a loan. Through the use of metadata, it is possible to complement the traditional assessment by knowing a person’s behaviour in real-time and obtaining predictive schemes, ultimately improving financial quality.‍

‍

Through the analysis of non-traditional data, including open banking data, e-commerce transactions and digital payments, it is possible to check and predict applicants’ affordability to repay a loan. Companies such as Credolab offer complementary products to score an applicant or predict the likelihood that the user will be approved for a financial product. The higher the Credolab credit score, the lower the probability of default.

‍

‍

How to Integrate Nontraditional Data for Credit Approvals 

Integrating non-traditional data into credit approvals works best as a phased process rather than an immediate replacement of existing models.

‍

First, define the business objective. This may be to strengthen fraud protection, improve marketing decisions, or enhance credit assessment. A clear objective helps determine which alternative data source will provide the most relevant insight.

‍

Next, select the right data source and a suitable partner. The chosen data should add predictive power to the information already available within the lending workflow.

‍

The alternative data scoring should then be layered alongside the existing traditional credit score. This hybrid approach allows lenders to complement bureau data rather than replace it, creating a broader view of applicant risk.

‍

Before full rollout, shadow-test the combined model against current decisioning processes. Track approval rates, default rates and model performance to understand the impact. Once results are consistent and measurable, lenders can introduce the approach more widely and continue monitoring performance over time.

‍

How do you get a hold of these alternative data insights?‍ 

Credolab analyses over 80,000 data points through an anonymous fingerprint by collecting device information, registered accounts, contacts, calendars, external and internal storage, application and in-app behavioural data to turn it into predictive data and improve financial quality.

‍

Credolab understands the importance of respecting privacy while extracting relevant information. At Credolab, we leverage anonymised, user-consented device and behavioural metadata to draw these insights.

‍

Alternative data offers insights that are simultaneously useful for reducing fraud risks, improving credit scores, and enhancing marketing and UX strategies. Through analysis and pattern identification, Credolab assists businesses in improving financial inclusion, making quick and accurate decisions, lowering customer acquisition costs, and segmenting and crafting future strategies. The scope is boundless.

‍

What are the main alternative data sources for insights? 

The main alternative data sources for insights include:

• Telecom and device data: Calls, SMS activity and contacts can support credit scoring, fraud protection and customer segmentation.
• Open banking and transactional data: Income, spending and payment patterns can strengthen credit decisions, fraud protection and marketing insights.
• E-commerce and digital payment data: Purchase behaviour can support credit scoring, fraud protection and personalised marketing.
• App installation and usage data: App-related patterns can inform credit scoring, fraud protection and UX decisions.
• Behavioural data: Device and behavioural interactions can strengthen behavioural risk scoring, fraud protection and marketing or UX strategies.