Transaction categorisation: 9 value-adding use-cases for retail banks and lenders
The growing number of account aggregation service providers has significantly reduced the barrier for banks and lenders to get access to customer transaction histories. As the PSD2 and Open Banking initiative gains more traction, more banks and lenders find themselves asking — “We have access to customer account information. Now what?”.
To help you navigate the maze that is account data, we’ve compiled a list of value-adding use-cases in lending that will help you understand the value of collected account information.
It all starts with transaction categorisation
Account data is noise — it contains abbreviations, numbers and strings of letters that present a great challenge when trying to understand them. The human brain is wired to be able to find patterns in noisy data, which is why for a long time banks and lenders were hiring numerous credit specialists to process the ever-growing number of bank statements, however today this data can be processed automatically using algorithms.
Algorithms make it possible to screen bits of information in the transaction descriptions or information fields and categorise the transactions according to specific pre-set categories.
A few examples:
Raw transaction:“PURCHASE ****1234 STR*BCKS 15.11 EUR”
Categorised: “Expense > Coffee shop > Starbucks > 15.11 EUR”.
Raw transaction: “SaveMyBaconNZ payment A***15513 566.00 NZD”
Categorised : “Expense > Payday loan > SaveMyBacon > 566.00 NZD”.
Categorisation allows transactions to be grouped according to “themes”, which can then be used in further analysis or lending decision making.
9 use-cases for account information
Once you are able to categorise transactions, the opportunities to use the new clean data are almost limitless. We’ve laid out some of the most useful applications for categorised account data.
Verify real income — often a loan applicant has multiple income streams, including freelance work or perhaps side jobs, such as driving for Uber. Account data allows you to identify the salary and other regular or irregular income for all new or existing customers.
Value: generate the missing data required to get the full information about customer’s income for credit scoring or pre-scoring purposes.
Verify activeloans— credit bureaus might not always provide the full picture of all liabilities a person has, either due to restrictions in local legislation or due to delays in how lenders report to the bureaus. Account data allows you to identify all incoming and outgoing loan payments, including loans that are not registered in credit bureau agencies.
Value: better insights on actual liabilities and can provide the missing data required to score or pre-score customers.
Discover risk-reducing behaviour — in lending, the context of the loan applicant is important and in many cases it might provide additional reasons to approve the applicant. Account data allows you to capture insights that give context to customers who have been previously defined as “high risk”. It can illuminate reasons for previously overdue loan payments or transfers on behalf of family members.
Value: reject fewer loan applicants due to misleading information from credit bureaus.
Check “red flags” — credit history and income gives a good first impression of a loan applicant, however to get the full picture of the applicant’s financial health, it’s a good idea to look for additional signals. Account data allows you to identify excessive gambling, frequent cash withdrawals or concealed bailiff and debt collection cases. among other factors.
Value: more insight to automatically identify high risk customers and reduce administration costs.
Identify unusual behaviour— recognising potential fraud cases is crucial to maintain a healthy loan portfolio. Account data allows you to spot potentially fraudulent behaviour, including frequent payments to accounts in other financial institutions (i.e. secret accounts), salary payments to or from other individuals (i.e. false salary payments), “1-cent” verification transactions to other lenders (i.e. unmentioned liabilities).
Value: more insights to automatically identify high risk customers and reduce administration costs.
Discover new features forcredit scoring models— building new scorecards requires a trustworthy data source with behaviours (or features) that can be predictive when assessing a loan applicant’s ability to repay the loan. Modelling features generated from transaction data capture effects that are not captured by traditional credit data.
Value: more predictive scoring or pre-scoring models, with expected GINI uplift of up to 15 percentage points.
Identify existing low-risk customers (for retail banks) — for retail banks to increase the conversion rates on marketing campaigns, it’s a great idea to know the segment of the bank’s customers that are likely to be approved for a loan. In cases where the bank already has a large number of account and card holders, account data allows you to calculate the probability of default for existing customers using only historic transaction data and historic loan performance.
Value: identify low-risk customers that were previously underserved.
Create new customer profiles (for retail banks)— customer segmentation is a great way to target specific audiences with the right message. Transaction categorisation can provide unique new characteristics to profile customers for campaigns (based on where and how they earn/spend their money), including identifying coffee-lovers, sports fans and tech-savvy individuals.
Value: build more tailored offers and marketing campaigns that deliver higher success rates.
Identify regular payments (for retail banks) — when on-boarding new customers, it’s important to make them feel at home from day one. Knowing what payments a customer makes on a regular basis allows banks to build solutions that can help a customer make these payments faster and improve the overall customer experience.
Value: onboard new customers from other banks or improve the UX for existing customers.
Use account data to generate value
Thanks to the development of the PSD2 and open banking initiative, it’s become possible to considerably increase the value generated from collected account and transaction data. These are just a few of the most immediate ways financial institutions can make use of and capitalise on categorised account data.
If you’re interested in diving deeper into transaction categorisation, get in touch with us. We’ll be happy offer our input and support you as you discover how to make the most of your collected customer account information.