Financial institutions have a wealth of rich data that can be the key to remain competitive, grow revenue, and provide unparalleled service. Institutions are now evolving their understanding of data as a strategic asset.
The relevant insights that live deep in transaction and payment data should be the foundation of the data strategy. This data strategy extends beyond marketing to:
It’s clear that data impacts many departments within a bank or credit union. Starting with clean, usable data can deliver insights that empower an institution to remain competitive. That’s why transaction data cleansing is crucial to a financial institution’s overall data strategy.
Transaction data cleansing is a tool that allows financial institutions to understand account holder transaction behavior and to model patterns in account holder spending. Transaction data requires cleansing because the content of transactions is often cryptic, with many different names for the same merchant. Raw transactions are unusable for data modeling, account holder analysis and integration with other solutions. That’s why transaction data cleansing should be the foundation for any institution’s data strategy.
Transaction data cleansing transforms a raw transaction string into a cleansed business or merchant name, and assigns valuable metadata, such as a detailed business category. This metadata provides context to the merchant, so you know whether Pirate’s Cove is a Miniature Golf establishment or a Themed Restaurant. A financial institution can use this tagged data to provide an organized online banking experience, avoid transaction disputes with better online banking statements, and reduce IT hours.
1. Targeted Marketing
Boost your cross-selling efforts. Identify account holders using competitive products and send them targeted messages about your products, all from insights derived from transaction data.
2. AI & Chatbots
Get more value from AI and chatbot investments. Tagged, categorized data from transaction data cleansing helps AI and chatbots respond faster to account holder requests.
3. Data Warehouse Enrichment
Enable more meaningful reporting and insights. Transactions in a data warehouse can be appended with enriched metadata from transaction data cleansing.
4. Data Science & Modeling
Leverage a proven and robust taxonomy to enhance your transaction data using the most accurate data tags and rigorous categorization.
5. Clear Descriptions in Digital Banking
Free up your call center. Limit the calls account holders initiate thinking there is fraudulent activity in their accounts. Transaction data cleansing provides easy-to-read transaction descriptions within online and mobile banking platforms.
Accurately matching transactions to the right merchants requires a lot of time and resources.
A data cleansing vendor that can accurately identify, tag, and categorize transactions including credit card, debit card, ACH, and bill pay can lead to profit opportunities and provide a competitive advantage.
Financial institutions should evaluate vendors carefully and find a partner that will deliver the match rate, speed, and quality to lay the groundwork for a solid data game plan.
Financial institutions should evaluate which vendor can deliver the best match rate on its data. Match rate describes the percentage of transactions for which a cleansed business name and category is provided. After all, who doesn’t want to get as many transactions cleansed and categorized as possible?
It’s an important metric, but it can be deceiving. A match rate near 100% sounds impressive, but it might be too good to be true. Make sure to get these questions answered:
How large of a data set was used to generate the match rate metric?
Small data sets may show a higher match rate but might not be reflective of the match rate that can be delivered on your data.
Was any data excluded from the calculation?
Typically 5-10% of the transaction descriptions are too vague to use.
Are all types of transactions being cleansed?
A transaction data cleansing solution should identify businesses where consumers make payments, but should also be able to identify payroll receipts, tax refunds, stimulus payments, investment accounts, and more. ACH, bill pay, and deposit account transactions are just as important as card swipes.
Granularity means the specificity and precision of cleansing a transaction. Let’s look at the merchant identification component.
Amazon is the most frequently transacted business in existence. However, not every Amazon transaction is the same. There are at least 27 different business units, like Amazon Music, Amazon Prime, Amazon Web Services, and Amazon Pay. Transaction data cleansing should treat all of these as distinct merchants.
It’s easy to classify a business to a general category, like restaurants. It takes more effort to categorize all the restaurants based on cuisine or style. When you need to understand the spending habits of your account holders, a high-end steakhouse is different from a family restaurant.
For transaction data cleansing, accuracy rate is measuring the percentage of enriched rows that were returned completely correct. The gold standard for evaluating accuracy is a manual review of the merchant name and categorization for each transaction. A statistically significant randomized sample of transactions should be pulled from a larger set of cleansed transactions and graded row by row.
Speed measures the time required to process a set of transactions, while scale is the number of transactions. Both are important. Leading transaction data cleansing vendors can deliver great match rate, accuracy, and granularity on a large data set in a short period of time.
Humans are critical to transaction data cleansing, but humans alone can’t deliver speed and scale. An extensive transaction knowledge base and human-supervised AI deliver accurate results within days.
Alkami’s Transaction Data Cleansing matches each raw transaction to the most relevant human-verified entry in the knowledge base. This ensures a human touch is part of every single enrichment, at speed and scale.
It also provides suggestions to the team of library scientists about where new transaction strings may fit best into the knowledge base. This creates a cycle of humans and machines working together to create an optimal result. As more transactions are received, the better the output gets.
Transaction data holds a wealth of valuable insights that can help a financial institution improve its bottom line. Every financial institution is going to need a transaction data cleansing partner to unlock the value of this data.
Financial institutions need to:
