Learn about customer financial behavioral patterns and distinguish valuable customers for your business
Deepen your analysis with additional dimensions
The labels are like hashtags that are used widely on the Internet. Each transaction can receive multiple “hashtags” that are supposed to characterize financial movements on the account.
We label transactions from bank accounts so you can easily identify them and detect trends in your clients' financial behavior.
The metadata contained in the Kontomatik labels allows you to determine, among other things, customer lifestyle and healthcare expenses, as well as payments for travel, children, nightlife and more. You will learn what payment methods your client uses, how much their salary is, and what obligations they have.
This helps with detailed understanding of specific areas like e.g.: insurance, scanning behaviour patterns and trend analytics. Having such data allows user profiling and precise targeting of your clients.
A typical categorization can create some issues when analyzing data. You can only have one category assigned, leaving you to decide which is more important or even whether any category is appropriate. That’s why we went for a more complex solution.
Each operation can receive multiple labels, making both the transaction classification and the information aggregation more accurate. This makes customer profiles and analyses of their financial behavior even more precise. Some transactions will not receive a label, which is far better than a forced assignment, in our experience. A forced assignment could skew the results and lead you to analyze behaviors that simply aren’t there.
The solution is based on a machine learning system. Our team of developers manually assigns labels to hundreds of transaction examples,then creates algorithms that are taught the patterns. As a result, we obtain a statistical model that reviews all fitting labels for a single transaction.
We’re taking into account data like the title, amount of the operation or its regularity. Every single transaction is reviewed for all labels separately.
Each label has different characteristics that assign numerical values to specific words, as well as amounts alongside some algorithmic rules that help verify the dates. If the final crosscheck value for the transaction and the label is high enough, the label gets assigned. Though the amount of factors taken into account is quite large, the labeling process takes just seconds, even for an enormous chunk of transactions!
And even though the amount of factors taken into account is quite large, the labeling process takes just seconds even for an enormous chunk of transactions!
The transaction labeling service isn’t the only thing you can benefit from as part of your collaboration with Kontomatik. Check out our other solutions:
Get a clear picture of the financial condition of your clients to make informed decisions about working with them.
Know the patterns of financial behavior of customers and isolate those valuable for your company.
Verify the customer's identity instantly and minimize the risk of fraud.