My data team was familiar with AWS solutions from previous job roles, so we believed we could build our analytics infrastructure on AWS in one month instead of up to three for an on-premises solution. We found AWS reduced our infrastructure costs by 20 percent compared to an on-premises infrastructure.
Azby Luthfan Head of Technology, Amartha

Amartha offers loans to entrepreneurs in rural areas of Indonesia to fund their business ventures. To request a loan, entrepreneurs must fill out a form at their nearest Amartha office. The form is then uploaded to the Amartha website, where investors select the ventures they would like to fund. Loan repayments are collected by Amartha personnel in the field.

Amartha faced challenges analyzing data on its loans to quickly identify areas in the country where the loans were non-performing. Azby Luthfan, head of technology at Amartha, explains, “Our data analytics and loan application ran on the same infrastructure layer. We needed to restructure the data in the database for detailed analytics but couldn’t do so without it impeding the performance of our loan application. As such, a dedicated analytics infrastructure was required.”

The business saw it was too expensive to build an on-premises infrastructure because of its capital expenditure costs. As a result, Luthfan looked at cloud services and decided on Amazon Web Services (AWS). “My data team was familiar with AWS solutions from previous job roles, so we believed we could build our analytics infrastructure on AWS in one month instead of up to three for an on-premises solution. We found AWS reduced our infrastructure costs by 20 percent compared to an on-premises infrastructure.”

Each day about 10 gigabytes of data from the loan and repayment database is stored in Amazon Simple Storage Service (Amazon S3). AWS Glue, a fully managed extract, transform, and load (ETL) service, restructures the data and moves it into an Amazon Redshift data warehouse. Amartha’s analytics team then queries the Amazon Redshift data using an open-source business intelligence tool to quickly identify loan defaults.

Luthfan says, “We can quickly see where defaults are increasing and send out Amartha staff to find solutions. With our analytics infrastructure running on AWS, we have seen the number of non-performing loans kept below 3 percent—much lower than the industry average.”

Learn more about Data Lakes and Analytics on AWS.