arvato Financial Solutions, which provides credit rating and fraud detection services throughout Europe, handles massive amounts of data—and wanted to be able to handle many times more. It accomplished this by adopting Microsoft Big Data solutions
and intuitive BI tools. With faster fraud analyses, arvato Financial Solutions may save clients €27 million (US$37 million) annually. And it has cut the time to deliver credit scores by a factor of 1,000, freeing resources for new applications and services—and
greater customer satisfaction.
The arvato Financial Solutions subsidiary of global communications company Bertelsmann wasn’t drowning in Big Data—yet. But it could see the tide rising rapidly.
The company is one of Europe’s largest providers of risk management and fraud prevention services. When an online consumer purchases a product or service from an arvato client, arvato responds in about a second with credit scorecard information that determines
what credit options the seller will offer that consumer. The company’s monitoring and analysis of its clients’ online transactions helps to identify fraud and reduce losses from it.
As the volume of online commerce grows, so does the demand for the company’s services and the pressure that puts on its data infrastructure and business intelligence (BI) technologies. arvato systems handle more than 275 million transactions and 100 million
credit checks each year and more than 10,000 processing steps a minute. The company’s data load grows by 2 terabytes per day.
arvato used data analysis tools including SAS and Microsoft SQL Server software. But none of them supported the in-depth analysis of Big Data at the near-real-time speeds that arvato requires to satisfy its clients and support its own growth. Fraud analyses,
for example, took a day or two of manual data manipulation. But the faster arvato could produce those analyses, the more it could help its clients to avoid fraud and save money. Similarly, the more intuitive it could make its reports, the faster its clients
and internal business customers—account managers, analysts, risk managers, and mathematicians—could act on them. More flexible reporting systems would make it easier for internal customers to adapt existing reports and create new ones in response to changing
“We don’t just want to keep up with growing data,” says Philipp Müller, Chief Data Architect at arvato Financial Solutions. “We want to do a better job of analyzing the data and managing the complex workflows that support our analyses, even as the volume
of data continues to grow.”
||We will launch a new generation of international fraud detection services powered by our risk management platform, which is built on SQL Server 2014 and Big Data technologies.
| Kai Kalchthaler
Managing Director, Risk Management
arvato Financial Solutions
arvato got its chance to do that better job when it learned about Microsoft Big Data solutions and business intelligence tools that plug into Microsoft Excel spreadsheet software and empower users with analytics capabilities without the need to burden IT.
The Microsoft approach to BI appealed to Arvato, according to Müller, because internal business customers—who already knew Excel—could use it immediately, without extra training. Also, they could use it to access data from both SQL Server and non-SQL Server
databases, including Big Data managed through open source Hadoop clusters.
The company began its use of Microsoft BI in a four-month pilot test that ran on a SQL Server 2014 data warehouse and analysis cube, fed in part by a Cloudera Hadoop cluster (arvato is contemplating a companywide migration to Windows Azure HDInsight for
arvato risk managers and analysts used Excel to discover and connect to internal and external data—including Hadoop data—and to transform and load it into Excel for analysis. Analysts used the tool to expedite the creation of complex credit scorecard models.
IT staff used the same tool in a very different way, to find and analyze configuration details on the sophisticated systems supporting arvato workflows. Internal business customers also used Power Pivot to create and run data models in Excel.
arvato internal customers used Power View to create highly visual, interactive charts and graphs that they accessed and manipulated within Excel. And with Power Map, they depicted geospatial data, such as the locations of specific transactions, to help identify
potential patterns of fraud. They also used Microsoft SharePoint to host data dashboards and to share and collaborate on their reports and models.
The pilot test was a success and Arvato plans to put the solution into production.
“We will launch a new generation of international fraud detection services powered by our risk management platform, which is built on SQL Server 2014 and Big Data technologies,” Kai Kalchthaler, Managing Director Risk Management, arvato Financial Solutions.
With Microsoft Big Data solutions and BI tools, arvato Financial Solutions looks to help its clients cut fraud losses in half, speed credit calculations by a factor of 1,000, and triple productivity for the development of complex models.
Potentially Saves €27 Million Annually in Fraud Losses
arvato is using Microsoft data solutions to provide faster service to its clients and help them reduce the incidence and cost of fraud. By replacing manual analysis of fraud-related data with highly automated analyses of Hadoop data through Excel BI tools,
arvato has reduced the time it takes to produce fraud reports from a day or two to just a few milliseconds.
Faster reports mean that clients can get their analyses in time to avoid many more incidents of fraud, and the losses that accompany them. If the faster analyses help to identify and stop 75 percent of fraudulent orders, arvato clients stand to save €27
million (US$37 million) a year. “We expect to have a great impact on our clients’ ability to combat fraud by using SQL Server 2014 and Excel,” says Müller.
Increases Supportable Transaction Volume by 1,000 Times
arvato also adopted Microsoft Big Data solutions and BI tools to gain an edge over the fast-growing volume of data. It has gained that edge—and more. Potentially fraudulent transactions are now identified in less than a millisecond, a reduction by a factor
of 1,000. The vast decrease in compute resources needed for running the rules engine means that arvato can support business growth without increasing the size or cost of its data infrastructure. Even better, after supporting that growth, arvato still has resources
Enables More-Accurate Fraud Detection
The success of the company's analysis and reporting business depends in part on the ability to continually develop fraud detection services of increasing accuracy, and to adapt them to various industries and individual clients. With the new risk management
platform using the latest Microsoft data solutions, arvato can adapt its fraud detection services quickly and even develop new, client-specific models in extremely short time frames.
Müller says, “We will use Microsoft data solutions to drive a more agile development process and work more collaboratively with internal customers to deliver the dashboards, reports, and services that work best for them.”
This case study is for informational purposes only.