Imagine it’s your job to manage billions of dollars in consumer mortgages. You’d want to know your risk position pretty much all the time. So imagine you had to wait till next week just to find out where you stand right now.
Before the crash of ’08, that’s how it was for most lenders. As portfolios grew, so did data volumes. Computers captured information faster than they could process it. And risk teams simply could not keep pace with demand for new and refined models.
But one of the industry giants has turned all that around. With SAS High-Performance Analytics, risk assessments that used to take a week are now ready in just 84 seconds. Analysts now have the time – and motivation – to iterate models many more times than previously they thought possible.
And the lender saves tens of millions of dollars.
From 167 hours to under two minutes
Previously, the firm’s risk-management organization operated a separate hardware environment to run a performance-intensive routine that identifies characteristics and candidates for modeling. In that environment, the average processing time was 6.5 hours, leading most analysts to compromise their data explorations due to simple pragmatics.
Worse, when the modeling team executed the same routines in its production environment, processing took 167 hours -- practically the whole work week.
These days, however, those routines are ready in just a couple of minutes. SAS High-Performance Analytics eliminates the ETL process (extract, transform, load) required to gather and prep data for analysis on other servers. A component of the solution, SAS Grid Computing, provides an exponential increase in throughput and performance.
New revenue from marketing
The lender faced similar “big data” challenges in its marketing operations. To minimize churn, maximize customer lifetime value and execute more profitable cross-sell and up-sell campaigns, the marketing team needs to target as many as 15 million recipients. But it couldn’t process all that data without high-performance analytics.
Now, the lender has achieved tremendous gains in the throughput of its database marketing – as much as 215 times faster – dramatically compressing the model-development life cycle and allowing its teams to test and validate additional variables for greater reliability. SAS removes the limits on observations and variables that the company can process, opening up the scope of questions to ask and creative ideas to explore.
As a result, the team’s productivity in executing campaign models has improved, and the models are more reliable. With 15 million prospects, even a minor improvement to the 1-percent response rate typical of direct-mail campaigns can translate into tens of millions of dollars in revenue.
Learn more from SAS "big data" customers in this special 32-page report on high-performance analytics.