MetLife improves actuarial calculation capability while generating savings
MetLife offers its customers many financial products tailored to meet their unique needs, and critical to its success is the ability to perform complex actuarial modeling of business data. To this end, the company shifted some of the high-performance computing and data processing this modeling requires to the Microsoft Azure cloud platform. MetLife benefited from the flexibility and scalability of Azure data processing capabilities to achieve faster, more accurate actuarial calculations and save significant infrastructure costs, which resulted in more value for the company, its customers, and other stakeholders around the world.
Insurers manage a wide variety of products with different financial profiles and requirements depending on the markets or policyholders they serve. These variables include inflation, investment returns over time, and the number of policies. Actuaries use stochastic models that estimate probability and potential financial outcomes through projections across thousands of generated scenarios over time. By running more scenarios, actuaries can progressively improve the analysis that they can perform by finding convergence on different areas of interest. That convergence, in turn, enables better valuations, projections, and risk analysis.
Running highly complex simulation models is extremely time-intensive and requires expensive, high-performance grid computing capabilities. In addition, workloads can vary based on reporting deadlines and special ad-hoc analyses. This creates a dilemma for insurers: provide sufficient IT resources for peak usage times and leave costly servers underused during other times, or have too little capacity for spikes and get slower, potentially less detailed results.
Equally important to MetLife is the ability to create enterprise solutions for decision makers in the company so they have access to the same high-quality data and results across the globe. Centralization makes it possible for people and infrastructure to be more effective, bring greater consistency to processes, and reduce the delivery time to make decisions.
To meet its business needs, the company wanted to extend its infrastructure without making large capital investments in servers. To accomplish this, MetLife created a processing environment called the MetLife Integrated Actuarial Modeling Environment (MIAME). This end-to-end solution, based on a high performance computing grid, takes advantage of Microsoft technologies including HPC Pack, Windows Server, Microsoft Analytics Platform System, big data, and Microsoft SQL Server. One of the technology goals of the MIAME program is to continuously improve capability and control costs by evolving with new technologies as they become available. This flexible framework allows for these technical paradigm changes without the need to rebuild processing from scratch.
Brian Cartwright, Assistant Vice President at MetLife, says, “MIAME uses parallelism to reduce time and cost while providing enhanced business capabilities. A seamless connection of grid and big data processing environments enables us to do things that were never done before.” MIAME users can analyze results by country, product, basis, and other criteria. “Our standards for using the environment are what makes this so powerful,” adds Cartwright. “Tasks such as building, upgrading, and support become easier. As well, our development cycles are greatly accelerated.”
MIAME continuously sources data from MetLife administrative systems and points them to a consolidated platform for performing calculations in a grid facilitated by Microsoft HPC Pack. Business intelligence and data insight become possible with the data processing resources in the MIAME environment, including Hadoop and the Microsoft SQL Server Parallel Data Warehouse. A centralized, consistent user interface standardizes the business inputs and processing across business teams and usage scenarios. The same user interface also makes the global environment transparent and controlled.
According to Adrian Parris, AVP Global Actuarial Modeling (GAM) team at MetLife, “MIAME has proven to be a very flexible and robust environment that GAM uses for a series of actuarial production and financial reporting activities. It has satisfied the organization’s insatiable desire for ever-greater granularity in analysis and reporting of financial results. What has also proven very advantageous has been our ability to assimilate and exploit new Microsoft technologies, so we can deliver a high-quality product to our Actuarial teams around the globe.”
Extending to the cloud
To increase computing capacity, speed, and performance while reducing costs, MetLife looked to the cloud. Extending MIAME to the cloud gives MetLife a uniform platform with virtually unlimited high-performance computing capacity for accurate and fast modeling, as well as savings opportunities. Ultimately, these advantages help MetLife advance its mission of delivering more value to customers, shareholders, and other stakeholders.
Grant Barrans, SVP & Actuary, GAM team at MetLife, says, “working with Microsoft has consistently paid dividends, they have proved to be reliable, innovative partners.”
Because MetLife was already committed to the Microsoft on-premises platform and wanted to use familiar tools and processes to deliver on its vision, the Microsoft Azure cloud computing platform was its first choice. By taking advantage of Azure for these complex monthly calculations, the company was able to achieve several goals.
Goal One: A consistent and central viewpoint
By having a single repository for data and results throughout the company, actuaries across the globe have access to the same information to make key decisions. “Today, MIAME is our strategic platform for actuarial processing,” notes Cartwright. “It’s big data processing and analytics provide us with a truly global view.” Business managers can view information from a central point to assess impacts across the company. MetLife developed a centralized analytics environment to aggregate and report on results from multiple countries and processing environments, running in multiple data centers on-site and off-site. Automated processing allows for a centralized view, which saves time and eliminates auditing concerns associated with the data handoffs that might otherwise be required to create such capabilities.
Goal Two: Infrastructure cost savings
By using Microsoft cloud offerings rather than continuing to build out its own calculation infrastructure, MetLife expects to save between 45 percent and 55 percent — as well as year-over-year savings — in infrastructure costs that it likely would have incurred to continue to meet reporting time frames. During peak times, MIAME can scale exponentially to meet business and regulatory needs, which enables faster delivery. “When processing is complete, we can scale the environment back down so that we only pay for the capacity we actually use,” says Cartwright.
Goal Three: Improved customer service
The increased speed at which MetLife actuaries can run models has given them the ability to deliver results faster in order to better serve its customers and other stakeholders.
Evolving the solution’s reach and capabilities
As powerful as MIAME was before the cloud, it has gained even greater effectiveness as it has evolved into a platform that can be scaled in real time and on demand. “MIAME was built to constantly evolve and improve,” says Cartwright. “To meet ever-increasing demand, we adopt new processes and technologies to simplify MIAME and enhance its capabilities. The key to making this work is our consistent approach to standardizing technologies and processing patterns, and ensuring that we have the capabilities to support more areas of the business. That strategy allows us to deliver best-of-breed functionality and support.”
And MetLife isn’t stopping there. The company is exploring Microsoft Power BI for data visualization and graphical processing for actuarial calculations. “In more than 10 years of working in large scale data analytics and processing, I have seen over and over how you address a series of bottlenecks,” says Cartwright. “After you solve one data processing problem, you move on to the next one. With the Microsoft technologies and cloud infrastructure supporting our MIAME environment, we can continue to make our processes smoother, faster, and more sustainable.”
Read more on the Microsoft Banking & Capital Markets and Insurance blogs.
