Managing Risk in a Data- and Model-Driven Culture

Imir Arifi, Senior Director, Business Solutions Consulting, Health Care Service Corporation
18
23
7
Imir Arifi, Senior Director, Business Solutions Consulting, Health Care Service Corporation

Imir Arifi, Senior Director, Business Solutions Consulting, Health Care Service Corporation

Key business decisions are increasingly driven by data. Not only can data identify what happened historically in the business and why it happened, it also helps predict what is likely to happen in the future and inform action. This is called modeling. Health insurers use models for a number of purposes, including helping members develop wellness plans or save money by avoiding unnecessary trips to the doctor.

As organizations explore emerging technologies like artificial intelligence and machine learning, which rely on a group of data from different systems, they are more dependent on these models. There is no question that using data and models has many potential benefits, such as decreasing costs and improving efficiency. However, reliance on models alone without properly infusing the value of human logic and intuition can lead to misinformed strategies that aren’t practical for the business or consumers. To help maximize benefits and mitigate the risk, two key elements are needed: accessing and integrating trusted data that can be used by models and establishing the right team that maintains focus on business strategies and objectives.

  As businesses continue to adopt a model-driven approach, they need to ensure they have the right data, expertise and processes​  

Accessing and Integrating Trusted Data

Successfully integrating the data for models can be challenging. For example, garnering quality sources of data, integrating the information and making it available for use in models often requires significant effort. There’s always the possibility of mathematical errors when building algorithms or misinterpretation of the information models produce without the broader business context.

To address possible data quality or calculation mishaps, analytics-as-a-service platforms, which build models through automation, support data scientists by handling the data capabilities itself. But emphasizing the importance of analyzing data in the broader business context and understanding the problem being solved at the forefront is likely to require an ongoing effort and even integration into formal processes.

Leveraging closed loop systems can also help more closely align models to the business. Applying a closed loop system allows models to learn from their predictions as more data is processed. Closed looped models are often used to help inform product development and give organizations a competitive advantage in understanding what consumers want and providing it before the market demands it.

Establishing the Right Team

When working with models, putting together a team and establishing strong processes can be just as important as the data integration process. The strongest digital capabilities closely align to the business strategy and consumers’ needs by addressing things that happen in the real world, which requires a team with a variety of skills and insights. As models bring efficiency to operations, organizations can shift their focuses to investing in specialized teams. But that also means they run the risk of operating in silos.

Model development teams should have diversity in core business experience, such as data scientists, model validators and data engineers that can bring different points of view. Team diversity can also help in developing the construct of certain models, like those around information security, which may require both cybersecurity and core business oversight given the constant stream of new data available.

In executing this concept, many organizations find a three-point process to be effective when applying models. This process includes developing the model with the right data, validating the model by ensuring accuracy and auditing the model through a review and analysis. The independent review of a validation team helps ensure the math is correct and there are no coding errors, while the audit team provides a third line of defense that also ensures the models are continually aligned to the business.

Ready to Manage

With organizations’ growing use of models to inform business strategies, understanding the risks and implications of technology and implementing the appropriate safeguards is critical to gain value from the investment. As businesses continue to adopt a model-driven approach, they need to ensure they have the right data, expertise and processes. When used consistently, this framework can allow businesses to feel more confident in their exploration of emerging technology to help devise new ways to serve consumers and anticipate their needs.

Read Also

Technology in Legal Industry, a Far Fledged Dream

Technology in Legal Industry, a Far Fledged Dream

Scott Rechtschaffen, Chief Knowledge Officer, Littler Mendelson
Knowledge Management - A Pragmatic View

Knowledge Management - A Pragmatic View

Christopher Harrison, CTO, Nova Southeastern University
Create Constant Growth for QA Professionals (or Perish)

Create Constant Growth for QA Professionals (or Perish)

Tim Harrison, Principal & Chief Knowledge Officer, SQAsquared
Current Market Trends Shaping The Knowledge Management Space

Current Market Trends Shaping The Knowledge Management Space

Dora M. Tynes, Chief Knowledge Officer, Sheppard Mullin Richter & Hampton LLP