Demonstrating the Value of Data

How to make the case for buying the data you need Timothy Paris and Michael Frings

Photo: iStock.com/MicroStockHub

More data is available than ever before. Actuaries love data! Our ability to analyze and use data has grown tremendously. However, we need to obtain data to use it—and data is not cheap. Making the case for buying the data is critical.

For the June/July 2018 issue of The Actuary, I wrote an article, “When Is Your Own Data Not Enough?” which focused on situations where using external data can strengthen your actuarial work. The article also demonstrated that actuaries’ technical skills, high professional standards and practical mindset uniquely position us to help our stakeholders better use data to manage the risks in their business.

Michael Frings joins me in this follow-up article, and we answer the critical question: How do you get the data you need? The answer is by quantifying and demonstrating the incremental value that the data provides to the business. In this article, we describe how to make the case for buying the data you need.

Data Helps Us Make Better Decisions

In our personal and professional lives, one of the defining features of today’s world is we are inundated with data.1 Businesses risk obsolescence when they do not seek competitive advantages by gathering, analyzing and better using data. This is not new, and there are many historical examples of businesses not using the data available to them and suffering the consequences. Examples include:

  1. The mid-20th century transition from unismoker to smoker-distinct life insurance pricing
  2. Subsequent refinements of preferred categories based on medical and lifestyle underwriting data
  3. The recent introduction of automobile usage-based devices and human wearable fitness trackers

However, data often has significant costs. In a business setting, a well-designed cost-benefit analysis before the start of the data initiative can mitigate the risks of over- or underspending on data.

Steps for Making the Case to Buy Data

1. Know Your Stakeholders

Whether you work for an insurance company for which you would like to advocate for the acquisition and use of more data, or for a vendor or consulting firm that is attempting to sell data and data-related services, you must know your stakeholders (i.e., management or potential clients, respectively) and what drives their decision-making process. There are several components to this:

  • Who is the approver? While it is tempting to try to go straight to the senior-level approver, this may not always be the best approach. The most direct reason for this is the approver often has staff to whom they have delegated detail work, so you need to respect their organizational structure—and you likely will need their support. This suggests the indirect reason, which is if you can win the support of these staffers—the data users and influencers—they can become internal advocates for you and the data. This in turn hopefully makes an approval more likely.2
  • What is their budget? Respect the stakeholder’s budget and select the data offerings that fit the budget. This will make for an easier first step, and you can build toward a larger scope following initial success.
  • How do they quantify value? Most companies have key performance indicators (KPIs), which define what the company thinks is most important and how it quantifies value. So, the most compelling case that you can make for data, the same as for any other business initiative, is to demonstrate how the data will improve their KPIs.3 What will this data allow them to do better or with more certainty, and how do they value each of these? Typically, KPIs cascade in importance based on immediacy and financial impact:
    • Improved current financial results. For example, earnings or surplus capital. Can you quantify that the data will result in higher expected levels, specific downside protection or lower volatility for current period financials?
    • Improved expected future financial results. For example, incremental present value of future profits or higher future surplus capital. Can you quantify the improvement that the data will provide? Is it due to higher margins, additional sales or entering into new markets? Will the use of this data provide the company strategic or reputational value?
    • Improved statistical measures. For example, smaller confidence intervals or higher R-squared measures. While these may be appealing to data users and influencers, such statistical measures typically are not KPIs for the business. Therefore, consider collaborating with the data users and influencers to translate these to financial KPIs so that, together, you can make a stronger case to the approvers.

Case Study

Suppose you are responsible for modeling policyholder behavior at your company, which is a large issuer of variable annuities. The guarantee features common to these products make them a complex combination of risks, and your company uses a capital markets hedging strategy based on the policyholder behavior models that you provide.

CONTINUE READING

2. Characterize the Cost of Data

Relative to the complexities of quantifying the value of data in terms of KPIs, quantifying the cost of data is usually straightforward. As a starting point, fixed amount or recurring subscription costs are common for accessing data. Estimating the cost to use the data is often more difficult, since you may need to consider the cost of hardware, software and staff to do the work, including acquisition and implementation time. Each of these may entail incremental spending or opportunity costs due to redeployment of resources.

Once you have calculated the costs of data, characterizing them is also important to the decision-making process:4

  • For how long will the data be useful? To what extent will you need to refresh the data going forward, and what will be the cost of this?
  • How reliable is the data? What are you and the data provider doing to acquire it? How will you scrub the data and ensure its integrity?
  • Who else has access to the data? Are you reflecting this appropriately in your quantification of benefits?

3. Estimate the Value of Data Before Purchasing

In an ideal world of intelligent stakeholders who are always ready, willing and able to consider new initiatives, you might only need to assemble the cost-benefit analysis components laid out so far in this article and let the results speak for themselves. But, in reality, you probably will need to lay the groundwork so your stakeholders have confidence that you understand the business landscape and how your data initiative would fit into it—even on an approximate basis—before engaging potential data providers on exact scope and costs.5

These approaches can be very useful in this regard:

  • Duplicate the data you have now. That is, if you think your current data is already fairly representative of the underlying relationships and mainly suffer from volume/credibility limitations, a simple hypothetical test is to copy (e.g., 2x, 20x) your current data to give a sense of potential statistical improvement due to higher volume/credibility. Express this relative to your current metrics and KPIs to develop a reasonable upper-limit estimate of the improvement you can expect by using the new data.
  • Simulate hypothetical data. In contrast, if you think your current data is lacking in specific key areas but is otherwise highly credible, you can test a range of hypothetical data in these key areas. Express this relative to your current metrics and KPIs to gauge the potential range of improvement that the new data can provide.
  • Research vendors and seek opinions from other users. More broadly, however compelling you think the case is for new data, stakeholders usually appreciate knowing that you have considered the full range of vendors and data options. Early in the process, spend the necessary time to research the options and document their respective merits to maximize the approver’s receptivity and minimize second-guessing and rework. Getting opinions from other users of the data will strengthen your case and might point out the potential strengths and weaknesses of the data.

Pulling Together The Cost-Benefit Equation for Data

As actuaries, our raisons d’être are risk quantification, cost-benefit equations and pricing “go/no-go” decisions, so the general approach needed here should be well understood. The challenge is to treat data initiatives like any other business initiative—usually not intrinsically right or wrong, but necessitating careful cost-benefit analysis across a range of scenarios before proceeding.

The most compelling way to evaluate the benefits or value of data is to use what each company defines as its KPIs as support. This helps to break through the internal silos that can obfuscate a comprehensive view of value and allows data to be viewed as a risk management tool and not merely an expense. If you can make a quantitative case that your new data initiative will improve the company’s KPIs, your stakeholders should be highly receptive.

Timothy Paris, FSA, MAAA, is chief executive officer at Ruark Consulting LLC, which aims to be the platform and industry benchmark for principles-based insurance data analytics and risk management. He recently served as the leader of the Assumption Development and Governance Subgroup of the SOA’s Modeling Section and as an elected member of the Reinsurance Section Council.
Michael Frings, FSA, MAAA, is a senior vice president in the Global Financial Solutions division of RGA Reinsurance Company, a leading global provider of financial solutions. He leads teams of actuaries who analyze risks and develop creative solutions for clients in the insurance industry. He is an expert in product development, data analytics, financial reporting and investment strategy.

References:

  1. 1. Silver, Nate. 2012. The Signal and the Noise, 9. New York: Penguin Books.
  2. 2. Hopkins, Tom. 1980. How To Master the Art of Selling, 6. Scottsdale: Made for Success Publishing.
  3. 3. Fisher, Roger, and William Ury. 1981. Getting to Yes, 81–94. Boston: Houghton Mifflin.
  4. 4. Pink, Daniel H. 2012. To Sell Is Human, 145–152. New York: Riverhead Books.
  5. 5. Drucker, Peter. 1967. The Effective Executive, 100–142. New York: Harper Collins.

Copyright © 2020 by the Society of Actuaries, Schaumburg, Illinois.