Designing Customer-Centric Products

Don’t forget the people–designing around human nature Arthur da Silva


The SOA ImpACT meeting last October was likely one of the first major conferences people attended since the pre-COVID era. I participated in this meeting in Orlando and had a great experience, which I detailed in another article. In addition to the conference networking experience, I learned a lot from the presentations. Therefore, I am sharing some of my reflections and takeaways focused on the human side of technological change.

With discussions frequently centered on tech-enabled innovation and change, it can be easy to forget that humans are still in the driver’s seat of implementing and managing change and making decisions as employees and customers. As companies and their people go through modernization, they should understand the following:

  • The customer journey for the purpose of designing customer-centric products and processes
  • The process of human decision-making and behavioral change to understand how to design their recommendations to produce the desired actions
  • The human emotions that can help or hinder the change journey, which is helpful in managing the risks associated with change

A common theme is the imperative to proactively consider and manage the human elements of any process and their associated changes, regardless of the technology implemented. This will help to design customer-centric products and processes, make effective recommendations, and minimize risk when implementing change.

Design Customer-Centric Products and Processes by Understanding the Journey of Your Customers and Influential Stakeholders

Gathering Customer Feedback Is Critical

Maybe you’ve heard the idea of designing customer-centric products and processes. But how does one figure out how to understand customers? In their presentation “Insurance Product Design Through Qualitative Research,” Amanda Turcotte, FSA, MAAA, and Carolyn McMahon detailed the importance of performing semi-structured interviews in addition to using surveys and data analysis to understand customers and their decision-making process fully.

The importance of detailed customer interviews is highlighted when a company finds out there’s a misalignment between its measurement of a product’s value versus the customers’ measurement of said value. Turcotte and McMahon described a case study of a company designing a new supplemental health product, hypothesizing—based on questions from distribution—that customers want detailed transparent disclosure of eligible conditions at onboarding. After performing in-depth customer interviews, the company observed that no one in the sample asked about the covered conditions.

Another example is that many actuaries think consumers use price as the main determinant for purchasing insurance. However, as Kevin Pledge, FSA, FIA, and Richard Pyper showed in their presentation “Personalization Within Digital-First Insurance,” less than one-third of customers compared products from multiple insurers. In response to these findings from both presentations, companies might want to tailor their consumer-facing marketing materials around the questions customers are asking and their values rather than answer questions people are not asking.

Understanding the Motivations and Experiences of Those Who Advise Your Customers

Although it is important to understand your customers, it is also vital to dive into the mindset of the stakeholders your consumers are interacting with directly, particularly those with influence over said consumers and their purchase decision—often, the agents or brokers. Good product design also addresses the needs and concerns of its salesforce. A company risks designing a great product for the end-consumer that does not get sold because the sales team is not incentivized or empowered to sell it. In addition to addressing pain points from distribution, companies can also leverage the branding of the distribution and empower them with personalized marketing to reflect the values of those communities more granularly.

How Actuaries Can Influence and Use In-depth Customer Research Interviews

Actuaries can contribute to qualitative research by asking questions around trends they observe in data, allowing them to dig deeper into why those trends are occurring. Conversely, in-depth interviews will help actuaries cast aside their assumptions regarding what customers want and provide real data on what customers value in their products. This feedback can be critical in product development and pricing, and it is important to get right early in the product design phase, well before launch.

Make More Effective Recommendations and Produce Desired Actions by Understanding the Process of Human Decision-Making

Insurance companies and actuaries need to sell ideas and influence people, whether convincing consumers to buy insurance or making recommendations to coworkers. Unfortunately, people, including actuaries, can fall into the trap of convincing themselves that more data leads to better recommendations and decisions. Yet, 85% of big data projects fail, according to Mico Yuk in her keynote presentation at SOA ImpACT.

Data projects often fail because they disregard the need to focus on the emotional aspect of the decision-making. Yuk described how studies have been conducted on the numerous failures of data projects. Many of these challenges often occur due to human issues rather than technology or data issues. This is because humans still make decisions with their gut, highlighting the need for projects to factor in the emotional side of decision-making despite the higher volume of data available.

Yuk developed a BI Data Storytelling (BIDS) Framework to help garner buy-in and trust when delivering recommendations supported by data. Under this framework, there are three core skill sets around data storytelling:

  1. What to ask
  2. What to write
  3. What to draw

In my experience, visual tools can help discover and deliver insights. Still, the fundamentals around storytelling should be prioritized over the visuals and tools, which Yuk outlines well with this framework. Let’s dive into the framework, layered with my own experience in developing dashboards and data-focused presentations.

What to Ask—Understand the Situation, Audience and Data

The key to understanding what to ask is ensuring you know the context and data. Unfortunately, the keynote at the SOA ImpACT meeting skimmed this step due to time constraints. To fill in the gaps, my personal experience around data storytelling has taught me that designing an effective dashboard or presentation starts with understanding the audience, their motivations, the intended usage of the presentation and, lastly, the context and data. Doing this ensures you are starting with the business purpose, which guides the rest of the decision-making and design choices.

What to Write—Define the Narrative

When it comes to what to write, the presenter should focus on writing the key story points to present to the audience. In summary, the process of data storytelling looks like this:

  1. Set your goals—what does success look like?
  2. Determine metrics to hit those goals.
  3. Review trends in those metrics and see if they are on track.
  4. Dig deeper into why you’re observing what you’re seeing.
  5. Present recommended actions needed to correct course, maintain progress or make a decision.

Furthermore, the words you use for the metrics matter and have specific associations with imagery and emotions in the brain. For example, some words, like “increase,” conjure images with neutral emotions, whereas a similar comment, like “grow,” produces images of nature, food and family, which are associated with positive emotion. Yuk noted that studies show that people are programmed to respond to and remember prompts containing words related to images and emotions. She also shared a cheat sheet as a reference guide containing examples.

What to Draw—Visualize the Data

Vague goals and metrics produce a vague story, which is inherently hard to visualize. This is intuitive to anyone familiar with specific, measurable, achievable, relevant and time-bound (SMART) goals (i.e., most people). Yet, many don’t reflect this in their visuals. This effect is exacerbated by dashboard developers cramming every metric onto a single screen. Once the goals and metrics have been appropriately designed and the story defined, the visuals should easily follow.

Ultimately, the goal of the dashboard or presentation is to guide the audience to an action or decision. Dashboards and charts can be great at pointing out trends and measurements against targets. However, they are often unable to recommend a decision or next action, which is where a person, such as an actuary, can add value by understanding the context, knowing the audience and inferring the result of a decision in relation to the metrics being presented.

Design Prompts to Drive Actions Using a Behavioral Science Framework

Sometimes, you might want to prompt people into action, such as paying overdue premiums, enrolling in voluntary insurance coverage and submitting expense claims. Matt Miller, FSA, MAAA; Seb Kleber, FSA, MAAA; and David Hopewell, FSA, MAAA, CFA, in their presentation “Managing Beyond NGEs,” talked about using the EAST behavioral science framework for nudges and changing behavior. Following this framework, any intervention used to evoke a behavior change should be:

  1. Easy—decisions requiring less effort are more likely to be chosen (e.g., pension auto-enrollment with opt-out rather than opt-in).
  2. Attractive—people are drawn to actions that are made more attractive (e.g., use of visuals, personalization and incentives or sanctions).
  3. Social—people are influenced by what others do and say.
  4. Timely—timing of prompts and costs or benefits influences response rates; helping people plan specific reactions to future events can prompt earlier responses when these events occur.

Companies may be able to improve their targeted outcomes in customer behavior by combining analysis from data, insights into behavior from interviews, applying the EAST framework, testing and iterating. Ultimately, understanding human behavior, nature and emotions is key to driving better decisions and desired actions. Therefore, when going about their daily work, actuaries should consider these concepts and tips to improve their communication and ability to influence others.

Manage Risks Associated With Change by Understanding How Human Emotions Help or Hinder the Change Journey

When it comes to enacting change in an organization, ignoring or underestimating the human element of change can sink the implementation project, particularly when it comes to people’s engagement or adoption of new tools and processes. Therefore, teams looking to implement successful projects should continually address the affected people as part of the project plan.

Considering Human Motivations Is Essential When Integrating Technology Into Their Workflows

Managing technology change, particularly where automated decisions are involved, involves modifying people’s workflows, which may affect their motivation at work and stresses the importance of structuring workdays around that motivation. For example, when implementing an automated underwriting model, it may seem more efficient to have an algorithm handle as many simple cases as possible. However, as humans, underwriters don’t always want to handle only complex cases, which often end in declines and result in demotivation. To address this, the models can be conservative to allow people to underwrite a mix of simple and complex cases, providing some variety to their day.

Another area to revisit is the culture and reward incentives for people affected by technology change. For example, if an underwriting decision is rewarded or punished based on specific claim outcomes, then underwriters would naturally want control over the decision-making. Taking away that control using automated tools may cause underwriters to feel unfairly punished since they would have to take on all the bad risks after the models have cleared most of the good ones.

Lastly, getting the affected individuals involved with the new technology’s development, implementation and maintenance will help with buy-in. For example, actuaries can help underwriters with the education of predictive models. At the same time, underwriters can provide valuable information to help tune the models, as discussed by Marc Cagen, FSA, MAAA, FLMI, CLU, AALU; Laura McKiernan Boylan, FSA, CERA, AALU; and Robyn Wallner, FALU, FLMI, CLU, in the presentation “Can We Talk?: Improving Collaboration Between Actuaries and Underwriters.” A helpful approach is to have the underwriter and the actuary look at specific cases together, particularly at the edge cases, to mutually explain the nuances of the models and why odd cases might be tripping up the models. Talking through counterintuitive examples might also add or change some variables that the model might have been missing or even remove some unimportant variables. Underwriters can then use the artificial intelligence (AI) models as a tool by layering business rules on top of the models and making adjustments based on the nuances in the data that might be hard for the model to capture.

Managing Change and Risk When Implementing New Technologies and Processes

Change management should be planned and start upfront and not be an afterthought at the end. With respect to this topic, Tom Fletcher, Ph.D.; Brigitte LaBrèche FSA, FCIA, MAAA; and Tanya Thompson, FLMI, AALU, described in their presentation “Successful Deployment of AI Models—How to Welcome a Robot to Your Team” that it is key to start with the business problem first rather than the shiny tool since people are more inclined to adopt change if it solves a pain point. Buy-in must be especially present within leadership to establish the purpose and provide direction for the project team to figure out how to implement the change.

Additionally, the presenters assert that project teams should proactively identify potential points of resistance in the organization. To manage potential backlash, teams should communicate early and often with affected stakeholders and ensure they understand the reasoning behind the change and any decisions made. Teams could also use change champions and internal marketing and images to help disseminate information and gather feedback.

Another key risk to manage that I’ve observed is the potential for a regulator or stakeholder not to approve the change or ask for additional information that is difficult to gather without upfront planning. While regulators might have some awareness of the implemented technologies, they might need more education to truly understand the details and data going into, say, the AI models being used. Therefore, it’s important to involve them early and walk through the processes using realistic, meaningful examples—as emphasized by Fletcher, LaBrèche, and Thompson—to get their buy-in rather than wait until all the work has been done. Actuaries should also keep in mind the Actuarial Standard of Practice (ASOP) No. 56 on Modeling. Specifically, actuaries need to vet and become comfortable with any “model,” especially if relied on downstream such as by the pricing team.

What should a company do if it hesitates to start a technology change? First, the presenters recommend that companies consider the risk of doing nothing or being last to the table. Sometimes, an organization should acknowledge that it is not ready for a change now and may require several years of effort before it is ready. One easy way to manage the change is by starting small, building a solid foundation and then adding building blocks.

Tying It All Together

Many of the presentations I witnessed during the 2022 ImpACT meeting explored similar themes around considering the people element of change, be it around launching a new product or implementing a new technology. My final four key takeaways include the following:

  1. Understanding customers and influential stakeholders informs better product and process design and is best done with a combination of data and in-depth interviews. Actuaries can gather qualitative information to explain observations in the data and influence what questions to ask based on trends in the data.
  2. Understanding human emotions and leaning on behavioral science frameworks can help actuaries craft better presentations and more effective recommendations through building more digestible storylines to drive action and desired behaviors.
  3. Planning upfront for change management and managing human reactions by continually maintaining buy-in when implementing new technologies makes the project more likely to succeed.
  4. Designing work processes around human motivation when integrating a new technology and partnering with affected individuals can create a superior implementation of the technology and a more delightful work experience for the employees.
Arthur da Silva, FSA, FCIA, is a senior manager and actuarial innovation lead in Deloitte’s Actuarial & Insurance Solutions practice in Toronto. He is also the chair of the SOA’s Entrepreneurial & Innovation section.

Statements of fact and opinions expressed herein are those of the individual authors and are not necessarily those of the Society of Actuaries or the respective authors’ employers.

Copyright © 2023 by the Society of Actuaries, Chicago, Illinois.