MARKET CASE STUDY: GLOBAL DATA COLLECTION

Supporting machine-learning advancement of underwriting processes using online surveys

How does online data collection help improve the insurance industry and premium calculations?
As fast technological advances reshape the insurance landscape, carriers are encouraged to adopt technologies to enhance operational efficiency and build ever more accurate underwriting models.

Project background:

Traditionally, most insurance companies use people working as actuaries to calculate insurance premiums. Actuaries are professionals who use mathematics and statistical modeling to assess the risk and estimate the likelihood of an insurance premium and an occurring claim based on multiple variables like age and gender, etc. They create actuarial tables provided later to the insurance company’s underwriting department, which uses the input to calculate insurance premiums. Insurance companies usually calculate and write programs on their own, but it becomes easier with the use of Machine Learning models and dedicated software.

Artificial Intelligence (AI), particularly Machine Learning (ML), is a promising technology that has proven successful in multiple disciplines related to the insurance process, such as image detection, sentiment analysis, anomaly detection, and more. However, they require the correct input to train the accurate model.
The common problem is that a large portion of on input comes from an applicant, which makes it susceptible to errors in modeling due to the declarative nature of the data. Therefore, having validation input for the model based and use in process something different than user-stated data would provide significant value for improving the quality of the future premium calculation models.

Therefore, one of our Clients approached us with a brief, asking us to support the process and help obtain actual image-based data to train the models supporting the calculation of insurance premiums. Commonly used platforms for Human-in-the-loop marketplaces (i.e., M-Turk) were not an option with their low-quality output, not to mention their geographical coverage and standardized approach.

Project objectives:

The objective was to train an AI model designed for the insurance industry and premium calculations using human input from countries of Asia (Taiwan, Japan, South Africa, India and Hong Kong).

Challenges:

Applying algorithms trained on images of fair skin could be inaccurate for persons of color, as certain lifestyle factors manifest differently in individuals with a skin color other than Caucasian. Since ML models are still in their nascency, by using proper data input, we had the opportunity to improve these studies by ensuring this lifestyle recognition technology can be inclusive to patients of all ethnic and racial backgrounds and later deployed for markets with diversified racial structures.

The markets selected were very diversified from a fieldwork perspective when it comes to the profile of the mobile devices used; this brought a significant amount of technical challenges to providing the quality of the input, which would be matching modeling expectations.

There were ethical concerns related to the use of data. It was extremely important to provide clear instructions and ensure the mitigation of legal risks.
Project took place in 5 diversified markets (Japan, Hong Kong, South Africa, India and Taiwan).

Methodology:

The project methodology consisted of two parts:
  • Country: Taiwan, Japan, South Africa, India, Hong Kong
  • Sample size: N = 2500 online interviews (N = 500 per country)
  • Part 1:
    Have participants answer a 15 minutes questionnaire about their health and daily habits (food consumption, BMI, smoking or drinking behaviors, exercising, deceases, etc.).
  • Part 2:
    Obtaining panelists/research participants' high-quality selfie photos.
Project organisation chart.

Survey questions:
  • Did you use makeup in this picture?
  • Do you currently smoke?
  • Have you ever been a smoker and smoked for more than a pack a month?
  • What is your current level of daily physical activity?
  • How many biological parents, grandparents, aunts, uncles, and siblings do you have that are aged 85 and older (including deceased relatives)?
Scope of work: recruitment, localization and translation of materials, data collection and data cleaning and processing.

Outcome:

The more selfies, images and data, the more the ML system becomes accurate and reliable. TGM Research has developed a unique technology for this custom research project, specifically for this client - integrating a validation module to activate the participant's webcam (works on all devices) with clear instructions to take the selfie (removing glasses, hat, have the face centered, no shadow and quality parameters, etc.). Only the participants with selfies and pictures validated according to the client's scope could go through and participate in the survey part.

This process allowed to map the selfies and various datapoint on face/skin alongside people's lifestyle information (to recognize if someone is a smoker only with a selfie, etc.).

All data was made anonymous, following ESOMAR-related principles and with clear instructions about the use of the submitted information – so no sensitive data was processed alongside the information that allows the identification of the participants; this was especially important to us as we abide strictly by all the codes of conduct.
Exhibits displaying part of the content collected during the research process.
With this methodology and unique validation approach, we rolled out the project in 5 countries, including South Africa and India, where the quality of smartphones and camera resolution is a challenge.

We were able to develop a fast and smart custom solution for one Client - a solution that no other panel companies have, to provide data of a total N=2500 (500 valid anonymized selfies per country) with additional questionnaire input to be used for data training.

Client's feedback:

“We were looking for a company to provide us with visual and questionnaire data in multiple regions around the world, and TGM delivered. TGM took the time to understand our difficult and challenging project and was flexible in meeting our dynamic goals. They stayed in communication with us through the process, and always returned data in a timely manner. We enjoyed working with the team at TGM and recommend them for your data collection needs.”
Client*
*Client information is not presented in this case study as per request. TGM received approval for publishing this project information from the Client.
As the leading online data collection agency, TGM Research conducted multiple market research projects in Asia and Africa. To learn more about our other projects and expertise, please contact us.

We hope you have found this short case study useful.

If you have any further questions about how best to set up your online research study in Asia or MENA region, please don’t hesitate to get in touch with us.

We are living in the digital world with new opportunities for the market research. Join the mobile journey. TGM Research specializes in mobile market research and online panels.

© 2022 TGM Research FZE

 

TGM Research FZE

A68 Circular Road #02-01, 049422, Singapore

A2-102B Building no. A2, Al Hamra Industrial Zone-FZ, Ras al-Khaimah 1005, United Arab Emirates

3rd floor, 100 Nguyen Thi Minh Khai Street, District 3, Ho Chi Minh City, Vietnam

ul. Długa 29/226, 00-238 Warszawa, Poland

Contact Us


TGM Research FZE

A2-102B Building no. A2, Al Hamra Industrial Zone-FZ, Ras al-Khaimah 1005, United Arab Emirates

3rd floor, 100 Nguyen Thi Minh Khai Street, District 3, Ho Chi Minh City, Vietnam

ul. Długa 29/226, 00-238 Warszawa, Poland

We are living in the digital world with new opportunities for the market research. Join the mobile journey. TGM Research specializes in mobile market research and online panels.

© 2022 TGM Research FZE