The Japan Times の 2025年12月25日付「スタートアップ・VC特集2025」 に、当大学関連のタイアップ記事「Polling experts aim to make tough decisions easier」（外部リンク）が掲載されました。
本記事では、本学発のスタートアップ「VETA」が取り組む、意思決定支援の新たなアプローチについて紹介されています。
同社が開発した、人々の価値観を引き出す「Value Elicitation」法による意思決定支援に焦点を当て、従来のアンケート手法とは異なり、回答者の潜在的な価値観や優先順位を可視化できる点が取り上げられています。本技術は、投票行動や進路選択、居住地選びなど、複雑な判断を要する場面での活用が期待されており、今後は政策立案や各種マッチング分野への展開も見込まれています。本学の研究成果が社会実装へとつながる好例として、国内外から注目を集めています。
記事全文・関連コンテンツのご案内
The Japan Times の承諾を得て、以下に全文を掲載しています。ぜひあわせてご覧ください。
また、The Japan Times に掲載された本学関連の過去のタイアップ記事は、以下のページからもご覧いただけます。
→ Japanese University Guide – Waseda University（外部リンク）
Polling experts aim to make tough decisions easier
‘Value Elicitation Method’ helps voters uncover true feelings on thorny election issues, Veta’s political scientists say
by Maiko Muraoka
Every day, people are forced to make choices with limited time. These decisions are often made under complex circumstances, with numerous factors intertwined and no single perfect choice, forcing people to weigh their conditions and priorities for everything from which cosmetics to buy to which political party to vote for.
But at such times, are people truly making satisfying choices that reflect their values and priorities? Most people have likely experienced some form of regret, suspecting there might have been a better option.
Veta Corp., a startup originating from Waseda University, offers a solution to this challenge in decision-making with a novel approach that fuses social science insights with technology. The company’s “Value Elicitation Method” is an innovative tool that evolves the latest survey methods and sheds light on individuals’ “true feelings.”
In a recent interview with The Japan Times, Veta’s Chief Knowledge Officer Airo Hino, Chief Science Officer Teppei Yamamoto and Chief Executive Officer Taketo Hara discussed the applications and future potential of the Value Elicitation Method. Hino and Yamamoto are both professors at Waseda University’s Faculty of Political Science and Economics. Hara cofounded Veta with them after he graduated from Waseda’s School of Political Science and Economics, completed his graduate work in economics and worked as a data scientist at IBM Japan Ltd.
At the core of Veta’s technology is “conjoint analysis,” a method that has long been used in marketing and business administration to study consumer behavior. This analysis technique is based on a survey method that uses computer-based randomization and repeatedly presents respondents with hypothetical choices that combine multiple elements. This reveals the respondents’ selection tendencies quantitatively.
In recent years, this method has been applied in political science to analyze how factors such as election candidates’ policies or personal attributes influence their vote share.
This means that the results of the analysis benefited the organizations, companies or researchers who conducted it, while nothing was returned to the individuals who provided the data. The respondents were merely “subjects for data collection.”
The three founders saw this as both a challenge and an opportunity. They developed a unique algorithm that transformed the conventional method into a tool that calculates the respondents’ underlying preferences and values and provides them with feedback.
In politics, for example, respondents can see a quantitative score indicating how important each issue is to them, which is then used to calculate their match with political parties or candidates. This allows respondents to visually confirm their own values and priorities, resulting in more informed and satisfying decisions.
Veta’s Value Elicitation Method was adopted by the Nikkei newspaper’s vote- match tool for the House of Councilors election in July, marking its first real-world implementation.
While conventional vote-matching tools only present compatibility scores based on how closely the respondents’ views of various policies match those of political parties, Veta’s algorithm provides scores indicating the importance respondents place on specific issues. These scores are then used to determine compatibility with political parties.
The key here is how to determine which of the many different political issues holds the highest priority for a particular individual.
“Of course, we could directly ask, ‘What is most important to you?’ but that would introduce various biases such as how one wants to be perceived by others or what is considered socially desirable,” said Yamamoto, a renowned political scientist and statistician who was also a professor at the Massachusetts Institute of Technology until last year.
The solution lies in embedding clues about a person’s priorities within a set of hypothetical choices to extract responses that reflect their true feelings. The more questions there are, the more accurate the results become.
In the vote-matching service for the July election, Veta combined both direct and indirect approaches to elicit each individual’s priorities because there was concern that asking too many questions would tire out users and cause them to quit midway.
However, the results showed some unexpected findings.
“The survey had 20 questions in total, but we designed it so that users could see their results after answering a minimum of five questions, even if they didn’t complete the whole thing. But in reality, the number of people answering all the questions far exceeded our expectations. More responses means more data for more precise analysis, and we plan to reflect this result for future improvements,” said Hino, a leading figure in vote-matching research and its use in Japan.
In a world where fragmented, unverified information is flooding social media and influencing people’s decisions, Yamamoto stated that Veta is advancing social implementation of the professors’ technology because they feel a strong sense of social mission in this project to help maintain democracy.
The Value Elicitation Method is effective primarily in situations where multiple factors must be considered while selecting a “package” of choices, so its applications extend beyond vote-matching.
For example, it can be utilized in evidence-based policymaking. When attempting to formulate policies based on citizens’ voices, simply starting with the most popular requests risks creating inconsistencies or contradictions.
Hara said that by using the Value Elicitation Method to quantitatively identify the elements citizens truly value and convey them to politicians, it could lead to more efficient and effective policymaking by taking feasibility into account.
“The job market is another area of potential application. During job hunting, job-seekers try to weigh various factors like training programs, benefits and possibilities of relocation. Clarifying their own priorities and confirming what they truly value would help them make the right choices. This kind of data would also be valuable for companies providing matching services to connect employers and job-seekers,” Hara said.
Hiring companies could also gain new insights by quantitatively understanding their own values and priorities, potentially enabling them to optimize their recruitment processes.
Furthermore, the Value Elicitation Method is expected to be effective in the real estate market. By visualizing the elements each person truly values about a property, the method could potentially lead to proposals that closely match one’s preferences more efficiently than a simple filter search based on various conditions and yes or no questions.
Veta aims to introduce the method across more fields through a series of proof-of-concept experiments. As one of the ways to expand its use, the company has been developing an application that companies can adopt with Veta’s support. This application functions as a customizable template for gathering the kind of data companies wish to collect or investigations they want to conduct. The data and results obtained will then be used to further improve the application.
This cycle of creating tools that are immediately useful in the real world and feeding the data gained back into research is a unique strength that allows Veta to continue providing updated technologies.
The social sciences, by their very nature, should inherently have high affinity with business ventures. However, university startups in Japan have historically centered on technological fields, such as engineering and the hard sciences. Yamamoto pointed out this is largely due to the traditional separation between the sciences and humanities in Japan.
“In the past, the pool of researchers in Japanese social sciences who conducted research across statistics, data science and mathematics was scarce,” said Yamamoto, who conducted research in the United States for 18 years.
This is changing, and according to Hara, research and education activities that focus on data are now flourishing at Waseda. Furthermore, support systems such as those provided by Waseda University Ventures are playing a critical role in promoting ventures originating from the social sciences.
Waseda University Ventures has made an initial investment of about ¥200 million in Veta while also providing management support. This venture capital firm has previously supported the establishment of startups originating from the Faculty of Science and Engineering that aim to commercialize research outcomes such as those in the fields of quantum computing and diamond semiconductors. The investment in Veta marks its first attempt to support innovations based on research achievements in the social sciences.
The three founders hope the success of social science ventures like Veta will inspire researchers who have never considered commercializing their research outcomes to see it as a viable option. Hara stated that by continuously creating services and products based on social science insights, Japan can move closer to becoming a more innovative society.
