No items found.
Success Case
Daerijubu, an O2O domestic service app, is a mobile app that matches users with domestic helpers. It was the first to be launched in Korea in 2015 and has been a pioneer in the industry. In the early days of the service, Daerijubu was struggling to match the needs of users with the needs of suppliers due to inconsistent service quality. Based on the data voucher support project, they were able to develop an 'AI matching algorithm' system that identifies user intentions. We introduce the story of Daerijubu, who not only achieved customer satisfaction through collaboration with DataHunt, but also received the honor of being selected as a best practice.
TheData Voucher Support Project is a government program that provides vouchers to early and mid-sized companies, small and medium-sized businesses, and budding entrepreneurs to purchase and process data. DataHunt has been selected as a supplier for four consecutive years since 2019, and was selected as a best practice in 2020, garnering attention from the data processing industry. Checkout some of the real-life examples of clients who have worked with Datahunt below.
Prior to the commercialization of domestic help services, there was a strong perception that the domestic service market was a "niche market."There was a stereotype that it was a service enjoyed by the upper class and would be expensive. However, as we operated the platform, we found that the patterns of users were different from the perception. We realized that customers were looking for a variety of detailed help rather than a low price.
However, the existing recommendation model, which evaluates providers based on questionnaire reviews and satisfaction scores, has limited recommendation accuracy.As a result, we wanted to improve our AI matching algorithm through the datavoucher support project. In this process, it was suggested that we utilize"natural language data" that expresses user intentions. After setting the goal, we collaborated with DataHunt to enhance the recommendation model overall.
In order to improve the algorithm by processing natural language data, we first needed to label the NLP data considering the recommendation model. Due to the large amount of natural language data, it is likely that it contains many unnecessary contents or typos to understand the intention, so we needed high-level labeling technology and know-how to ensure that the accuracy of the algorithm does not decrease. Also, the performance of the personalized recommendation model varies depending on the quality of the data, so the technical skills of the engineers were important at this point.
DataHunt had a lot of technical experience and know-how, so they responded well.
We started by refining more than four types of text data, including customer review data, reservation requests, chat history, and introductions and greetings from housekeepers. We performed algorithmic annotation to ensure accurate and smooth matching between customers and housekeepers.
We successfully improved the matching algorithm, which was our top priority. We succeeded in building high-quality natural language data, which simplified and streamlined the data collection process, which was previously done manually by surrogate housewives like a menial labor task. User satisfaction naturally increased as we understood their intentions and needs in context.
Datahunt's meticulous consulting was also instrumental in the actual selection of the business. When it was finally completed and released to the public, the matching AI garnered widespread attention. It was said to have increased satisfaction between buyers and sellers, facilitated long-term transactions, and laid the foundation for more high-quality suppliers to be hired directly.In 2020, the joint product between the agency and Datahunt was finally selected as one of the best examples of data voucher support projects.
For companies looking to transform their business, the Data Voucher Support Program should be an opportunity to take a new leap forward. Simply meeting the deadlines set by the contract is not enough to achieve the essential goal. We need to select a partner who can think about our business with us, and who can have a long-term positive impact on our vision.
Datahunt has been the number one partner in our journey to becoming a successful datavoucher support business. Not only did they make the process of data collection and processing quick and clear, but the quality was comparable to that of a large company.
The domestic help market is still dominated by suppliers. However, I believe that if Daerijubu and DataHunt continue to strive to improve quality, both suppliers and consumers will be satisfied. Mirroring Daerijubu's progress through the datavoucher support project, we plan to continue to collaborate with Datahunt on AI model and data construction.
.svg)
