From 60 to 18 minutes, Tokopedia’s implements advanced optimization algorithms

The Institute of Operations Research & Analytics (IORA) at the National University of Singapore, and Tokopedia, a technology company with the leading marketplace in Indonesia, have recently signed a Research Collaboration Agreement.

The collaboration strives to provide better insights through R&D to enrich Tokopedia’s marketing campaigns by devising better coupon assignments policies with advanced optimization algorithms.

The research team at NUS IORA will work closely with the Data Science team in Tokopedia on data analysis, optimisation formulation and design of software architecture.  The results of the research will be implemented and deployed onto Tokopedia’s system.

The technologies are vital to Tokopedia’s core operations and will help to provide a competitive advantage to its marketplace platform. Tokopedia believes that the way to measure its success is when its merchants have better lives through selling their products and reaching more customers with technology that they provide. The technologies developed though this collaboration is a step forward this belief, with improved sales through the optimisation of coupon assignment.

The collaboration is led by Professor Teo Chung Piaw of IORA and Paul Condylis, Associate Vice President of Data Science Tokopedia. Professor Toh Kim Chuan (Department of Mathematics, Faculty of Science and affiliated faculty member of IORA), will co-lead the project.

Tokopedia has incorporated the algorithm developed by NUS into their marketing campaign.  With this, they have achieved significant improvement in the efficiency in solving the coupon assignment problem. The splitting algorithm reduce the computation time from over 1 hour to 18 minutes with a near-optimal solution. Moreover, the online algorithm is capable to reduce the computation time from over 1 hour to 1.7 minutes with a small optimality gap, and can be adopted in Tokopedia online decision-making system for coupon assignment and product recommendation.

Professor Teo, IORA Executive Director, said, “This industry-academia partnership between IORA and Tokopedia is a win-win for both.  Tokopedia is able to build on NUS’ growing research expertise in operations research and data science, and our researchers and students have the opportunity to work with industry to solve real-life problems, and to create products that have high value for commercialisation and industry adoption.”

Paul Condylis, Associate Vice President (AVP) of Data Science Tokopedia, said, “Tokopedia’s collaboration with IORA can further enrich the design and development of Tokopedia’s marketing campaigns, which can subsequently upscale our technology and better help our sellers achieve more through Tokopedia’s platform in the long run. We hope synergies like this can continuously accelerate our efforts to democratize commerce and digital access for MSMEs across Indonesia.”

 

Project Team:

From NUS:

Prof Toh Kim Chuan (PI)
Prof Teo Chung Piaw (Co-PI)
Asst Prof Lyu Guodong
Asst Prof Lin Meixia
Dr Xiao Nachuan
Ni Tu

 

From Tokopedia:

Paul Condylis, Associate Vice President of Data Science
Liu Wang-sheng, Data Science Senior Lead