Integrasi Progressive Web App Dengan Payment Gateway Midtrans Pada Website E-Commerce Umkm (Studi Kasus: Butik Busana Fashion)
Abstract
Butik Busana Fashion currently relies on marketplaces such as Shopee and Tokopedia as its primary sales channels.
While this approach expands customer reach, it faces significant margin compression due to rising commission and
service fees, with platform deductions ranging from 4.25% to 10% on Tokopedia and up to 10% on Shopee.
Concurrently, the boutique's existing website remains underutilized; functioning merely as an inactive, static catalog
without transactional capabilities, it fails to serve as an independent sales channel. This study aims to develop a
website-based e-commerce system for the MSME Butik Busana Fashion, integrated with the Midtrans payment
gateway and utilizing Progressive Web App (PWA) technology to enhance the user experience. The system, previously
a static catalog, has been transformed into a fully transactional platform featuring product management, shopping
cart, checkout, and automated payment processing. The system development employs the Agile Scrum methodology,
encompassing product backlog, sprint planning, sprint execution, sprint review, and sprint retrospective phases. The
Midtrans integration facilitates payments via multiple methods, including virtual accounts, QRIS, and e-wallets,
supported by an automated transaction status notification system via webhooks. Meanwhile, the implementation of
PWA enables push notification capabilities to deliver payment reminders to users. Implementation results demonstrate
that the integration of Midtrans and PWA successfully addresses the need for seamless on-site payments and real-time
payment reminder notifications. Lighthouse testing yielded scores of 93 for Performance, 96 for Accessibility, 92 for
Best Practices, and 100 for SEO. This indicates that the Midtrans and PWA integration significantly enhances
accessibility, information delivery speed, and the overall user experience. Furthermore, functional testing using the
Black Box method across 37 test scenarios in 15 modules achieved a 100% success rate, confirming that all functional
requirements of the system have been fully met.

