
Preparing for
high-traffic events
How the Wise team, entrusted with engineering optimization, made sure the Paddle8 auction platform has efficient performance and offers a better experience for their art-loving users.
Paddle8, a New York-based auction house, approached us with a request to prepare an auctioning software for high-load online events and improve the experience for several user personas. Wise team’s tasks as engineering partner were to build a new admin panel for easier auction management, improve site search for collectors, create an in-house app for offline events, and optimize the overall project infrastructure.
Front end: React
Back end: PHP, Python, Django, GraphQL, MySQL, ELK Stack
Amazon Web Services: Elastic Load Balancing, Amazon Elastic Compute Cloud, Amazon Relational Database Service, ElasticSearch Service, Amazon S3
Google Cloud: Google Kubernetes Engine, Compute Engine, Compute Storage, Vision Artificial Intelligence
As the product started to scale, more visitors took part in the actions and more events were about to take place, the website performance began to decrease. Besides, admins needed an easier way to start auctions, and collectors – a more user-friendly way to navigate the art object catalog.
Taking into consideration industry standards, we dived deeply into the client’s product vision and business goals. Based on our technology expertise, we offered a list of the technical environment, infrastructure, and security improvements, as well as a roadmap for the requested features – admin panel and site search.
From the beginning, three engineers started work on the project, cooperating with the CTO, the Marketing department from the USA, and the distributed development team from Poland. In several months, another three engineers joined the team.
We started with moving all services from legacy unsupported Vagrant VM setup to Docker Compose. Our team took care of the code clean-up, refactoring, and supporting an old codebase within the new functionality. It allowed us to reduce the ownership cost and continue work on infrastructure optimization.
With every new feature realized, the Wise engineers also created comprehensive tech documentation and simplified the onboarding process for new teammates joining the project, both from the Wise team and teams from other countries.
To help Paddle8 users find art objects on the website faster, we performed data migration and integrated the software for auction with ElasticSearch Service.
Due to code clean-up, huge parts of the legacy code were deprecated, so we also made sure all the remaining parts are compatible with the newly integrated website search.
We decided with the client to launch a new Bid Retraction feature because deleting a bid was a time-consuming manual process for the admins. To delete a bid, someone from the tech team had to run a SQL query in the database and perform additional testing to make sure that it didn’t affect other auctions/bids.
With the new feature, admins could delete any bids placed accidentally or that are no longer valid directly from the auction admin page. It helped Paddle8 to unburden the tech team from multiple requests and made the overall process of removing misplaced bids more secure.
Using the Paddle8 website users could both take part in the auctions or buy collectibles at a fixed price. Before collaboration with the Wise Engineering team, administrators were able to add the item to the storefront only by hard-coding images, data in the code. This was a very time-consuming and insecure process.
The Wise team helped to implement the storefront management system to the Paddle8 online auction software. Now admins could publish new artworks, add images, data, and select the page they had to be published on via the web interface.
There were two mobile applications the Wise team delivered. The first one was an iOS application on React Native that basically duplicated the functionality of the website from the collector’s perspective. There was a personal user profile with a history of bidding, users could buy items at a fixed price, and place bids taking parts in the online auctions.
The second app for iPad was aimed at in-house use in offline events exclusively. The app, also written on React Native, allowed guests to place bids for the artworks that were presented during the particular offline auction.