Notebookrank was playing an important role in consumer decision making in Poland, generating millions in revenue for its partners. This was time to go global and apply proven business model to multiple countries around the world.
Doing great on a local market, platform wasn't yet prepared to handle significantly more users. That was our goal - to make it ready for scaling from relatively small Poland to countries like USA or India. We were also asked to optimize usability and conversion rate, to make Notebookrank competitive on more demanding markets.
- To achieve high performance and scalability we added important improvements to code and platform architecture:
- Introduced stateless, microservice architecture
- Migrated compute heavy AI algorithms to Azure cloud
- Added multiple cache layers betweens parts of the platform
- Delegated all the static content to CDN's
- Added server-side rendering to keep UI fast and SEO friendly
- Adding observability and setting up various conversion goals allowed us to A/B test every aspect of the platform. As a result we implemented multiple tweaks to UI&UX that helps now users to make their purchase decisions faster.
- Global launch of Notebookrank required also adding internationalization features. Multiple markets, languages and currencies were introduced with both UX and SEO in mind.
New architecture allowed Notebookrank to scale form Poland to multiple countries around the world. Right now they operate also in UK and India whilst US and Canadian versions are soon to be launched.
Thanks to performance optimization users get their personalized rankings in less than 35 milliseconds now, and overall page load time was reduced by over 80%. Better page speed & redesigned usability improved conversion rate by 65% compared to the local version, giving this startup a solid foundation to conquer further markets.
The whole process took approximately 8 months.
- Angular 9
- Azure cloud
- GitLab CI/CD