Kueez Analytics Platform
Kueez Analytics Platform (KAP) takes traffic analysis to the next level by enabling users to configure custom reports with multiple tabs, various filters, and metrics sliced over different dimensions. A/B tests and multiple compare modes encourage data-driven improvements, and snapshots allow re-evaluating historical decisions anytime.
Challenge
The platform was meant to replace the existing analytics dashboard. It had to be much better to justify the switch.
User experience was a top priority. The application should’ve looked modern and had to be fast when dealing with large amounts of data so users would do their job instead of waiting for report calculations all the time. Users should’ve been able to dig into the data quickly and efficiently investigate suspicious results.
Must Haves
When making changes to the websites, monitoring real-time data helps ensure it works well on all devices, browsers, and locales everywhere in the world.
Enabling users to create their custom reports by themselves was a big one. Previously, they had to ask developers to code everything to the existing dashboard, slowing down the process a lot. Developers are always busy.
Optimizing the A/B tests workflow was essential to encourage everyone to perform more tests. Users had to see the summary of every running test and analyze the results deeper by investigating detailed reports. Reviewing this exact data anytime in the future was necessary to improve the decision-making process inside the company.
Involvement
LeanyLabs was recommended to the client by his colleague. After the initial tech kickoff, sharing technical and business needs for the project, the client had chosen us as their technical partner. Trust, technical skills, and dedication were among their main criteria.
We developed the front-end entirely from scratch, closely cooperating with the client’s back-end team and the designer. Initial MVP was ready in roughly two months, and we continue development, adding more valuable features every weekly sprint.
Keeping good quality standards wasn’t easy without the QA and unit tests covering only the data calculation module. We’ve managed to accomplish that with careful developer testing for every task with its impacted areas and mandatory sanity checks procedure for each release.
Solution
To make reports truly customizable, we had to avoid hard-coded logic and develop a generic reporting framework instead. Analyzing the data from millions of daily sessions isn’t easy and adding total flexibility is even more challenging.
We had no room for performance compromises and developed the BI engine from scratch, carefully evaluating every library and optimizing all critical paths of code. The engine works in a separate WebWorker thread to make UI responsive all the time.
MaterialUI React library for UI components matched the client’s vision of a modern Material Design look. The project leverages Recharts for rich customizable data visualization and various libraries for virtual scrolling, drag&drop, forms, etc.
The grid itself was also a challenge. It had unique requirements, like multiple levels of sticky rows and the ability to render massive amounts of data in three modes: hierarchical, flat, and pivot. We had to build our own for this.
Accomplishments
SimilarWeb recently recognized our client as one of the top-10 fastest-growing publishers in the United States in 2020. Although they were mainly using the old platform back then, after the MVP version was finalized (Q2 2020), they started using the new platform incrementally, and every new feature release increased the usage dramatically. They are now using both platforms in a hybrid mode.
The platform accomplished its primary goal of providing improved experience and performance when digging into data. The ability to easily customize the reports allowed them to find valuable insights they weren’t aware of before.
The client is happy, and hopefully, we would build other exciting projects together in the future!
