Leading Provider of Biopharma Precision Medicine Intelligence
The Problem
The current platform was not scaling well, was costly to run, and hard to operate.
Inaccurate Ideal Customer Profiles (ICPs) and the need for frequent user modifications strained development resources.
While users could filter ICPs to refine results, the process of defining and applying these filters was resource-intensive, causing cost spikes, slow performance, and user frustration.
This was compounded by the daily rebuilding of large data sets and an app architecture not designed for high transaction volumes.
The Solution
Our team set out to rework the entire architecture to improve performance, reduce costs, and overhaul critical functionality.
The team optimized Ideal Customer Profile (ICP) management by shifting from customized data sets to regional ICPs (e.g., global, North America, Europe).
This change enabled faster filter creation for users and allowed the infrastructure to efficiently rebuild and cache daily results after data ingestion, significantly improving display speed.
The core of the architecture was reworked, migrating logic from Google Kubernetes Engine (GKE) to Dataflow, a Google Cloud Platform (GCP) serverless service. This resulted in significantly faster, less costly, and more resource-efficient processing with native autoscaling.
The Result
The delivered solution significantly increased client satisfaction and performance, leading to increased usage and revenue.
Immediately following deployment, the new architecture garnered widespread user praise for its enhanced ease of use, performance, and result quality.
This success translated into higher client satisfaction, increased system adoption, and greater revenue.
Furthermore, strategic use of autoscaling capabilities optimized costs, reducing average CPU and memory usage by half.
