Lifesight is an AI-powered full-stack marketing platform that uses client data to help them make marketing decisions.
The platform enables marketers to track and measure campaign performance accurately, focusing on first-party and zero-party data. It bridges gaps left by third-party data, enabling exceptional marketing for a privacy-centric, cookie-less future.
Build a SaaS CDP (customer data platform) solution aimed at improving the productivity of enterprise marketing strategies. The target product was required to have the following features:
The CDP (customer data platform) is a complex class of products lacking established industrial standards for its design and development.
The Xenoss teams faced many challenges during the project that had no off-the-shelf patterns to resolve. Here are a few examples of problems and solutions that were found along the way:
Dynamic data formats
The system was to be able to operate with a huge variety of input-output data formats. System operators were to be able to add various data formats during the runtime. Thus, there was no possibility of designing a static data validation and processing solution.
Solution: The solution architecture was designed with unlimited extensibility and adaptability potential. The new data formats could be added through system configuration without coding and rebuilding the system.
Big data to operate with
The system was required to process data in various patterns: individual events and batch data ingestion, browsing the individual profile and segment insights, activating individual profiles, and segment export.
Solution: This system was designed to implement a particular approach for every processing pattern, providing extreme efficiency for processing the data.
Multipattern data processing
Real-time data activation
Several usage scenarios implied real-time data consumption, creating aggressive response time requirements.
Solution: The engineering team decided to use Aerospike, a special AdTech-focused in-memory data storage solution purposed for ultra-high-speed data operations. We implemented the Lifesight storage as a core element of the CPD’s hot storage, which enabled us to meet response time requirements and made the product usable in real-time scenarios.
Architectural highlights
Business logic scalability
Each system component sensitive to data size or load was made horizontally scalable, allowing multiple instances of the same component to handle varying data sizes or loads.
Real-time processing
Data was processed in a soft, real-time manner, ensuring that it was processed immediately upon entering the system without any delays.
Efficient data sharding
The data sharding process was optimized for efficiency. To maximize the system’s performance, the data pipeline was designed to recognize that each data processing stage should efficiently handle a portion of the data size, ensuring fast and effective processing without the need for data preparation at any stage.
Comprehensive monitoring
A robust monitoring service was implemented, enabling continuous monitoring of all services and processes. This comprehensive monitoring system tracked the system’s health, performance, and statistics, allowing timely intervention in case of any issues during production operations.
HUGE LOAD
Zero system crashes
During trial tests, the system demonstrated remarkable load tolerance. It proved capable of supporting thousands of clients across dozens of workspaces, each accommodating millions of users.
PROVEN IN PRACTICE
The cloud-agnostic approach
2 months after we started the development, the client asked to switch from Google Cloud to Amazon, and we managed to do that in 2 days.
INTEGRATED
High level of integrability
The solution was seamlessly integrated with prominent e-commerce platforms such as Shopify and Magento, as well as mobile SDKs and cloud storage services.
MVP DELIVERY
8 weeks
We developed a multifunctional CDP solution, which enabled Lifesight to start early- stage sales and beta-client onboarding.
COST-EFFICIENT INFRASTRUCTURE
TCO
Ended up being below customer expectations. It includes the cost of cloud hardware and internet traffic.
TEAM CAPACITY
9 people
The team of 9 developers was built in 3 weeks, which helped to kick off the development process quickly.
Sounds interesting?
