Revitalizing B2B eCommerce for Scalable Growth
Product Development
Key Takeaways
- Transformed a struggling B2B eCommerce platform into a scalable powerhouse in just 3–4 months.
- Redesigned ETL process with robust data validation and error handling.
- Slashed data processing time by 55%, from 12+ hours to 5.5 hours.
- Doubled catalog size from 200,000 to 400,000 SKUs, eyeing over 1,0000,000.
- Cut AWS costs by 70%, from $1500 to $300–400 monthly.
- Delivered nine flawless releases, building a trusted partnership for growth.
The Challenge
The B2B eCommerce platform has been in business for over five years, powering a complex supply chain with products sourced from eBay and Asian partnerships. The client’s front-end team was capable, yet they lacked a skilled back-end team to address systemic failures.
The client wanted to scale their catalog to over 500,000 SKUs, streamline their operations, and expand into new markets—without disrupting current operations. The technical challenge involved was like rebuilding a car’s engine while it’s still running. But without a serious overhaul, those goals were out of reach.
Their biggest challenge was a back-end drowning in outdated scripts and unstable processes, choking on the demands of a growing catalog. Three prior teams had failed, with one claiming the platform could never scale beyond 200,000 products. Every attempt to scale triggered crashes, delays, and data errors.
Taazaa faced a formidable task under tight timelines and evolving demands. Our team was faced with 12-hour processing times, buggy code, and an inefficient AWS setup generating 350 MB logs per vendor. With no documentation from past teams and a tight four-week release deadline despite a typical 6–8-week discovery period, Taazaa faced a daunting challenge.
The Solution
Our team partnered closely with the client to turn a failing system into a model of efficiency. Initially tasked with online marketplace to third party live catalogue price and quantity updates, the team adopted a strategic, multi-faceted approach to overhaul the entire code and catalog.
Without initial AWS access, they meticulously analyzed 12–16 hour logs to map the system, replacing legacy Python scripts with a modular ETL architecture using AWS Lambda, SNS, Python and PostgreSQL for robust, incremental processing.
They rewrote inefficient online marketplace API scraping logic, adding advanced SKU matching, validations for product types and prices and a quarantine system for unmatched items. Real-time third-party integration ensured consistent updates with automated markup by price tiers for live product catalogue.
The Results
In just 3–4 months, Taazaa turned a project struggling with backend issues for the last 6–7 years into a springboard for growth. The meticulous audit and implementation optimized logs, enabling real-time fixes for a massive catalog.
The future-proof design resulted in processing time dropping from 12 hours to 5.5 hours, while structured logging reduced log sizes from 350 MB to 8 MB. This enhanced logging plus an Apache Superset dashboard offers the client real-time insights into their operations.
The Taazaa solution resulted in faster processing and more cost-efficient operations, empowering the client to scale confidently. The catalog grew to 400,000 SKUs, with 1,000,000 in sight, and inventory value soared to $365 million. The vendor base expanded from 10 to 25, onboarding 15 new partners.
Taazaa’s delivery of nine impeccable releases, including one delivered in four weeks, resolved years of setbacks. With unwavering commitment and technical excellence, Taazaa not only revitalized a stalled project but also earned the client’s trust, paving the way for expanded partnerships with our design and AI teams.
