E-commerce Platform
Lifted checkout conversion 30% and average order value 20% on an e-commerce platform for car-detailing products, built on AWS serverless.
01.2019 — 12.2020Contract React developer — LTV Commerce.
ReactAWSE-commerceRecommendations
- Recommendation engine increased conversion rates by 30%.
- Upselling and cart recovery raised average order value by 20%.
What I did
- Architected a full-featured e-commerce platform for car-detailing products on AWS serverless.
- Built 10 custom checkout templates with a dynamic product-recommendation engine.
- Implemented cart-abandonment recovery and real-time upselling flows.
The problem
LTV Commerce needed a full e-commerce platform for car-detailing products — not a storefront theme, but a system they could run promotions and experiments on: custom checkout flows, product recommendations, and recovery of abandoned carts.
My approach
- Platform architecture — I architected the platform on AWS serverless, which matched the traffic profile: bursty campaign-driven spikes with quiet periods in between, and no servers to babysit.
- Checkout as a product — I built 10 custom checkout templates so the client could match checkout flow to campaign and audience instead of forcing every customer through one funnel.
- Recommendations & recovery — I developed a dynamic product-recommendation engine for cross-sells at checkout, plus cart-recovery and upselling flows that bring back customers who left mid-purchase.
Challenges
- Checkout is hostile territory for experiments — every variation had to keep the payment path boring and reliable while the merchandising around it changed; templates share one hardened order/payment core underneath.
Outcome
A full-featured store the client could actively tune — checkout variants, recommendations, and recovery flows — running on AWS with serverless economics.