10,000 SKUs, 60 category leaders, one operating system.
The client acquires Amazon-marketplace brands and runs them at scale. The portfolio reached 10,000+ products, 60+ category leaders, and a $1B valuation. None of that is feasible without a unified product-data platform.
Three operating constraints: unified product data across thousands of SKUs; real-time analytics for buy-vs-pass and pricing decisions; agility during the Amazon events (Prime Day, Black Friday, Cyber Monday, Lightning Deals) where everything spikes at once.
A data ingestion platform with Amazon at the centre.
Runtime built a data ingestion platform integrating Amazon MWS APIs alongside third-party enrichment (Keepa, SentryKit). The system pulls historical data on acquisition and streams real-time updates, powering inventory forecasting, pricing optimisation and anomaly detection.
The hard problem isn't any single data source — it's reconciling the flow when fifty stores, ten thousand SKUs and a dozen enrichment vendors all push at the same time. The architecture is the product.
A data platform that ingests, enriches, decides.
The product is a data platform — and the stack reflects that. Two backend frameworks for different services, a React frontend for operations dashboards, and integrations into Amazon's MWS APIs alongside third-party data enrichment.
Operating visibility through a $1B run.
The platform enabled rapid scaling across acquired brands while maintaining operating visibility and competitive positioning. Prime Day and Black Friday went from "all-hands fire-fight" to "automated pipeline doing its job".
Team Integration at scale.
Fifteen Runtime engineers embedded with the client's product organisation across five years. Multi-leader structure, same Jira board, same release cadence. Long engagements, not transactional handoffs.
