Building a "Gachapon" Inventory Bundling Tool
2,500 mystery boxes, hundreds of SKUs, zero duplicates. An algorithm and operator workflow for "fair randomness" at scale.
| Timeline | Feb 2022 |
|---|---|
| My Role | Engineering, UX design |
| Domain | Ops tooling, Optimization, Ecommerce |
| Tech | TypeScript, Shopify API, Google SSO, Node.js |
Challenge
Game Grumps’ merch team had hundreds of SKUs aging out of relevance and tying up capital. Mystery boxes were the play to bundle slow-moving inventory at a discount, but assembling them by hand was tedious and didn’t scale.
The real constraint was fairness: every box needed to land within a tight value range, avoid duplicate items, and still feel worthwhile to the customer. That had to hold across 2,500+ boxes and hundreds of SKUs.
Approach
I designed and built an internal tool that let non-technical operators configure a run, pull live Shopify inventory, generate pack lists against those constraints, and export the results for packing — all behind Google SSO with run history and summaries.
Impact
- ~$202k of slow-moving inventory routed through the first batch with successful sales and positive customer feedback.
- 2,500+ boxes generated per run, up from a manual ceiling of a few dozen.
- Fully operator-owned after handoff. The merch team continues to run subsequent batches without developer involvement.
