understanding•5 min read
Problem Statement & Context
What product bundling means in large retailers
Design Product Bundling at Retail Scale
Retailers like Amazon, Best Buy, and GameStop sell bundles—pre-defined kits (console + game + controller), mix-and-match combos (pick any 3 accessories), and algorithmic bundles surfaced as "Frequently bought together." Your interview answer must connect merchandising intent, pricing rules, and inventory truth without treating a bundle as a single opaque box unless fulfillment truly ships that way.
Why bundling is its own system-design problem
A bundle is not just a discount coupon on multiple SKUs. It introduces:
- Composite availability: the bundle is sellable only when component constraints are satisfied (all required SKUs in stock, or substitute rules apply).
- Dual identity: merchandising sees one bundle offer; warehouses often see component picks; finance may need both views for revenue recognition.
- Pricing surfaces: list price, bundle price, member price, and stackability with cart-level promos can conflict if you recompute naively at checkout.
Personas and journeys
| Persona | Goal | System pressure |
|---|---|---|
| Shopper | One-click "Add bundle" with clear savings | PDP latency, accurate "in stock" badge |
| Merchandiser | Launch seasonal kits without engineering deploys | Rule DSL, preview, rollback |
| Warehouse | Pick components efficiently | Exploded order lines, kitting vs separate picks |
| Finance | Correct tax and refund allocation | Line-level audit trail |
Scale anchors (stated assumptions)
- 80M monthly active shoppers on the storefront
- 12M bundle PDP views/month; 8% add bundle to cart
- 500K active bundle definitions; 20% change weekly during campaigns
- Peak Black Friday: 15× read traffic on bundle catalog APIs
These numbers drive cache sizing, event throughput for bundle rule updates, and the need for precomputed availability on hot bundles.
Key Highlights
- •A bundle is not a multi-SKU coupon: it carries composite availability, dual identity (one merchandising offer vs many component picks), and conflicting pricing surfaces that break if recomputed naively at checkout
- •Three archetypes need different machinery: fixed kits (console + game), mix-and-match N-of-M combos, and algorithmic 'frequently bought together'
- •Scale anchors: 80M MAU, 12M bundle PDP views/month at 8% add-to-cart, 500K active definitions with 20% weekly churn, 15x Black Friday read peak
Bundle-specific
Anchor Problem Statement & Context to component explosion, not generic cart patterns.
Numbers
Cite ATP min() logic and pro-rata allocation when discussing this section.
Section Rescue Kit
Buzzwords to use:
Virtual bundle SKUComponent explosion
Safe statements:
- "I would treat bundle availability as the minimum available quantity across required components, with explicit rules for optional add-ons."
- "Pricing can be snapshot at add-to-cart so bundle discounts do not drift during checkout."