MBA
New feature

Out of stock?
Return a ranked answer, not silence.

Most stores hit an out-of-stock SKU and either show a sad empty-state, hide the product, or hand-pick generic replacements once a quarter. MarketBasketAnalysis ships a substitution engine that mines your basket history and answers the question every agent and shopper actually asks: what else fits the same purpose?

Live on Shopify · BigCommerce · WooCommerce · Magento · OroCommerce · Surfaced through MCP for agents

How it works

Three signals, one ranked answer.

The substitution engine runs on the same basket data that powers our cross-sell rules. Same mining, different question.

Context similarity

Two products are substitutes if they appear in similar baskets (paired with the same other items) but rarely co-occur themselves. Jaccard overlap on per-item context vectors, computed at mining time.

Inverse co-occurrence

True substitutes are bought one-or-the-other. The composite score subtracts direct co-occurrence so “customers who bought BOTH” pairs (which are cross-sells) don't pollute the substitution result.

Metadata boost

Light category + vendor boost on top of the data signal. Returns a reason code (context_similar / category_match / vendor_match) so the consumer can re-rank or filter.

Where it shows up

One feature, six surfaces.

The substitution engine populates a per-job table on every platform we ship. Each surface consumes it differently.

MCP tool

find_substitutes(product_id, limit?) — part of Basket AI's Discovery Agent, available on Claude Desktop, Claude Code, Cursor, Cline, OpenAI Agent SDK.

REST API

GET /api/v1/substitutions?product_id=X — Bearer auth, JSON in/JSON out. Same shape as our recommendations endpoint.

Storefront fallback

When checkout hits a backordered line, surface the top substitute inline. We provide the data; your storefront decides the UX.

B2B quote workflow

Plug into the RFQ / requisition flow so agents auto-resolve unavailable SKUs to the closest fit, no human escalation.

Ready to turn your order data into revenue?

Install on your platform in under 10 minutes. Or book a consulting call and we'll do the launch for you.