Basket Composition
Date: 2026-08-16
What’s actually in the order. Mix moves contribution more than price does, and it moves silently — a stable AOV can hide a margin collapse if the composition underneath it shifted.
What it is
Basket composition is the mix of products, categories and margin profiles within an order.
Two baskets, same value, different businesses:
BASKET A BASKET B
1 × premium item £50 1 × entry item £20
2 × accessories £30
─────
value £50 value £50
contribution £22.50 contribution £9.00
Same AOV, two and a half times the contribution. Any metric that stops at revenue treats these as identical.
Why mix beats price
Price changes are visible, debated and slow. Mix changes are invisible and continuous — driven by merchandising, search ranking, stock availability and which campaign ran.
month 1 60% high-margin / 40% low → contribution £15.00/order
month 2 45% high-margin / 55% low → contribution £12.75/order
revenue flat. Contribution down 15%.
Nothing was discounted. The mix moved, and only a per-category view shows it — Simpson’s Paradox in its commercial form.
What drives it
- Merchandising and collection ordering. What’s first in a listing sells most, and the ordering is rarely set by margin — Merchandising
- Search and filter defaults. A default sort by price-ascending pushes mix down permanently
- Stock availability. An out-of-stock high-margin line diverts demand to a lower-margin substitute, or loses it — Stockouts and Availability
- Promotions. A discount on one category shifts mix towards it for the duration and often afterwards
- Acquisition channel. Paid social often brings entry-level demand; branded search brings intent for specific items — Channel Mix
- Free-shipping thresholds. Customers pad baskets to reach them, usually with cheap accessories — which is the intended mechanism and a mix cost — Shipping Thresholds
What to measure
Beyond AOV and units per order:
| Metric | Reveals |
|---|---|
| Contribution per order | The number AOV was standing in for |
| Category share of revenue | Mix drift over time |
| Category share of contribution | Which categories actually pay |
| Attach rate | Proportion of orders including an accessory |
| Single-item order share | How much of the business is one-item purchases |
Category share of contribution is the one to watch. A category can be 30% of revenue and 8% of contribution, and every decision made on revenue share will be wrong about it.
Where it matters for CRO
The connection people miss: conversion optimisation changes mix.
- A test that raises conversion by promoting entry-level products lowers contribution per order
- A cross-sell module that works raises attach rate — check whether it raises attach of margin or of cheap padding
- A shipping-threshold change shifts basket construction directly
Which is why contribution per order belongs as a guardrail on any test touching merchandising, pricing or basket — a conversion win with a contribution loss is a real and common outcome — Guardrail Metrics, Contribution Margin.
Practical
- Report contribution share by category alongside revenue share, monthly. The divergence is the finding
- Watch attach rate as the cheapest mix lever — it adds units without discounting
- Check what the threshold is padding with. If customers reach free shipping with 20%-margin accessories, the threshold is buying volume at the expense of mix
- Segment new versus returning. Returning customers buy differently, so an acquisition surge changes mix with no merchandising cause