Boutique basket analysis can help you check whether two clothing options are actually bought together before you commit to more wholesale stock. A top and trousers may look convincing in an outfit photograph, but that does not establish that customers purchase both. Equally, a popular item can appear in many paired purchases simply because it appears in many orders overall.
This is a suggested transaction-review method for UK boutique buyers, not a sales forecast or a claim about any particular reporting dashboard. It adds a specific question to your buying review: does the observed pairing merit closer investigation, or are you mistaking general popularity, a promotion or a compulsory set for independent customer choice?
Define a basket before counting clothing pairs
Use one completed retail transaction as the basket, with its original reference and selling channel. An online order dispatched in two parcels remains one order, not two baskets. Duplicate exports and payment records must not create extra purchases. Do not combine separate purchases by the same customer into an invented transaction, or collect extra personal information for this exercise.
Choose a dated sales window and record the extraction cut-off. Exclude cancelled and wholly unfulfilled orders. For partly fulfilled orders, count only completed items and retain a note of what was not supplied. Keep later returns and exchanges linked to the original transaction rather than quietly rewriting what was bought together.
For each item, preserve the exact style, colour, labelled size and selling unit. Decide whether the question concerns exact options or broader styles before aggregating. A style-level pairing does not prove that the available colours coordinate or that the size combinations serve the same wearer.

Separate optional purchases from sets and offers
A co-ord sold as one complete set is one selling unit for this review. Its top and trousers appearing on separate component rows do not establish that customers independently chose both. Preserve the original set identity and do not turn its components into a successful cross-selling pair.
Mark transactions involving compulsory bundles, multi-buy discounts or a particular outfit promotion. They may be useful evidence about that offer, but they are not interchangeable with purchases at ordinary separate-item prices. If several garments were shown together, record that presentation too. The records establish an association, not that the display caused the purchase.
Count a selected option at most once per basket for the pairing calculation, even when a customer buys two units. Keep unit quantities separately for stock planning. Otherwise a single large order can look like several independent pairing decisions.
A basket example: check the comparison group
Hypothetical example, not Catwalk trading data: a boutique reviews 40 completed transactions in one channel during a period when options A and B were both available. The options are independently sold garments, not components of a set. There are no cancellations, returns, exchanges, compulsory bundles or promotional differences in this simplified example.
- 10 baskets contain A. Of these, 6 also contain B and 4 do not.
- 30 baskets do not contain A. Of these, 18 contain B and 12 do not.
- Across all 40 baskets, 24 contain B; 6 contain both A and B.
The share of A baskets that also contain B is 6 ÷ 10 × 100 = 60%. That may sound persuasive on its own. However, the share of non-A baskets containing B is 18 ÷ 30 × 100 = 60% too. Within this small extract, B is no more frequent in baskets containing A than in the comparison group.
The pair appears in 6 ÷ 40 × 100 = 15% of all reviewed baskets. Looking in the opposite direction, only 6 ÷ 24 × 100 = 25% of B baskets contain A. These percentages answer different questions. Do not swap their denominators, describe them as unit sell-through or average them into a single score.
The correct conclusion is limited: these records do not show a stronger observed association with B when A is present. They do not prove that the garments are unsuitable together, that no customer wanted an outfit, or that future demand will match these percentages. If a comparison group is empty, its percentage is undefined, not zero.
Check whether customers had a comparable choice
Before interpreting a pairing, check that the exact options could be bought at the same time in the selected channel. One size selling out can prevent a pair even while both style pages remain visible. Use the boutique stockout analysis guide to distinguish known availability from unknown intervals. Recalculate within an explicitly selected common window where the records support it; keep the original whole-window result alongside it.
Also check prices, promotions, time since launch and presentation. Comparing shop purchases during an outfit event with ordinary online orders introduces several differences at once. Split those groups before drawing conclusions, and show the raw basket counts beside every percentage. A striking percentage from a handful of transactions remains limited evidence.
Record returns separately: did the customer keep both items, return one or return both by the review cut-off? An exchange can change the colour or size without creating a second independent outfit purchase. Where the original sale cannot be linked reliably, mark the retained-pair result unknown. The clothing returns review explains how to retain the event history without guessing the reason or cause.
Turn the pairing into a bounded wholesale buying question
Start by checking existing and confirmed incoming stock. If both garments are already available, the useful next action may be a clearer outfit presentation or a staff question, not another wholesale order. A frequent pairing also does not establish that you need equal quantities of the two garments: each may have substantial demand on its own.
For a proposed addition, write the unmet outfit requirement rather than buying a visually similar garment and transferring the old results to it. The wholesale clothing buying brief separates a genuine assortment gap from missing product information. Verify the actual colour relationship, fit compatibility and independently purchasable options. Two garments bought in the same order need not have been intended for the same wearer or occasion.
Then expand the proposed purchase into its actual pieces or complete packs. Do not assume a supplier will split a pack, mix colours or accept unsold trial stock back. Check the exact product terms and the full payable commitment against the dated wholesale buying budget. A basket percentage is not an order quantity, a margin calculation or permission to spend.
Keep a review note that another buyer can reproduce
Save the source extract, transaction-selection rule, dates and cut-off; the exact A and B definitions; basket counts for both, A only, B only and neither; unit quantities separately; and any set, promotion, availability or return exclusions. Keep source references in your existing controlled business records rather than copying customer contact details into a general buying sheet.
Finish with one next action and its reason: investigate the pairing, improve presentation of existing stock, assess a specific complementary candidate or make no purchase. If you run a presentation trial, record what changed and compare reasonably similar periods without promising a causal sales result. Preserve earlier extracts when later returns revise the picture. Useful boutique basket analysis makes an outfit-buying hypothesis testable; it does not turn attractive combinations into guaranteed wholesale demand.