The exchange-first playbook: how to build the ladder

Everyone nods along to exchange-first until they open the config panel. Then the questions start. Which rung comes first? Do we ship the replacement before the old item lands? What happens when a shopper in the EU clicks refund on day three? Should sale items get store credit only? The philosophy is easy to agree with and hard to wire up correctly. This post is the wiring diagram. It assumes you already believe a refund is a sale you lost, and it focuses entirely on how to build the outcome ladder so it actually holds revenue instead of annoying the people you are trying to keep.

We will go rung by rung in the order you should configure them, then cover the instant versus deferred decision, the portal UX that makes exchange the obvious choice without hiding the refund, per-policy and per-product rules, the legal edges you cannot cross, how the ladder compounds with size recommendation, and the four metrics that tell you whether any of it is working. Every section is a setting you can turn on this quarter.

Configure the ladder from the top rung down

The outcome ladder is not a menu you present all at once. It is a fallback chain. Each rung fires only when the rung above it is impossible for this specific item and this specific shopper. Configure it top-down, because every rung you get right at the top removes volume from the rungs below.

Rung 1: right size, same product

This is the highest-value outcome and the easiest to earn, because the shopper already chose the product. The only reason they are here is fit. Wire this rung to your live inventory and your size model so the portal can say the exact replacement size and show it in stock. If you cannot check stock in real time, this rung collapses into a promise you might break, and a broken exchange promise sends people straight to refund. Gate nothing here except availability.

Rung 2: variant swap

Same product, different color or style variant. The shopper likes the line but not the specific version they received. Configure this to surface sibling variants of the returned SKU, priced at parity so the swap is one tap with no payment step. Price deltas kill this rung; if the variant costs more, you have introduced a checkout and lost the momentum.

Rung 3: product swap

The product missed, but the brand did not. Let the shopper exchange for anything of equal or greater value, collecting the difference at the top end. This is where a good recommendation surface earns its keep, because you are asking the shopper to reshop. Constrain it to same-category or all-catalog depending on how much reshopping friction you want to absorb.

Rung 4: store credit plus a bonus

Nothing in the catalog fits right now, but the money can still stay in the store. Attach a real incentive, typically five to fifteen percent on top of the item value, and issue it instantly. The bonus is not generosity, it is math: retained revenue at a fifteen percent premium still beats cash out the door at one hundred percent. Make the bonus visible before the shopper chooses, not after.

Rung 5: refund, and only here

Refund is the floor, not a hidden trapdoor. It stays reachable on every screen. The entire point of the four rungs above is that most shoppers find a better answer before they reach this one, not that they are prevented from reaching it. If your refund rate does not fall after launch, the problem is almost always that rungs one and two are not wired to live inventory.

Instant versus deferred: pick per profile, not per store

A deferred exchange ships the replacement only after the returned item arrives and passes inspection. It carries almost no inventory or fraud risk, but the shopper waits days, and every day of waiting nudges them toward asking for a refund instead. An instant exchange ships the replacement the moment the request is made and collects the old item separately. The experience is dramatically better and lifts exchange conversion, but it exposes you to fraud and double-shipment risk.

The mistake is treating this as one global switch. Configure it per shopper profile. Offer instant exchanges to trusted accounts, judged by order history, low prior return rate, and payment standing, and hold new or high-risk profiles to deferred. A card pre-authorization on the instant path covers the item value until the return lands, so you get the conversion lift without eating the fraud.

Portal UX: make exchange obvious, keep refund honest

Done wrong, exchange priority reads as a company hiding your money from you, and that torches trust faster than any refund ever would. The goal is to make the exchange more attractive, never to make the refund harder. A few rules that decide whether the portal converts or frustrates:

• Ask the reason first. Fit, color, changed my mind, and defect each route to a different top rung. A too-small return should land on the exact right size, not a generic menu.
• Show the answer, not a form. If the size model knows the fit, the portal should say for your measurements this runs a size up rather than asking the shopper to guess again.
• Put the perks on the choice screen. Instant approval, free return shipping, and the credit bonus should be visible at the moment of decision, not buried in a confirmation email.
• Never bury the refund. Keep it on the same screen, plainly labeled. A refund that takes five clicks to find is a support ticket and a chargeback waiting to happen.
• One screen, one glance. Each rung should show what the shopper gets in speed, money, and bonus, so the comparison is instant.

The refund button is not your enemy. A refund the shopper had to fight for is.

Per-policy, per-product, and legal-basis rules

A single global ladder breaks on contact with real catalogs. Final-sale and clearance items should skip rungs one through three and offer store credit only. Hygiene-sensitive categories may block same-product exchange entirely. High-margin lines can afford instant exchange and a fatter credit bonus; thin-margin basics cannot. Build the ladder as a default and then layer overrides by product tag, collection, and price tier.

The legal basis is not optional and not something you can design around. In jurisdictions with a statutory withdrawal right, the EU distance-selling rules being the common example, the shopper is entitled to a cash refund within the cooling-off window, and no ladder may obstruct it. The clean pattern is to detect the legal-basis return and keep refund fully first-class on that path, while still offering the exchange and the bonus as a genuine alternative. You are competing for the shopper's choice, not removing it. Anything that removes the legal option is a compliance problem, not a returns strategy.

1. 1Set the global default ladder for standard, in-policy returns.
2. 2Override by product tag for final-sale, hygiene, and margin tiers.
3. 3Override by profile for instant versus deferred eligibility.
4. 4Flag legal-basis returns and guarantee the statutory refund path.
5. 5Log which rung every request resolved on, so you can tune the rules with evidence.

Where the ladder compounds with size recommendation

The ladder holds revenue on this return. A size model stops the next one, and the two feed each other. When a shopper starts a fit exchange, the portal captures a structured signal, the item ran small, in the exact size they owned. That signal does two jobs. It picks the correct replacement size for this exchange, so rung one is accurate and does not bounce back as a second return. And it updates the fit-graph, so the next shopper buying that SKU sees the corrected recommendation before they ever check out.

That is the loop a refund can never start: better recommendation leads to accurate exchange leads to fewer second returns leads to richer fit data leads to better recommendation. Configure the two together or you get half the value. A ladder without a size model keeps guessing at the replacement; a size model without a ladder has no exchange event to learn from.

The four metrics that prove it works

Return rate alone will mislead you, because it does not care how a return resolves. Track how returns end, not just that they happened.

Read them together. A high exchange success rate with a rising second-return rate means the ladder is holding shoppers but shipping the wrong replacement, which points straight at the size model. A healthy outcome distribution skewed to rungs one and two means your inventory wiring is solid. Watch the lifetime value of shoppers who exchanged against those who refunded; the exchangers are almost always worth more, which is the whole argument in one number.