Refund Recovery

When most customers who get a refund never order again and support keeps fielding when will my money arrive questions, this journey is worth building. Before you build it, confirm that refund approved, refund sent and funds arrived can be read from the payment source against the same order, that expected timing ranges exist per payment method, that you can generate alternatives to the returned product, and that email subscription and in-app message permission are stored. If refunds come from one product defect or a misleading product page, remove that cause first; an alternative with the same flaw will not bring the customer back.

Incremental repurchase rate vs control

The journey's contribution is how many more refunded customers who receive alternatives and the incentive order again than the control group that receives refund information only; an opened alternatives email or a product page click is not a recovery on its own.

How it is counted: Count every refunded purchase assigned at entry to either group as the denominator; the numerator is the customers who purchase again within the same follow-up window after the refund completes, and the gap between the two groups is the journey's effect. Delivery or clicks on the alternatives message do not change the denominator; show customers handed to support because the money never arrived separately in both groups.

Also track

• Alternative product view rateAmong customers who received the alternatives message, the share who viewed at least one suggested product; the intermediate step before repurchase, and a sign that the suggestions fit the return reason.
• Refund completion and non-arrival shareTrack the share of refunds that reach the account within the expected range and the share handed to support because they did not, per payment method; a rising non-arrival share points at the refund process, not at the commercial recovery.
• Contribution of the incentive branchRead repurchases after the alternatives message and after the incentive email separately; if the incentive brings no extra orders, or only pulls forward orders the alternatives would have won, limit the offer.
• Unsubscribes and support contactsAfter the alternatives or incentive message, watch email unsubscribes, in-app message opt-outs and refund-related support contacts; recommending products to a customer whose refund went wrong turns into a complaint.

• Trigger and entryThe journey starts for an order when its refund is approved and sent to the bank. Customers are split at random into the journey group and a control group that receives the refund status emails but not the alternatives or the incentive; refund information is withheld from nobody.
• Who must not enterOne refunded purchase is one journey instance: no second flow opens for the same order, and a customer who has already purchased again after the refund gets no recovery message. A payment that failed to come in belongs to Payment Failure Recovery; the two journeys share a payment record and neither announces the other's movement.
• Timing and channelsFirst comes the refund approved email, then the refund on the way email 1 to 3 days later, and the completed email after 3 to 7 days by payment method, once the money arrives. With no repurchase 3 to 5 days after completion, alternatives go in-app in an active session, otherwise by email; with none 2 to 3 days later, one incentive email goes out.
• Exit and handoffA repurchase ends the journey and cancels the pending alternatives and incentive. If the money has not arrived, an email with what to check and a support link goes out and the record moves to refund support until the payment is resolved; after the incentive, or with none defined, the journey simply ends.
• Frequency and consentEach refund gets at most three status emails, one alternatives message and one incentive; no alternatives are offered while the money has not arrived. Status emails are transactional; alternatives and incentive are marketing, need an email subscription or in-app message permission, and the incentive goes out only where an offer is defined for that customer.

1. 1A repurchase rate in the journey group that does not beat the control group means the alternatives reached customers who were coming back anyway; views and clicks do not make up for that. Weigh the order gap between the two groups against the change in unsubscribes, and keep reporting the refund status and the commercial recovery as two separate results, never reading one as proof of the other.
2. 2Change the waiting ranges, the channel selection or the defined incentive rule and the result no longer applies. An order placed later by a customer handed to refund support is not this journey's win, and a payment recovered by Payment Failure Recovery is not counted here either.
3. 3When refunding customers are few and repurchase is rare, a handful of orders changes the picture; enter each group's customers and repeat orders into the A/B test significance calculator to see whether the gap is significant, and if it is not, extend the follow-up window past the incentive email and run the same read again.