Signal
A second skip in a row
Daniel skips his coffee delivery for the second cycle. Last month's delivery arrived two days late and he rated that order three stars.
Solution · Churn prevention
Catch churn while it is still a signal.
Ayra prevents ecommerce customer churn by treating it as a sequence of signals rather than a cancellation: a skipped delivery, a late order, a low review, an unspent balance, an open ticket. Each one updates the customer's profile and can open a churn-risk opportunity, and the engine answers with one next best action: a pause instead of a cancel, a fix before a discount, or nothing while a complaint is open.
Private beta. Free migration from the apps Ayra replaces. Pricing
Churn risk
- Daniel P1
SavePause 1 month · or 250g bag
- Mara P2
SaveCard update · retry Thu
- Jon Hold
SaveFix first · marketing held
- Elif P2
SaveHand to winback
No discount until a pause, swap or fix has been offered.
01 The merchant problem
By the time churn shows up in a report, it already happened.
Churn rate is a lagging number. The leading signals are spread across the subscription app, the help desk, the review tool and the order history, and nobody is adding them up per customer.
-
Skips are read as losses, not messages
A skipped subscription delivery says something about timing, product or budget. The usual response is a discount, which answers none of those.
-
Complaints and marketing never meet
The review request goes out while the delivery complaint is still open, because the help desk and the review app have never heard of each other.
-
Risk is scored per app
The subscription app's dunning, the loyalty app's inactivity flag and the email tool's engagement score are three risks that never meet on one customer.
-
Churn is counted on subscriptions only
One-time buyers who quietly stop ordering are not counted as churned anywhere.
What to measure
- Open churn-risk opportunities
- Customers currently in CHURN_RISK, by the evidence that opened it: skips, failed payment, low rating, open ticket or lapse.
- Subscription churn and saves
- Cancellations, pauses and swaps per month, and the share of cancel attempts that ended in a save.
- Cohort retention curve
- The share of each first-order cohort still ordering after 30, 60, 90 and 180 days.
No benchmarks. Ayra compares your store to itself, by cohort and segment.
02 How Ayra solves it
How Ayra catches churn early.
- 01Explore Signals
Churn risk from the customer's own signals
Skips, failed payments, late deliveries, low ratings, open tickets and lapses are typed signals on one profile. Risk is a rule you can read, not a hidden score.
- 02Explore Subscriptions
Save before cancel
Save Engine meets a cancel attempt or a second skip with a pause, a swap or a date change before any discount.
- 03Explore Reviews
Fix first, then ask
Trust Recovery holds review requests and promotions while a complaint or refund is open, and releases them after it is resolved.
- 04Explore Subscriptions
Dunning that does not read as churn
Card-update flows with retries on the customer's preferred channel, before a subscription is marked as lost.
- 05Explore Analytics
Churn you can see by cohort
Retention curves, subscription churn and MRR, and the revenue kept by each save, in one report.
03 The workflow
One customer, start to finish.
Daniel, subscriber · 8 months. Signal to outcome, one decision, every reason recorded.
Daniel
subscriber · 8 months
5 stages
Customer
Eight months, first wobble
Subscriber since February, $392 lifetime value, 310 points, no open tickets today, email preferred. Risk rule: two skips plus a rating of three or lower inside 60 days.
Opportunity
CHURN_RISK opened, P1
Evidence: two consecutive skips and a low rating. It expires before his next billing date, because after that the decision is his.
Action
A pause and a smaller bag, no discount
Save Engine offers a one-month pause or a switch to the 250g bag, by email, with a note about the late delivery. The 20% 'don't go' discount in the old flow is not sent.
Outcome
Swapped, not cancelled
Daniel switches to the smaller bag and keeps the subscription. The save is credited to Subscriptions, and his next review request waits for the next delivery.
04 Use cases
What it looks like in your store.
Failed payment 01
A card expires the day before renewal.
Ayra decides
Dunning sends the card-update link on the customer's channel and retries on a schedule. The subscription is not marked churned until the retries are done.
Low review 02
A first-time buyer leaves a two-star review about sizing.
Ayra decides
The review is routed as a support issue first. Marketing to that customer is held until it is resolved, then a size-exchange note goes out instead of a promotion.
Quiet lapse 03
A one-time buyer passes twice their expected gap with no signal at all.
Ayra decides
A lapse opportunity opens and hands over to winback. Churn on one-time buyers is counted, not ignored.
The modules behind it
All features- Explore Subscriptions
Subscriptions
Subscriptions on Shopify's own contracts, with retention built in.
- Explore Reviews
Reviews & UGC
Reviews, photos and Q&A that live on the customer profile.
- Explore Flows
Lifecycle Flows
Email, SMS and WhatsApp journeys that respect the customer's timeline.
- Private betaExplore Analytics
Retention Analytics
Repeat rate, LTV and cohorts, with revenue attributed per module.
Read next
Switching?
Dated, sourced comparisons against the apps Ayra replaces. Migration is free during private beta.
05 FAQ
Questions merchants ask first.
How does Ayra detect churn risk?
From typed signals on the customer profile: consecutive skips, failed payments, late deliveries, low ratings, open tickets and lapses against the customer's own order gap. The rules are explicit and editable.
Is this only for subscriptions?
No. Subscription churn is where the signals are richest, but lapse detection covers one-time buyers too and hands them over to winback.
Does it predict churn?
At launch it is rule-based. Churn prediction is layered on later, on the same guards and the same timeline.
What does the Save Engine offer?
A pause, a swap, a date change or a smaller size before any discount, in an order you set.
Will a customer with a complaint still get marketing?
No. Trust Recovery holds retention actions while a ticket or refund is open, and the hold is visible on the timeline.
06 Related solutions
Same engine, different outcome.
- See the workflow
Customer winback
Detect lapsed customers from their own rhythm and bring them back once.
- See the workflow
Retention automation
One loop that decides per customer instead of a stack of triggers.
- See the workflow
Increase repeat purchases
Turn deliveries, wishlists and balances into second and tenth orders.
Every solution runs the same loop. How it works
Private beta
See your churn signals before the churn.
Private beta merchants get churn-risk rules on their own subscribers and buyers, with every save and hold recorded on the timeline.