Getting started.

Install Retrics from the Shopify App Store and choose the store you want to connect. Shopify walks you through a standard OAuth approval screen listing exactly what Retrics can access.

Retrics connects read-only. We read orders, customers, and products through Shopify's API to compute analytics — we never write to your store, never touch payment credentials, and never message a customer on our own. Email addresses are hashed before they leave Shopify's protected-customer-data boundary.

Approve the scopes and you land in your new workspace. There is nothing to install on your storefront and no theme code to edit.

As soon as you connect, Retrics begins importing your order, customer, and product history. This first sync is the one part of onboarding that isn't instant — it reads your full backlog so the analytics are built on real history, not a sample.

You can close the tab and come back. When the import finishes, your workspace opens and every screen below fills with your own numbers. From then on Retrics keeps itself current in the background as new orders arrive.

With your history loaded, Retrics computes the foundations: customer lifecycles (new, repeat, at-risk, lapsed, won-back), purchase cohorts by first-order month, and a health read on repeat rate, retention, and revenue concentration.

It then scores each customer's lapse probability against your store's own patterns and surfaces the first set of opportunities — specific, ranked places where retention is leaking or a segment is worth re-engaging. Nothing here is a guess pulled from another store; every number is derived from your data.

Home opens on the Morning Brief: a short, plain-language digest of what changed since yesterday. It flags cohorts drifting off pace, customers crossing into at-risk, health metrics moving, and the opportunities worth your attention today.

The Brief is designed to be read in under a minute and to point you at the one or two things that actually matter, rather than a wall of dashboards. Everything it mentions links straight to the detail behind it.

Open an opportunity and Retrics shows you the reasoning: which customers qualify, why they were selected, the segment's size, and the retention math behind the estimate. When you open it, the audience freezes — a snapshot is taken so the list you review is exactly the list you act on, unaffected by later syncs.

Review the members, adjust if you want, then approve. Approving is your explicit action — Retrics never contacts your customers itself. From the approved snapshot you export a clean CSV or hand the audience off to your ESP (Klaviyo, Mailchimp, and the rest) to run the campaign in the tool you already use.

After a campaign runs, the recovered-revenue ledger tracks what came back. For intervention audiences — overdue and win-back sends — Retrics measures against a holdout so the number reflects lift over the customers who would have returned on their own, not gross returns credited to the send.

The result is a running, defensible tally of revenue Retrics helped recover, with the methodology shown so you can trust it. Predictions and estimates are statistical instruments, not guarantees; the ledger is where they get checked against reality.