Shopify Customer Segmentation: RFM, Churn and CLV Guide

Your customer list contains at least four completely different audiences: people who buy constantly, people who used to, people who bought once and vanished, and a small group quietly responsible for most of your profit. Segmentation is just the discipline of telling them apart before you spend money talking to them.

Start with what Shopify gives you free

Shopify's native segment editor filters customers on attributes and events: location, tags, number of orders, amount spent, email subscription state. It is genuinely useful for static rules ("VIP tag AND subscribed") and it feeds Shopify Email directly.

Its limits show up fast. "Amount spent over $500" is not the same as "top decile of customers", and no native filter answers "who is overdue for their next order?". For that you need scoring, not filtering.

RFM: the scoring model that has never stopped working

RFM grades every customer on three axes: Recency (days since last order), Frequency (orders in the window), and Monetary (total spent). Each axis gets a score, usually 1 to 5 against your own customer distribution, and the combination lands each customer in a named segment. The classics:

• Champions (high on all three): protect them. Early access, thank-you rewards, referral asks. Never discount them; they were buying anyway.
• Loyal (frequent, moderate spend): grow basket size. Bundles, threshold free shipping, volume incentives.
• At Risk (good history, fading recency): the money segment. Win-back offers cost a fraction of the acquisition spend it takes to replace them.
• Lapsed (long gone): one strong reactivation attempt, then suppress them from paid audiences so they stop costing money.

The point of the names is action. If a segment does not change what you send, it is analytics theatre.

Churn risk: the forward-looking signal

RFM describes today; churn prediction models tomorrow. The core idea: every customer has their own rhythm. Someone who orders every 20 days and is now at day 45 is at risk even if their RFM still looks healthy. A churn model compares each customer against their own history, which is exactly the computation you cannot do with segment filters.

This is the job Miko Claude AI was built for: it computes RFM segments, churn risk, and predicted lifetime value for every customer automatically, keeps them updated, and explains each insight in plain English rather than dumping scores on you. Its comparison against Chupper covers how it stacks against other intelligence apps.

Lifetime value: deciding how much a segment is worth

CLV closes the loop: predicted future value tells you what you can afford to spend per segment. A $40 win-back discount is rational for a customer whose predicted CLV is $900 and irrational for a one-time $25 buyer. Segment-level CLV is also the honest way to judge acquisition channels, because channels that deliver Champions beat channels that deliver one-and-done buyers at half the CAC.

Wiring segments into your stack

• Email: sync segments to Klaviyo (or Shopify Email) as tags so flows target Champions and At Risk differently by default.
• Loyalty: segment movement should drive tier decisions; Champions belong in your top tier before they ask.
• Wholesale: the same at-risk logic applies to B2B accounts; a wholesale buyer whose ordering rhythm slows is revenue walking out slowly. Miko B2B's quiet-accounts check is this exact signal for trade customers.
• The admin itself: with Sidekick connected, "which customers are at risk of churning?" and "who are my top customers?" are one-line questions.