Sourcing Accurate Retention Rate and Purchase Frequency Inputs From Shopify
A step-by-step walkthrough for sourcing retention rate, purchase frequency, and AOV from Shopify — including where the built-in reports mislead and which cohort export to trust before running an LTV model.
Quick answer
Don't use Shopify's headline 'Returning customer rate' for LTV math — it's a lifetime ratio, not a cohort retention rate. Pull the Customer cohort analysis report, export to CSV, and compute 12-month repeat rate and orders-per-customer per cohort. Use net AOV (after discounts, before shipping/tax) as your value input.
Sourcing Retention Inputs From Shopify
The operational process for extracting trustworthy retention rate, purchase frequency, and AOV figures from Shopify before feeding them into an LTV or retention-lift model.
Shopify exposes several customer metrics, but most stores plug the wrong one into LTV math. The headline 'Returning customer rate' on the Customers dashboard is a cumulative lifetime ratio across all orders ever placed — it conflates a 2019 buyer with last week's. For any model that compares cohorts or projects future revenue, you need cohort-aligned retention, purchase frequency measured over a fixed window, and an AOV definition that matches how you'll spend against it. This page walks through where to pull each input in Shopify admin, which defaults to ignore, and how to sanity-check the numbers before they land in the calculator.
The reason this matters: an LTV model is only as honest as the retention number you feed it. Overstate 12-month repeat rate by 5 percentage points and your blended LTV inflates by 15-25%, which quietly justifies CAC bids you can't actually afford.
Shopify's admin reports were built for merchandising decisions, not finance-grade LTV inputs. The numbers are there, but you have to know which report to open and which column to trust.
Where Shopify's defaults mislead
The Customers dashboard shows 'Returning customer rate' as a single percentage. This is total customers with 2+ orders divided by total customers ever — a lifetime cumulative ratio. It will keep climbing as your store ages, even if recent cohorts are getting worse.
The same dashboard's 'Average order value' is also lifetime-blended. If you ran an aggressive discount push six months ago, that period still drags the average down today, masking what a fresh acquisition is currently worth.
Don't paste the dashboard number into your LTV model
If you're modelling forward — for a paid-media bid, a subscription forecast, or a retention experiment — the lifetime 'Returning customer rate' is the wrong input. It will be 5-15 points higher than your current 12-month cohort rate on a store older than two years.
The report to actually use: Customer cohort analysis
In Shopify admin, go to Analytics → Reports → Customers → 'Customer cohort analysis'. This groups customers by acquisition month and shows repeat purchase behaviour over time. It's the only built-in report that lets you isolate a clean acquisition cohort.
Export to CSV. The default view aggregates by month, which is fine for stores doing 200+ first orders per month. Under that threshold, switch to quarterly cohorts — monthly noise will swamp the signal.
For each cohort, the columns you care about are: cohort size (customers acquired), repeat orders in months 1-12, and revenue attributed to that cohort. From these three, derive 12-month repeat rate, orders per customer, and cohort AOV — the three inputs the Retention Lift LTV Calculator expects.
Typical ranges to sanity-check against
Typical 12-month cohort retention ranges on Shopify by vertical
| Vertical | 12-mo repeat rate | Orders per customer (yr 1) | Net AOV (€) |
|---|---|---|---|
| Beauty & skincare | 32-48% | 1.6 - 2.3 | 45 - 75 |
| Apparel & accessories | 22-35% | 1.4 - 1.9 | 65 - 110 |
| Supplements / consumables | 40-58% | 2.0 - 3.1 | 35 - 60 |
| Home & lifestyle | 15-25% | 1.2 - 1.5 | 80 - 160 |
| Electronics & gadgets | 10-18% | 1.1 - 1.3 | 120 - 280 |
| Pet products | 38-52% | 1.8 - 2.6 | 40 - 70 |
If your Shopify export puts you well above these ranges, you've almost certainly grabbed the lifetime ratio instead of a cohort figure. If you're below, check whether subscription orders are being counted as separate purchases or rolled up under the parent order.
Aligning the time window
Pick a cohort window that has had time to mature. A cohort acquired 4 months ago hasn't had a chance to make a second purchase in many categories — using it will understate retention. For 12-month repeat rate, only count cohorts acquired 12+ months ago; for 90-day frequency, only count cohorts 90+ days old.
A clean default: average the four most recently matured quarterly cohorts. That smooths seasonality (Q4 acquisitions skew different to Q2) without dragging in three-year-old behaviour that no longer reflects how you sell.
Defining AOV consistently
Shopify's default AOV is gross — it includes discounts as positive line items and excludes shipping and tax. For LTV math, use net order value: gross merchandise value minus discounts, minus refunds, before shipping and tax. This matches the contribution margin denominator that CAC sits against.
If you run a high-discount calendar (sample sale, friends-and-family, BFCM), compute AOV both with and without those windows. The 'normal' AOV is what your steady-state model should use; the blended figure is what your annual finance forecast uses. Once these inputs are clean, feed them into the Retention Lift LTV Calculator to model the upside of a retention experiment.
Frequently asked questions
The dashboard figure is a lifetime cumulative ratio — total customers with 2+ orders divided by total customers ever. Cohort retention measures one acquisition group over a fixed window. On a store older than 2 years, the dashboard number is typically 5-15 points higher than your current 12-month cohort rate.
Yes. Analytics → Reports → Customers → Customer cohort analysis, then the 'Export' button gives you CSV. Shopify Plus accounts can also pull via the GraphQL Admin API using the customer object's order history, which is more flexible for custom windows.
Aim for 200+ acquired customers per cohort bucket. Below that, single-customer behaviour swings the percentage 1-2 points. If you're doing 50-150 first orders a month, use quarterly cohorts instead of monthly to get the count up.
It depends on your model. For raw retention behaviour, yes — they're real revenue. For predicting active customer choice, separate them: subscription recurrences inflate apparent retention because the customer isn't re-deciding each time. Most LTV models track both side by side.
Subtract them from both revenue and order count before computing AOV and frequency. Shopify's cohort report doesn't net these out by default — you have to pull the Refunds report and adjust. For apparel and electronics with 15-30% return rates, ignoring this overstates AOV materially.
The calculator takes three inputs: current 12-month repeat rate, average orders per customer, and net AOV. The Shopify cohort export gives you all three. Use the four-most-recent-matured-quarters average for each, and you have a defensible baseline to model retention experiments against.
Shopify identifies repeat customers primarily by email match. Guest checkouts using the same email across orders are correctly attributed. Different emails (a common B2C pattern) are recorded as separate customers, which slightly understates repeat rate — usually by 2-4 points.
Quarterly is enough for steady-state LTV models. Refresh after major catalogue changes, a pricing shift, or any retention experiment graduates. Monthly refreshes introduce more noise than signal unless you're scaling fast.
Pull cohort data per market and weight by acquisition volume. Repeat rates and AOV vary 20-40% between markets due to shipping economics and currency. A blended global figure will mislead any per-market CAC bid decisions.
Yes. Customer cohort analysis is available on Shopify, Advanced, and Plus plans. Basic Shopify shows a more limited view. If you're on Basic, the Orders export plus a pivot in your spreadsheet gets you the same numbers — just more manual.
Test ideas before you ship them
Run unlimited A/B tests, attach hypotheses to outcomes, and build a searchable archive of what works — and what doesn't.