The eCommerce retention playbook: using your CRM after the first order

Acquisition costs rise every year. The CRM work that pays best in eCommerce happens after the first order.

Get order data into the CRM properly

Sync orders with line items, order value, first and last order dates, discount used, channel and returns. Retention analysis with only "customer created" as a data point is guesswork.

Segment by behaviour, not demographics

Four segments carry most of the value: one-time buyers within 60 days of purchase, repeat buyers, high-value customers, and lapsed customers past their expected reorder window. Each needs a different message and a different offer economics.

Time replenishment to the product cycle

Calculate the median days between orders per product category. Trigger the reorder sequence at 70–80% of that interval, not at a fixed 30 days. This single change usually outperforms any creative improvement.

Build second-order conversion as a programme

The largest retention gain is the first repeat purchase. A dedicated post-purchase sequence — usage guidance, complementary product, modest incentive on day 21–30 — moves the metric that compounds every other metric.

Win-back with honesty

For lapsed customers, acknowledge the gap, ask what changed and offer one clear reason to return. Test a no-discount version: many lapsed customers left for product or service reasons, and a discount does not address either.

VIP treatment that is actually operational

Identify the top decile by lifetime value and give them something real: early access, priority support routing in the service desk, a named contact. Tag it in the CRM so support and marketing behave consistently.

Measure what matters

Repeat purchase rate, time to second order, revenue per customer cohort by acquisition month, and lifetime value by acquisition channel. Cohort views expose whether your growth is durable or rented.

Deliverability underpins all of it

Retention runs on email. If SPF, DKIM and DMARC are not correct and list hygiene is loose, every programme above underperforms for reasons that have nothing to do with strategy.