A weak repeat rate does not automatically mean the brand needs more lifecycle emails.
The customer may not be due to return. They may have bought a product that rarely repeats. The first experience may have disappointed them. The product may work, but no habit formed. Or acquisition may have brought in a large group of one-time deal seekers.
All five can produce the same dashboard headline. They require very different work.
What you’ll learn
- How to define a realistic repeat window by product and customer.
- How to separate timing, product fit, experience, habit, and customer-mix problems.
- How to build a simple cohort comparison that guides action.
- Why more messaging can make a poorly diagnosed retention problem worse.
Start with the buying cycle, not a generic benchmark.
A coffee subscription, skincare treatment, suitcase, and piece of furniture should not share a repeat-purchase expectation. Even within one brand, a trial size and a durable accessory have different jobs.
For each meaningful first product, estimate the earliest sensible replenishment or next-product window. Use product consumption, customer interviews, order history, and category behavior. Treat an estimate as an estimate.
Then view cohorts only after they have had enough time to repeat. A group acquired 30 days ago should not be called weak if the normal second purchase occurs closer to day 75.
Only compare customers who have reached the same age since first purchase. State the window: “percentage of first-time buyers who placed a second order within 90 days,” for example. A blended returning-customer share answers a different question.
Test five explanations before prescribing retention.
1. Timing: the customer is not ready yet.
Orders may be healthy at 120 days and look poor at 60. Customers can also delay replenishment because the product lasts longer than expected. Inspect the distribution of days to second order, not only a single cutoff.
If timing is the issue, an earlier discount may waste margin and train the customer to wait. Better product-use guidance and a message near the natural replenishment moment may be enough.
2. Product fit: the first item does not lead anywhere.
Some acquisition products create an easy first sale and a weak second step. Customers may like the item but have no obvious reason to explore the wider range. Compare second-order behavior by first SKU, bundle, size, and use case.
If one starter product produces many first orders and few valuable follow-ons, the problem sits in merchandising and proposition as much as CRM.
3. Experience: the promise did not survive delivery.
Returns, refunds, support contacts, delivery delays, damaged products, sizing issues, confusing use, and low review sentiment can all interrupt repeat. Join those signals to the customer cohort where possible.
A win-back flow cannot repair an unresolved service problem. Fix the experience, recover the customer honestly, and measure whether the next eligible cohort behaves differently.
4. Habit: the product worked, but the behavior did not stick.
Many products require routine. The customer needs to remember when, how, or why to use them. Look for signs that successful repeat customers adopt a particular usage pattern, consume related content, or buy a complementary product.
The intervention may be onboarding, a product ritual, packaging guidance, or timely education. More promotional frequency is only one option, and often not the first.
5. Customer mix: acquisition changed who entered.
A large promotion, new channel, creator partnership, or market expansion can bring a cohort with different intent. Compare repeat by acquisition period, source where trustworthy, discount use, first product, geography, and first-order value.
A weak aggregate repeat rate may hide healthy core cohorts diluted by a new group. The right response could be an acquisition change, a different onboarding path, or a revised payback expectation for that segment.
Build a cohort table that can change a decision.
You do not need a complex model to begin. Build rows for first-purchase cohorts and columns for the factors most likely to explain behavior:
- first purchase month or week;
- first product or product family;
- discount band;
- market or channel;
- eligible customers;
- second order within the chosen window;
- median or distribution of days to second order;
- second-order product and contribution where available;
- return, refund, or service-contact signal.
Start broad, then segment only where the difference is large enough to matter and the sample is usable. Tiny slices invite confident stories from weak evidence.
A worked example: the welcome flow was not the constraint.
Suppose a supplements brand sees its 90-day second-order rate fall after a strong acquisition month. The lifecycle team proposes a new discount series.
The cohort view shows that customers who started with the core 30-day product repeat much like earlier groups. The decline is concentrated among buyers of a heavily promoted trial bundle. Those customers take longer to begin the product, contact support more often about usage, and rarely move into the full-size range.
The stronger first actions are to improve trial onboarding, clarify the path to the full-size product, and review whether the acquisition creative set the right expectation. A blanket discount would reach healthy and unhealthy segments alike while teaching the team little.
This is an illustrative pattern. The real brand would need to check fulfilment, stock, cohort age, product changes, and message exposure before accepting the explanation.
If every customer receives the new flow, offer, and onboarding at once, improvement may be visible but the reason will remain uncertain.
Match the intervention to the cause.
- Timing: adjust the message window and replenishment cue; avoid premature discounting.
- Product fit: redesign the first-to-second product path, bundle, or recommendation.
- Experience: fix fulfilment, product-use, service, or expectation gaps before increasing pressure.
- Habit: support the routine with onboarding, education, reminders, and visible progress.
- Customer mix: change acquisition qualification, segment the journey, or set a cohort-specific economic expectation.
Choose one leading behavior and one commercial outcome. For a habit intervention, the leading signal might be onboarding completion or product-use engagement; the commercial outcome remains eligible second purchase or contribution. Review both.
Retention improves when the second purchase becomes the natural result of a good first experience.
Frequently asked questions
What is a good repeat-purchase rate?
It depends on product life, replenishment, first-product role, price, channel, and the measurement window. Use comparable cohorts and your product’s real buying cycle before looking outward.
Should we use discounts to drive the second order?
Only when the incentive addresses a real barrier and the economics support it. A discount can accelerate timing, but it can also subsidize customers who would have returned anyway.
Can returning-customer revenue measure retention?
It is useful context, but it mixes cohort sizes, timing, frequency, and order value. Cohort-based second-purchase behavior answers the retention question more directly.
What if the product is naturally low-repeat?
Measure the next valuable behavior: accessory purchase, referral, review, service adoption, replacement cycle, or movement into another category. Do not force replenishment logic onto a durable product.
Do not send more messages until you know which customer, product, moment, and experience are failing to create a reason to return.