Customer retention strategies: fix the leak before the campaign

What customer retention strategy means in e-commerce

A customer retention strategy is the set of deliberate actions a brand takes to make existing customers buy again. The standard playbook is loyalty programs, email and SMS lifecycle marketing, subscriptions, replenishment reminders and win-back campaigns, measured through repeat purchase rate, retention cohorts and lifetime value.

My rule from running these brands: the loyalty battle is won or lost in the 3 to 7 days after checkout. That window is the delivery experience, and no points program reaches into it.

All of that is real and worth doing. My argument is that it is the second half of retention, and most brands are running it on top of a leaking first half.

Every order is a promise. Keeyu keeps it. Too much of the energy in e-commerce goes into how to catch customers, and not enough into how to keep them. Most brands are competing on acquisition. The smart ones are competing on retention.

The retention event nobody puts in the strategy deck

The most reliable predictor of whether someone buys from you again is whether the last order arrived as promised. Not the loyalty tier, not the win-back discount. Whether the thing they paid for turned up when you said it would.

At Papinelle I found complaint tickets running at 110% of order volume, and essentially all of them were about orders. Every one of those was a retention event happening in the support queue while the marketing team ran reactivation campaigns at the same people.

The shopper who got a bad delivery rarely sends the angry email. They paid once, the parcel went wrong, and they went quiet. Protect the shopper experience after the buy button and they don't churn, so there is nobody to buy back with a discount.

Cost of a broken promise versus cost of a campaign

Work the arithmetic on your own numbers. Take your monthly order volume, the share of orders that hit an operational break, and the share of those customers who do not return. Compare the lost lifetime value with what you spend on retention marketing in the same period.

Here is the mechanism most retention decks miss. For years cheap acquisition hid the damage: if a customer churned after a bad delivery, you bought a replacement and called it just e-commerce. That world ended. Now it is a choice to pay to acquire a customer, lose them to a late parcel, then pay again to replace them. More in the post-purchase operations playbook.

So when you look at a brand's retention numbers, the better question is not how to handle more tickets or run another win-back flow. It is why that ticket exists at all. A win-back campaign pays for the churn every time it runs. Fixing the cause means you stop paying for the same customer twice.

What retention looks like when operations hold

EHP Labs chose Keeyu to centralize orders, inventory, fulfillment and customer communication. What that did for retention is in their own words in the EHP Labs case study: "On the occasions where something comes up, we notify customers proactively. The feedback has been incredible - they actually love us more for it."

We built Keeyu because we lived the other side of this: growing fast, dealing with endless complaints, always playing catch-up. Once we flipped from reactive to proactive, everything changed. Fewer issues, happier customers, less wasted time. None of that shows up as a line in a retention report, and it is still retention work.

Customer retention examples: where the operation held and where it broke

A customer retention example is a specific case, described in enough detail to copy, of what a brand did that kept or lost an existing customer. Most published examples are campaigns: a loyalty tier, a win-back email, a subscription offer, a VIP early-access list.

Start there. Shopify's customer retention guide documents live ones, such as skincare brand Jaxon Lane giving repeat buyers free shipping, surprise gifts and handwritten notes alongside a points program. The four campaign examples most stores run, with the condition each depends on:

Read from the operations side, every one of those lists assumes the last order arrived as promised. The industry built a thousand tools to win customers. Almost none to keep them. In Sifted's 2025 survey of 500 US consumers (a vendor survey, so treat it as directional), 76% of shoppers said a positive delivery experience influenced their decision to repurchase. Each delayed or missing order is a small crack in loyalty, and the campaign examples never show the crack. By the time a ticket reaches support, the order behind it has already gone wrong, so none of the next four examples starts in the helpdesk: two where the operation held, two where it broke, each with a rule. Helpdesk manages complaints. Keeyu prevents them.

Held: a late order, and a make-good sized to a five-year customer

A recovery offer is the make-good a brand gives when an order misses its promise. Size it by the shopper's history: Keeyu can pull a loyal customer's record from Shopify and offer compensation that fits it, rather than one flat code for everyone.

Here is the example I walk brands through. An order gets rerouted to another warehouse and will land a day late. Keeyu detects the reroute against the delivery promise, and the shopper is told before they ask. The remedy is the brand's call, set as a rule in the workflow, and the shopper's order history in Shopify is what sizes it. For someone who has bought every week for five years, that can be a $10 coupon loaded to their next purchase. The delay still happened. The operation held because the shopper heard it from the brand first, with the remedy attached, and the remedy points at the next order instead of refunding the last one.

Held: VIP early access run by the service team

VIP early access means giving your best customers first claim on a launch. It reads like a marketing example. It runs on capacity, because someone has to have the hours to do it.

At P.E. Nation we did a drop every week and sold out every week. We turned the customer service team into advocates. They used their spare time to reach out to VIP customers, load their carts, and say: "we're launching at 7:00 a.m. tomorrow morning with your carts preloaded." Tracy's team could do that because they had the tools to detect issues in real time. Engagement went through the roof once they stopped firefighting, and the service team became a profit center instead of a cost center, because those hours went into exactly that outreach. That only works when the team is not buried in order-status tickets. Keeyu automates the manual, repetitive order work so a team has those hours. The rule: before you plan a VIP program, count the hours your service team spends chasing order status. Those hours are the budget for it.

Broke: the loyal buyer who went quiet

A lapsed customer is one whose buying pattern stops. Win-back campaigns usually find them by days since last order.

Jason Greenwood, host of The Ecommerce Edge, gave me this one when I was a guest on his show. For about three years he spent hundreds of dollars a month on bourbon with one online liquor store in New Zealand. Then he moved to Mexico and the orders stopped overnight. What he wanted from that store was a call from a person who had noticed, before any discount code. No order broke here. The detection did: a three-year regular going silent is a signal, and it only helps if something is watching for it. The loudest customers aren't the risk, because they care enough to complain. It's the silent ones who just disappear, and the real retention gains come from shifting from waiting for them to reach out to spotting disengagement before it happens. One rule follows from it: measure each repeat buyer's normal gap between orders, flag anyone who goes well past it (twice the gap is a starting point to tune, not a benchmark), and have a person look before the generic win-back flow fires.

Broke: a quarter of one brand's orders had issues

Customer lifetime value at risk is the future revenue attached to customers whose orders hit a problem. As soon as Keeyu detects an issue on an order, we attribute that customer's lifetime value to it, and a running total on the platform shows the customer lifetime value at risk. If the issue isn't resolved and the service level isn't met, that is the value that churns.

In one account I logged into, 25% of orders had issues, which put about $800,000 of customer lifetime value at risk. Replacing those customers costs more than keeping them: Harvard Business Review reports that, depending on the study and the industry, acquiring a new customer costs five to 25 times as much as retaining one. Value at risk is the exposure. To estimate what you will lose, multiply it by the share of those customers who do not come back:

Expected retention loss = customers whose order hit an operational break x share of them who do not reorder x average remaining lifetime value

With illustrative round numbers: 10,000 orders a month with 5% hitting a break is 500 orders, so call it 500 customers. At $200 of remaining lifetime value each, $100,000 is at risk. If 30% of them do not reorder, $30,000 walks out each month.

Every order is a promise. Keeyu keeps it.

How to tell whether a retention example is working

Customer retention rate and repeat customer rate show whether a retention example is working. Shopify gives the formulas: customer retention rate = (customers at end of period minus new customers acquired) / customers at start x 100, and repeat customer rate = customers who purchased more than once / unique customers x 100. The walkthrough is in how to measure Shopify customer retention, and the argument behind this section is in post-purchase operations is the new loyalty program.

Practical sequence

• Instrument the breaks first: how many orders hit an exception, and how many customers found out before you did
• Fix the highest-frequency operational failure before adding another lifecycle flow
• Communicate proactively when something does go wrong, with the remedy already in motion
• Then layer loyalty, subscription and lifecycle marketing on an experience that holds

The order matters. Lifecycle marketing compounds on a good experience and accelerates churn on a bad one, because it keeps inviting people back to something that disappointed them. We scored the retention tools brands buy, Keeyu included, on six published criteria in our ranking of customer retention tools.

How to measure customer retention

Customer retention metrics are the numbers that show whether existing customers keep buying, how often, and what they are worth over time. A store tracks them across a fixed period, such as a quarter or a year, so each reading compares like with like.

Six metrics cover most of what an ops team needs. Run all of them over the same period.

Lifetime value has its own walkthrough in how to calculate e-commerce customer lifetime value. The other five you can pull from a standard order export.

A worked example with round numbers

Say a store starts the quarter with 1,000 customers. It acquires 400 new customers during the quarter and ends with 1,100.

CRR = (1,100 - 400) / 1,000 x 100 = 70%. Of the 1,000 people it started with, 700 are still buying.

Now look at the year. The store had 5,000 unique customers, and 1,500 of them ordered more than once. Repeat customer rate = 1,500 / 5,000 x 100 = 30%. Seven in ten buyers never came back for a second order.

What good looks like

Benchmarks depend on category and model, so treat them as a rough range. As reported by Saras Analytics, general e-commerce retention runs around 25 to 30%, and subscription e-commerce around 40 to 60%. Shopify's guidance is to compare against businesses with similar customers and buying patterns, and against your own history. Your own trend line is the benchmark that matters most.

My take: the gap is operations, not loyalty

We spent 12 months analyzing 100+ Australian e-commerce brands. The average repeat rate sat at 30 to 40%. The best operators hit 70 to 80%, consistently. Those are Australian brands, not a US benchmark. The pattern still travels.

The brands with 70 to 80% repeat rates aren't running better loyalty programs. They're running better operations after checkout. So when your repeat rate is stuck, look at what happened to the last order before you look at the next campaign.

The metric I would add: the silent cohort

The standard metrics tell you how many customers left. They do not tell you why. Tickets will not tell you either, because they only sample the shoppers who still bother to complain. As reported by Shopify, Qualtrics found that 30% of customers do not tell anyone and switch brands silently.

You can measure that group with the order and exception data you already have:

• Count the customers whose order hit a broken promise in the period: late, lost, short-shipped or out of stock
• Subtract the ones who opened a support ticket about it
• What is left is the silent cohort
• Compare that group's repurchase rate with the repurchase rate of everyone else

If the silent cohort reorders well below the rest of your base, that gap is retention you are losing in operations. No win-back flow will show it to you. Track it next to CRR and repeat customer rate every period, and watch it close as you fix the breaks upstream.

Where this sits

This is proactive e-commerce operations: detect the break across store, warehouse and carrier, decide the remedy, and act before the customer feels it. Detect. Decide. Act. Keeyu gets shoppers what they want, on time, as promised, which is the cheapest retention mechanism available.

Every order is a promise. Keeyu keeps it. See how our customers run this or book a demo.

Related reading

For the experience layer, read customer experience management. For the ticket type that signals a broken promise, see WISMO. For the operational picture, read post-purchase operations.

• Customer satisfaction metrics: measuring recovery, not operations
• CX optimization: start from cause, not from feedback
• Delivery exception: what it means and how to act on it

• Harvard Business Review, 1990: the original zero-defections case for retention economics.
• NRF and Happy Returns, 2025 Retail Returns Landscape: 19.3% of online sales returned, and the loyalty cost of poor returns.
• The Breakout CEO #86, The Pivot This Founder Made After an Investor Called It Impossible - one in five US shoppers not getting orders on time, and the ticket ratios behind it.
• Marketing for SMEs, Jevon Le Roux on the E-Commerce Mistake Costing Millions - the orchestration layer, and the EHP Labs story.
• The 9-5 Exit Plan, How AI Is Making Customer Support Invisible - proactive AI and support that disappears.
• Harvard Business Review, 2014: acquiring a new customer costs five to 25 times more than retaining an existing one.
• Sifted, 2025 Consumer Survey: vendor survey of 500 US consumers; 76% said a positive delivery experience influenced their decision to repurchase.
• Shopify, Customer Retention: formulas for customer retention rate and repeat customer rate, and the Jaxon Lane example.
• The Ecommerce Edge Podcast, E637: Proactive Post-Purchase Ecommerce Ops & CX - host Jason Greenwood's story of the liquor store that lost a three-year regular.
• Saras Analytics, eCommerce Customer Retention: as reported by an aggregator, general e-commerce retention around 25 to 30% and subscription e-commerce around 40 to 60%.