The Retention Math: Why Ecommerce Customer Retention Beats Chasing New Buyers in 2026
Two brands each do $5M a year. On the traffic report they look identical: same sessions, same conversion rate, same average order value. One repeats 20% of its customers. The other repeats 40%. Over a full year, at the same cost to acquire, those two businesses sit roughly $2.5M apart on the P&L. Same top line today. Wildly different companies.
That gap is the entire argument for ecommerce customer retention, and it's the number most brands never actually run. Across our client portfolio, having managed more than $450M in Amazon revenue across 100+ brands, we've found the same thing again and again: the brands that scale profitably aren't the ones with the lowest CAC. They're the ones whose second and third orders make the first one cheap.
The retention math most brands never run
Start with the economics, because they're more lopsided than most operators assume.
- A 5% lift in retention increases profit by 25% to 95%. That's the old Bain and Reichheld figure, and it stays the anchor because it keeps holding up.
- Acquiring a new customer costs roughly 5 to 25 times more than keeping one you already have.
- Returning customers spend about 67% more per order than first-time buyers.
- Around 65% of revenue comes from existing customers, and the top 5% of customers drive close to 35% of total ecommerce revenue.
Put those together and the $5M example stops being hypothetical. A store at a 40% repeat rate generates roughly 50% more revenue than a store at 10% on identical traffic. You don't need more sessions to grow. You need more of the sessions you already paid for to come back.
Here's why this matters more in 2026 than it did three years ago. Average ecommerce CAC now sits around $68 to $84, up close to 40% in two years. When it cost $40 to buy a customer, a mediocre repeat rate was survivable. At $80, the first order is often a loss and the profit lives entirely in orders two, three, and four. Retention stopped being a loyalty-team nicety. It became the margin lever that decides whether paid acquisition is solvent at all.
The probability ladder: why the second order is the whole game
Repeat purchasing isn't linear. Each order a customer places makes the next one far more likely. We think about it as a ladder, and the rungs get easier to climb the higher you go.
| Customer's order history | Probability of the next order |
|---|---|
| After the 1st order | ~27% |
| After the 2nd order | ~45% to 54% |
| After the 3rd order | over 54% |
| By the 9th order | over 80% |
The steepest jump on that ladder is the first one. Getting a buyer from one order to two roughly doubles the odds they keep buying at all. That single transition, the second order, moves lifetime value more than anything else you can influence.
Timing decides it. Customers who place a second order within 60 days of the first are about 3 times more likely to become long-term customers than those who wait 120 days or more. The window is early and it's short. Most brands spend their budget re-acquiring these exact people through paid ads weeks after a well-timed nudge would have brought them back for free.
Retention isn't equal across categories
Now the honest part. Retention is a powerful lever for some brands and a weak one for others, and the difference is the product, not the marketing.
| Category | Typical repeat purchase rate |
|---|---|
| Consumables (supplements, food, pet) | 30% to 45% |
| Beauty and skincare | 25% to 40% |
| Apparel | 20% to 32% |
| Home and durables | under 18% |
| Subscription boxes | 60% to 70% |
| Luxury fashion | about 10% |
The average DTC store retains about 30% of its customers. Top performers with real lifecycle programs hit 45% to 55%. But if you sell mattresses or furniture, no email flow turns a five-year replacement cycle into a monthly habit. For one-time-purchase and high-AOV durable categories, the smarter play isn't chasing a repeat rate that physics won't allow. It's raising average order value on the first purchase, adding attachment and warranty revenue, and building referral. Retention advice that ignores your category will have you funding flows that can't pay back.
LTV:CAC and why 3:1 isn't always the right target
Most operators have heard that a healthy LTV:CAC ratio is 3:1. Every dollar of acquisition should return three dollars of lifetime value. It's a fine starting point and a bad finish line.
| Vertical | Workable LTV:CAC range |
|---|---|
| Supplements and health (monthly refill) | 3:1 to 6:1 |
| Skincare and beauty subscription | 3:1 to 5.5:1 |
| Wellness consumables | 2.5:1 to 5:1 |
| Apparel (mid-market) | 2:1 to 4:1 |
| High-AOV durables (furniture, appliances) | 1.5:1 to 3:1 |
Two things break the universal 3:1 rule. First, margin. DTC brands run 40% to 60% gross margins where SaaS runs 70% to 85%, so a "good" ratio depends on what each order actually contributes, not on a blanket benchmark. Second, growth stage. A 5:1 ratio can be a warning sign rather than a trophy: it often means you're under-investing and leaving growth on the table. A 2:1 with strong contribution margin and a fast payback period can be exactly right while you're scaling. Read the ratio against your margin structure, your payback period, and your stage, or it'll tell you the wrong thing with complete confidence.
This is also why LTV belongs upstream of your ad accounts, not buried in a quarterly report nobody acts on. If you know a supplements buyer is worth six months of margin, you can afford to bid past the first-order breakeven on Amazon, Google, and Meta that a competitor bidding to a single transaction can't touch. We walked through the mechanics of that in our piece on bidding Google Ads to margin instead of revenue. Retention is the input that makes aggressive, profitable bidding possible in the first place.
The retention levers we actually pull
We manage acquisition. We're not going to claim we'll run your Klaviyo flows, because that isn't our lane and you should be wary of an ads team that says it is. What a full-service ecommerce partner should do is make retention the strategic input to every channel it operates. In practice, that comes down to a few specific levers.
LTV-informed bidding across channels
We model contribution margin and repeat value by customer segment, then feed that into how we bid on paid media. The accounts we've scaled furthest are the ones where the media plan is built on lifetime value, not first-order ROAS. It's the difference between buying a customer and buying a customer you know how to keep.
Amazon Subscribe & Save, modeled honestly
Subscribe & Save is real. About 23% of US Amazon customers held an active subscription in 2024, roughly 35% have subscribed at some point, and subscribers shop Amazon weekly at 61% versus 34% for everyone else. It's one of the strongest loyalty signals on the platform.
It's also not free recurring revenue, and this is where we push back on clients. Average Subscribe & Save lifetime in CPG and supplements runs only about 2.4 to 3.8 deliveries before cancellation. A high subscription share tells you a SKU is a franchise product, but the standing discount plus short lifetimes mean you have to model it on contribution margin per delivery, not treat it as an annuity. We run that math as part of Amazon brand management, alongside brand loyalty analytics and Brand Tailored Promotions for winning specific cohorts back.
Returning-visitor conversion
Returning customers already convert at 4.5% to 6% against 1% to 2% for first-timers. The site experience for someone on their third visit shouldn't be identical to a cold visitor's, yet on most stores it is. Tightening that path is part of how we approach website and conversion optimization.
One more lever is worth naming even though it sits just outside what we run day to day, because it's the cheapest LTV win most subscription brands ignore: billing. Moving subscribers from monthly to annual billing cuts effective churn by 60% to 80% on the identical product. And involuntary churn from failed payments is 30% to 40% of total churn, sometimes half for low-AOV brands. Fixing dunning and payment retries recovers revenue you already earned and spent nothing new to get.
Loyalty programs get pitched as a cure-all, and the reported numbers look genuinely good: around 90% of programs claim positive ROI averaging roughly 4.8x, and tiered VIP structures tend to outperform flat point systems by about 80%, with top-tier members showing meaningfully higher AOV and more frequent purchases. Read those figures with one eye open. Loyalty ROI is self-reported and selection-biased, because the brands still running a program are the ones where it happened to work. Points and discounts also carry a real margin cost that rarely shows up in the case study. A loyalty program earns its place when it deepens a repeat habit you can already see in the data, not when it's bolted on to manufacture one.
Where to start
Run one number this week. Take your repeat purchase rate and ask what a 5-point improvement would add to profit at your current traffic. For most brands past $5M, that number is large enough that retention stops being next year's project and becomes the highest-return work on the table. If you want a partner that buys the traffic and has a point of view on what it's worth over a lifetime, let's talk.