Customer lifetime value: why keeping her beats chasing the next one

“Retention” is one of those words that sounds like homework. It isn’t. It’s just the difference between a brand that has to win every customer from scratch every month and one that gets to keep the ones it already earned. The maths behind it is genuinely on your side, but it’s buried under a lot of vague “loyalty matters” talk, so I went and traced the real numbers, kept what has a name and a year against it, and pulled it into something you can act on. Here’s why keeping her is almost always cheaper than chasing the next one, and how to decide where your next pound goes. Founder to founder.

The short version

1. CLV is her worth over time, not per order. Roughly what she spends per order, times how often she buys, times how long she stays, after your margin. The first sale is rarely where you make the money.
2. Keeping her is cheaper than finding her. Acquiring a new customer runs 5 to 25 times the cost of keeping one you already have (Harvard Business Review, 2014).
3. Small retention gains compound hard. Lifting retention by 5% can raise profit by 25% to 95% (Frederick Reichheld, Bain & Company; cited in Harvard Business Review, 2014). It’s an old finding, but it’s the honest one.
4. Acquisition got dearer, so the customer you own got more valuable. CPMs rose and ad signal thinned after the iOS privacy changes, so the cheapest growth most months is the customer already on your list.
5. Email and SMS own lifetime value, paid owns the first sale. Owned channels do the repeat work at near-zero marginal cost. Don’t pay twice for a customer you already have.
6. Judge growth on blended economics, not last-click ROAS. The flattering retargeting number is mostly revenue your own flows would have closed for free.

---

1. What CLV actually is (and why the first order rarely pays you back)

Plain version. Customer lifetime value is what a customer is worth to you across the whole relationship, not the single order in front of you. A rough back-of-envelope version: average order value, times how often she buys in a year, times how many years she stays, then taken down to your gross margin so it’s real profit and not vanity revenue. You don’t need a data team for it. You need to stop judging a customer by her first basket.

Here’s why the first basket lies to you. By the time you’ve paid for the ad, the discount that got her over the line, the shipping and the cost of the product, the first order on a small beauty brand often barely breaks even, sometimes worse. The money lives in the second, third and fourth order, where there’s no acquisition cost to pay again. That’s the whole game. The first sale buys you the right to a relationship, and the relationship is where the profit is.

And the relationship is genuinely easier to sell into. The probability of selling to an existing customer sits around 60 to 70%, against 5 to 20% for a fresh prospect (Marketing Metrics, Farris et al., 2010). She already trusts you, she’s already tried the product, she doesn’t need converting from cold. That’s why the economics tilt so hard: acquiring a new customer costs anywhere from 5 to 25 times more than retaining one you have (Harvard Business Review, 2014).

Then it compounds. The most-quoted figure in retention, and it’s old so I’ll flag it, is that lifting your retention rate by just 5% can increase profit by 25% to 95% (Frederick Reichheld, Bain & Company; cited in Harvard Business Review, 2014). I’d treat the top of that range as a ceiling, not a promise, because the original work was finance-sector and varies a lot by industry. But the direction is rock solid, and it’s the reason every serious operator eventually stops obsessing over the top of the funnel. A small improvement in how many customers come back moves profit more than almost anything you can do to acquisition.

2. Acquisition got more expensive, so the customer you already have got cheaper

The reason this matters more now than it did five years ago is that buying a stranger’s attention has quietly got dearer. By the agency Rareview’s read, median direct-to-consumer customer acquisition cost has climbed roughly 60% over five years (Rareview, 2025), driven by rising CPMs and more brands bidding for the same feed. On top of the price, the targeting got blunter. Apple’s privacy changes zeroed the ad identifier for the roughly three-quarters of users who don’t opt in to tracking (ATTN Agency, 2026), so the platforms see less of who actually bought, and the old “spin up a store, buy Meta ads, scale” loop no longer carries the unit economics on its own.

Put those together and the conclusion is almost arithmetic. If the cost of a new customer keeps rising and the first order barely pays it back, then the cheapest growth available to you most months isn’t a new customer at all. It’s the next order from a customer you’ve already paid to acquire. That order has no ad cost, needs no fresh discount, and has a far higher chance of closing.

The problem is that most small brands leak exactly there. The average ecommerce repeat-purchase rate sits around 28% (Mage Loyalty, 2026), which means roughly seven in ten first-time buyers never come back. For consumable beauty and wellness, where she will physically run out of the product, that’s money left on the floor. The fix isn’t clever. It’s having the basic owned-channel machine actually built: a welcome flow, a post-purchase flow, a replenishment nudge timed to when she runs low. Most leaks are simply a flow that was never turned on.

3. Email and SMS own lifetime value, paid owns the first sale

This is the line we hold at Mad Social, and it’s the most useful way to split the work. Paid social owns customer acquisition cost and acquisition quality: it buys the first customer and decides who enters the machine. Email and SMS own lifetime value: they’re the repeat engine, and they run at near-zero marginal cost once built. In Klaviyo, automated flows generate close to 41% of total email revenue off just 5.3% of sends (Klaviyo 2025 Benchmark Report), and email overall is credited with around a quarter of ecommerce revenue on average (Klaviyo ecommerce benchmarks; treat as directional). That’s not a channel you bolt on at the end. It’s where lifetime value actually gets captured.

Two brands worth pointing at, both doing the achievable version of this rather than the fancy one. Rheal Superfoods runs a clean subscribe-and-save at 20% on its consumables, with the customer choosing the frequency and skip, swap or cancel kept free and self-serve. That’s exactly the model for something she re-buys anyway. You remove the chore, the discount feels like a reward for a habit, and the relationship renews itself. A step up, Wild Nutrition runs monthly supplement subscriptions with flexible skip and modify and even subscriber nutritionist consults, which turns replenishment into a service rather than a charge that surprises her. Both keep leaving easy on purpose, which is the point. Retention that survives scrutiny is about being worth coming back to, never about trapping her into staying.

Now the bit that quietly defunds retention if you let it. The most flattering number in your ad account, the 6 to 10x return on retargeting your own site visitors and past buyers, is mostly an illusion. On a proper holdout test it collapses to something closer to 1.5 to 2.5x, and roughly 30 to 40% of those “attributed” purchases are revenue your owned channels would have closed for free anyway (Stella, 2025; WorkMagic, 2025). Treat those exact numbers as directional, they’re modelled incrementality estimates rather than gospel, but the pattern is well evidenced: in-platform ROAS tends to read perhaps 2 to 3x above the true incremental number, and it reads highest exactly on the warm, retargeting spend brands lean on hardest. So you can end up paying Meta to re-sell a customer your welcome flow already had. That’s the definition of paying twice.

4. How to weigh CLV when the next pound is in your hand

So where does the next pound go? Here’s the test I’d actually use, founder to founder.

Ask what the pound buys. If it acquires a genuinely new customer, that’s paid’s job, and you judge it on blended economics, not the last-click number on the dashboard. Look at blended marketing efficiency across the whole account, contribution margin after product and shipping, and the cost of a new customer specifically, rather than the warm-traffic ROAS that’s mostly your own audience reflected back at you. Creative does most of the heavy lifting here anyway. By NCSolutions’ analysis, nearly half of incremental sales (around 49%) trace to the creative itself, more than targeting, reach and recency combined (NCSolutions, 2023), so a new pound on acquisition is usually better spent on a sharper founder-led ad than on narrower targeting.

If instead the pound just re-reaches a customer who already opens your email, stop. That’s the work your flows do for free. Spend it on the owned machine instead: build the flow that isn’t built yet, the replenishment reminder timed to run-out, the win-back pegged to her real repurchase cycle. A lightweight loyalty layer earns its place here too, the kind Thunderbird Skin runs with its Good Skin Rewards points and a simple referral, because members who actually redeem tend to spend meaningfully more than those who don’t (Rivo, 2025, vendor-reported, so treat it as directional). And there’s a real concentration worth protecting. In most stores the top tenth of customers drives something like 40 to 50% of revenue (a long-standing Pareto pattern, directional), so an hour spent keeping them close usually beats an hour spent at the cold top of the funnel.

The north star, if you want one, is Beauty Pie, which built retention right into the business model: a membership club where the subscription, not product margin, is the revenue, so the whole thing is engineered around keeping her rather than re-selling to her. You don’t need the buying-club infrastructure. You need the principle. Design for the second year, not just the first order.

What this means for your brand

1. Stop judging customers by their first basket. Track what she’s worth over a year, net of margin. The first order buys the relationship; the relationship is the profit.
2. Build the boring flows first. Welcome, post-purchase, replenishment, win-back. They capture lifetime value at near-zero marginal cost, and most leaks are simply a flow that was never switched on.
3. Don’t pay twice for the same customer. If paid is just re-reaching people who already open your email, that spend belongs in the owned machine. Retargeting your own list is mostly non-incremental.
4. Judge paid on blended economics. Blended efficiency, contribution margin and new-customer cost, never the warm-traffic ROAS the platform shows you.
5. Make staying worth it, never make leaving hard. Subscribe-and-save, flexible skip and cancel, a loyalty layer she actually redeems. Authentic retention survives scrutiny; lock-in doesn’t.
6. Protect your best customers like the asset they are. A small lift in repeat rate among the people who already love you compounds harder than almost anything you can do to acquisition.

A note on the evidence

A couple of honesty flags, because you might quote this. The famous “5% retention lifts profit 25% to 95%” line is real and traces to Bain, but it’s decades old and the top of that range came from specific industries, so use it as direction, not a guarantee. The retention-versus-acquisition cost multiples (5 to 25 times) are widely cited through Harvard Business Review and are best read as “much cheaper,” not a precise law for your store. The email share of ecommerce revenue (around a quarter) is a Klaviyo benchmark that moves with store size and season, so treat it as directional too. The retargeting holdout numbers (the collapse to 1.5 to 2.5x, the 30 to 40% that’s non-incremental) are modelled incrementality estimates, so they show you the direction, not a guaranteed figure for your account. And anything I’ve marked vendor-reported or directional, like the loyalty spend multiple and the top-10% concentration, points the right way but shouldn’t go in front of an investor as hard fact. Everything else here has a name and a year against it.

That’s the case for keeping her. The rest of the library is about how.

Madison, Mad Social

---

Sources

• Harvard Business Review, The Value of Keeping the Right Customers (Amy Gallo), 2014 (acquisition vs retention cost, 5 to 25 times; and the Bain 5% retention lift)
• Frederick Reichheld, Bain & Company, customer-retention and profit research; cited in Harvard Business Review, 2014 (5% retention lift, 25% to 95% profit; classic finding, treat as directional)
• Marketing Metrics, Farris, Bendle, Pfeifer and Reibstein, 2010 (probability of selling to an existing vs new customer)
• Rareview, The DTC Reckoning, 2025 (median DTC CAC up roughly 60% over five years; agency estimate)
• ATTN Agency, iOS privacy and Meta ads optimisation, 2026 (ATT opt-in around 25%, the resulting signal loss)
• Mage Loyalty, beauty and skincare repeat-purchase benchmarks, 2026 (around 28% average ecommerce repeat rate)
• Klaviyo, 2025 Benchmark Report (AMER), 2025 (flows around 41% of email revenue off around 5.3% of sends)
• Klaviyo, ecommerce email benchmarks (around a quarter of store revenue attributed to email on average; directional, varies by store size and season)
• Stella and WorkMagic, incrementality / iROAS benchmarks, 2025 (retargeting holdout ROAS; non-incremental share of attributed purchases; modelled, directional)
• NCSolutions, Five Keys to Advertising Effectiveness, 2023 (creative around 49% of incremental sales; modelled meta-study)
• Rivo, loyalty-program statistics, 2025 (redeemer spend uplift; vendor-reported, directional)
• Brand examples: Rheal Superfoods, Wild Nutrition, Thunderbird Skin, Beauty Pie (Mad Social verified brand-example bank, June 2026)

Prepared by Mad Social. Figures are point-in-time (2010 to 2026) and attributed to the named source. Directional and vendor-sourced figures are labelled as such in the text.