Retention and repurchase benchmarks by category

Set one 12-month retention target across a house that sells both mascara and mechanical watches, and one team will clear it without trying while the other misses it by construction. A Sephora Beauty Insider who bought a serum in March is expected back before the bottle runs out. A Breguet client may not transact again in the same decade and is still, on any sensible reading, retained. The number is rarely the problem. The category attached to the number is.

What follows is the comparative table, the arithmetic that makes categories comparable, and the specific ways a blended figure lies to you.

Why one benchmark doesn't fit all of luxury

Luxury is a set of categories with different purchase cycles, price points and reasons to return:

• Beauty and fragrance: consumable, replenishment-driven, closest to mass-retail dynamics. Estée Lauder Companies (which sells the very products under discussion here, from Clinique to La Mer to Jo Malone London) runs on the fact that a serum is finished in a season.
• Leather goods and accessories: frequent, semi-habitual, engineered around entry price points.
• Ready-to-wear: seasonal and trend-sensitive, moderate repurchase.
• Watches: mechanical, durable, serviced rather than replaced.
• Hard luxury more broadly (fine and high jewelry): often singular, milestone-driven, sometimes inherited rather than bought.

Swatch Group is the clean test case, because it holds most of that spread inside one owner: around eighteen brands running from a plastic Swatch at roughly €100 to Breguet and Blancpain at five and six figures, plus Harry Winston in high jewelry. A Swatch collector may buy four times a year. A Breguet client may buy twice in a lifetime. Averaging them produces a number that describes no one.

Defining the core metric: repurchase rate

Repurchase rate = (customers in a cohort who buy again within a defined window) / (all customers in that cohort) x 100.

Two things break this in practice. The first is the window: 24 months is reasonable for jewelry and far too generous for fragrance, where it flatters the figure by folding in three replenishment cycles. The second is the denominator. Plenty of cited luxury retention numbers are computed only over clients who already came back at least once, which is survivorship bias with a decimal point. The famous "80% of sales come from repeat clients" style claims are share-of-revenue facts, not cohort retention facts, and the two get quoted interchangeably.

Worked example. A beauty division recruits 10,000 first-time buyers in a quarter. Within 12 months, 6,500 buy again. Repurchase rate = 65%. Now split the cohort: the December recruits, mostly gift buyers purchasing for someone else, come back at closer to 30%. Blending them into the annual figure hides the fact that your Christmas acquisition is buying strangers, not clients.

Category benchmarks (estimates, recent industry reporting)

These are directional estimates drawn from industry commentary (Bain & Company's annual Luxury Goods Worldwide Market Study is a good free primary source for category-level trend data), not audited disclosures. Maisons do not publish repurchase rates.

Why the gap exists: purchase logic, not brand health

Consumability. Beauty repurchase is functionally required. Sephora built its Beauty Insider tiers around that certainty: the top Rouge tier turns on annual spend, which only works when the product runs out on a predictable schedule.

Durability and servicing. A mechanical watch is bought to outlive the buyer. Most manufacturers recommend a full service somewhere between five and ten years, so the natural repeat contact in watches is a service appointment, not a second watch. Houses that count only transactions record a decade of silence from a client who has been in the salon twice.

Frequency ceilings at high price points. A client with a five-figure annual jewelry wallet usually deploys it in one purchase, not five.

Resale leakage. In hard luxury, a client who sells a piece and buys a different reference on the secondary market has repurchased the category and churned from your books. Beauty has no equivalent; watches and jewelry have a large one.

What to measure instead of pure repurchase rate

For low-frequency categories, repurchase rate alone understates loyalty. Useful companions:

• Share of wallet: what share of a client's total category spend comes to you. A jewelry client who buys once every four years but always from the same maison is a 100% share-of-wallet client with a 25% repurchase rate.
• Referral and gifting rate: in watches and jewelry, a satisfied client often produces the next sale through a spouse or a child rather than buying again.
• Second-purchase conditional rate: the jump from first to second purchase is the hard one. Once a client buys twice, the probability of a third is materially higher in every category. Report the two separately, because a house with a healthy third-purchase rate and a weak second-purchase rate has an onboarding problem, not a loyalty problem.
• Lifetime valueLifetime valueLifetime Value: the total revenue (or profit) a customer generates throughout their entire relationship with your business.View full definition →, computed the way the sibling lesson on clients who buy twice a decade sets it out, reorders the table completely: the categories at the bottom of the repurchase column are frequently the ones with the highest value per client, which is why budget allocated on frequency alone drifts to the wrong place.

The second-order consequence of a badly chosen window is a misallocated budget. Measure watches on 12 months, read a retention rate near zero, and the honest conclusion looks like "retention marketing does not work in watches", so money moves to acquisition, where the payback threshold this module's acquisition lesson sets is hardest to clear. The same error in reverse makes a beauty brand with a 68% repurchase rate look excellent when the category median sits close to it.

Setting realistic retention targets by category

1. Segment the CRM (customer relationship management) database by category before setting any KPI (key performance indicator). The arithmetic of blending: 100,000 beauty clients at 65% and 2,000 watch clients at 18% gives a house-level 64%. That number moves only when beauty moves, so the watch business becomes invisible to management reporting.
2. Match the window to the natural cycle. Six to 12 months for beauty, 12 to 24 for accessories, 36 to 60 for watches and high jewelry. Publish the window next to the figure every time.
3. Exclude gift-driven cohorts, or report them separately. Anything acquired in the two weeks before Christmas or Mother's Day is a different animal.
4. Set the target against the category, then against your own prior cohorts. Your Q1 2024 cohort at month 12 versus your Q1 2023 cohort at month 12 is a cleaner read than any published benchmark, because the methodology is identical by construction.
5. For long-cycle categories, read the leading indicators the engagement lesson scores rather than waiting four years for a transaction to confirm a campaign.
6. Weight retention spend by margin and value per client, not by frequency. A jewelry client returning every fourth year can justify more spend per head than an annual fragrance repeat.

🎬 [VIDEO: "Bain Luxury Report Explained" - youtube.com - search for recent Bain & Company luxury goods market briefings, which break down category-level growth and client behavior trends discussed by Bain partners]

A note on data limitations

Treat any repurchase statistic quoted without a window and a denominator as unusable. Most maisons hold client data as proprietary; the group-level disclosures that do exist (Swatch Group and Estée Lauder both report by brand or segment rather than by client cohort) tell you about revenue mix, not retention. Industry estimates from Bain, McKinsey's State of Fashion series and Altagamma are directional and worth using, provided you check the methodology and the date before a figure reaches a client-facing deck.

Key Takeaways

• Repurchase benchmarks vary by category: an estimated 70 to 85% for leather goods against 20 to 30% for fine jewelry and 15 to 25% for mechanical watches. There is no luxury-wide target.
• Always publish the measurement window and the denominator alongside the rate. Survivorship-based figures and share-of-revenue claims get quoted as retention and are not.
• Blended house-level rates are dominated by the highest-volume category, which is how a watch or jewelry business disappears from the dashboard.
• Low repurchase does not mean low value. Share of wallet, referral rate and the second-purchase conditional rate carry more signal in long-cycle categories.
• Service appointments and resale activity are real retention events that transaction-only reporting misses entirely.
• Your own prior cohorts, measured the same way, beat any published benchmark for setting a target.