Average Jewelry Return Rate: 2026 Benchmarks

If you sell jewelry online, the return rate is one of the few numbers that quietly decides whether your contribution margin holds. Jewelry returns less often than almost any other category, but each returned piece carries some of the highest AOV in retail, so the dollar damage per return is outsized. This page sets the benchmarks by category and channel, anchors them to the one clean public-company figure we can pull from SEC filings, and lays out the policy levers that actually move the number. The read is built for operators running roughly $2M to $50M in jewelry GMV. For the broader financial profile of jewelry brands at this scale, the jewelry financial benchmarks guide covers margins, AOV, and operating ratios across the sector.

What's the average jewelry return rate?

There is no single jewelry return rate, and any source that gives you one number is hiding the spread. The honest answer is a range that depends on subcategory, price point, and channel.

For fine jewelry sold online (rings, necklaces, fine gemstone pieces), the benchmark range is roughly 8% to 15%, with a working midpoint near 10% to 12%. Fashion and costume jewelry runs higher, 12% to 20%, because it is lower-priced, more appearance-driven, and bought on more of an impulse. As a sanity anchor, the NRF 2025 Retail Returns Landscape puts the all-retail online return rate at 19.3% across every category, so jewelry sits below the all-retail online average on a unit basis.

The single biggest swing factor is channel. Online return rates run 2x to 3x the in-store rate for the same merchandise, because a customer who walks into a store sizes the ring, sees the luster, and inspects the stone before paying. Remove that and the rate roughly doubles. In-store fine jewelry sits near 3% to 5%; the same pieces sold online land in that 8% to 15% band.

When I talk to founders running a jewelry brand at this size, the first thing they do is plug a return rate into the P&L as an assumption, usually 5% or 6%, and build the unit economics on top of it. That is fine as a placeholder, but the gap between an assumed 5% and an actual 12% is the difference between a healthy contribution margin and a leaking one. Measure it before you model it.

Subcategory	Online return rate	Midpoint	In-store (est.)	Notes
Fine jewelry (rings, necklaces)	8% to 15%	~10% to 12%	3% to 5%	Higher AOV, more considered purchase
Bridal / engagement	10% to 15%	~12%	2% to 4%	Sizing and gifting amplify the online rate
Fashion / costume jewelry	12% to 20%	~15%	5% to 8%	Lower AOV, appearance-driven, impulse buy
All retail (NRF 2025 baseline)	19.3% (online)	~15.8% (all, NRF/Happy Returns)	8.9%	Every category, not jewelry-specific; 19.3% online verified on NRF page

The cleanest public number belongs to Brilliant Earth (NASDAQ: BRLT), a pure-play online fine jewelry brand. From its FY2024 Form 10-K, the company processed about $25.8M of returns against $422M of net sales, an implied return rate near 6.1%. That rate has risen in each of the last two reported fiscal years as net sales softened, which is its own warning: a rising return rate on falling sales squeezes margin from both ends.

Treat the BRLT 6% as the floor for a disciplined, premium, online-first brand with a strict 30-day window, not as the number a fashion-jewelry operator should expect. Your mix, price point, and policy decide where in the band you land.

What actually drives jewelry returns

Three forces explain most jewelry returns, and only one of them is fully in your control.

The first is sizing, and it is the single largest driver for rings. Industry estimates put ring-sizing returns at roughly 15% of all ring returns. A buyer ordering for a gift, or for themselves without trying on, is guessing, and a guess on ring size is wrong often enough to matter. This is the lever you can actually pull: complimentary resizing, a clear sizing guide, and at-home or AR try-on tools all chip at it directly.

The second is gifting, and it is structural. The gift-giver cannot know the recipient's exact taste or size, so a meaningful share of gifted jewelry comes back regardless of how good your product page is. That pressure concentrates around major gifting moments, which is why the calendar matters so much (more on that below).

The third is expectation mismatch. Photos cannot fully convey scale, luster, or subtle color variation in a fine gemstone, so some buyers open the box and feel the piece is smaller or different than they pictured. Better product content, video, and AR try-on narrow the gap; one benchmark found AR try-on lifted add-to-cart by roughly a third while cutting returns.

The pattern we see again and again: operators benchmark their online return rate against an in-store number, their own retail floor or a peer's, and the gap is eye-opening. That gap is almost entirely sizing and expectation mismatch, which means it is partly addressable, not fixed.

The January spike: holiday gifting compresses your return calendar

Jewelry returns are not evenly spread across the year. They are heavily back-loaded into January, and for a gift-heavy category that concentration is brutal on cash flow and labor planning.

Narvar's January 2026 data found post-holiday returns climbing nearly twice as fast as holiday sales, and earlier benchmark work put roughly a third of all holiday returns landing in January. For jewelry the effect is amplified: engagement rings and bridal pieces given over the holidays, gift recipients returning the wrong size, and extended holiday return windows all expiring at once in Q1.

The operator implication is concrete. If you staff your CX and returns team to a monthly average, January will swamp you. If you forecast cash to a smooth curve, the January refund wave will surprise your bank balance. Build the spike into both. One founder I worked with treated returns as a flat monthly line in their model and got caught flat-footed in January two years running before they finally indexed the forecast to the seasonality.

The true cost of a jewelry return

A jewelry return is a cash event, not a rounding error, and the high AOV is exactly what makes it bite.

All-in, a return costs roughly $10 to $65 per item depending on your size and logistics setup: reverse shipping, inspection labor, payment-processing fees you do not get back, and restocking. On top of that is markdown risk, because only about 48% of returned jewelry resells at full price. The rest takes a price cut or sits, which is additional COGS exposure that never shows up in the headline return rate.

AOV	Return rate	Returns per 1,000 orders	Cost per return (low)	Cost per return (high)	Cost per 1,000 orders
$200	8%	80	$10	$65	$800 to $5,200
$350	10%	100	$10	$65	$1,000 to $6,500
$500	12%	120	$10	$65	$1,200 to $7,800

Run the math on a typical case. At a $300 AOV and a 10% return rate with $65 all-in cost, you are eroding about $6.50 of margin for every order you ship, before markdown loss. On fine jewelry gross margins of 40% to 70%, that is often the single largest variable deduction from contribution margin after customer acquisition cost. It is the kind of line that looks small per order and turns into a six-figure number across a year. If you want help wiring returns into the rest of your unit economics, that is exactly the work we do in our fractional CFO services for ecommerce brands.

Jewelry returns less often than almost anything else in retail, but the dollar damage per return is among the highest. A brand that treats its return rate as a 5% assumption and never measures it can be quietly giving back a chunk of contribution margin on every order. The number is worth measuring, not guessing.

Policy levers that actually hold the rate down

You cannot eliminate jewelry returns, and trying to can cost you more in conversion than it saves. The goal is structure, not suppression. Four levers do most of the work.

Window structure comes first. The default that works is a 30-day refund window paired with a longer exchange or resizing window, which is exactly what Brilliant Earth runs (30-day return plus 60-day complimentary resizing) and what the Signet brands (Kay, Zales, Jared) run with their 30-day return and 60-day exchange structure. A longer refund window lifts purchase rate, but a 30-day refund plus 60-day exchange captures most of that lift without leaving you exposed to late returns.

Retailer	Return window	Exchange window	Resizing	Final-sale exclusions
Brilliant Earth (BRLT)	30 days	30 days	60 days, 1 complimentary	Custom orders, estate
Kay Jewelers (Signet)	30 days	60 days (in-store)	Per brand terms	Custom, engraved, altered
Zales (Signet)	30 days	60 days (in-store)	Per brand terms	Custom, engraved, altered
Jared (Signet)	30 days	30 days (in-store)	Per brand terms	Custom, engraved, altered

Final-sale carve-outs come second. Engraved, personalized, and resized pieces should be final sale; this is industry consensus and customers accept it. Third is exchange-first flow design: 30% to 50% of would-be refunds can convert to an exchange or store credit when the return flow nudges that way, and that is the single largest recoverable leakage point for most jewelry operators.

The fourth lever is authentication, and it is the one most DTC operators underweight. Lab-grown diamond swap fraud is a real and documented risk: a natural stone is removed during appraisal or weighing and replaced with a lab-grown stone of matching carat weight at a fraction of the value. A 2024 New York case put roughly $460,000 of swapped stones in front of a grand jury, illustrating the mechanism. The same substitution risk applies when accepting customer returns above a value threshold; without authentication, there is no practical check. Require a certificate-number match against the original order, verify the laser inscription, and inspect condition before issuing a refund. It is cheap insurance against an expensive loss.

One honest caution on the suppression instinct. A founder I worked with pulled free returns to get their rate down, and it worked on paper, but their women's segment softened at the same time and they could never cleanly isolate whether the policy change was the cause. The safety net is part of why people buy expensive things online. Tighten structure, not the willingness to stand behind the product.

Related reading. For the true landed cost of a return, see our true cost of apparel returns breakdown and, for the gifting-season spike, our holiday return rate spike benchmarks.

Sources and methodology

Brilliant Earth (BRLT) FY2024 Form 10-K provides the cleanest public benchmark. As a pure-play online fine jewelry brand, its audited filing discloses net sales of $422M and a sales-returns reserve roll-forward showing about $25.8M of returns and allowances processed, an implied return rate near 6.1% (calculated as returns processed divided by net sales; the "returns and allowances" line includes minor price adjustments in addition to physical returns, which may fractionally inflate the implied rate relative to a pure unit-return measure). The FY2022 figure is estimated from the opening balance of the FY2023 roll-forward. Filing: BRLT FY2024 10-K on SEC EDGAR.

The NRF 2025 Retail Returns Landscape anchors the all-retail baseline. The NRF page confirms a 19.3% online return rate for 2025 across all retail categories. The 15.8% overall figure cited in the table above comes from the NRF/Happy Returns 2025 partnership report rather than the NRF landing page directly; treat it as an approximation. Neither source breaks out jewelry as a standalone subcategory, so both figures are used here only as benchmark anchors, not as jewelry-specific numbers. Report: NRF 2025 Retail Returns Landscape.

Platform and aggregator benchmarks set the category ranges. The 8% to 15% fine-jewelry and 12% to 20% fashion-jewelry ranges are synthesized from multiple retail-returns benchmark reports, each using a different merchant sample. The Loop Returns Winter 2024 Benchmark Report (22 million returns, 4,000 merchants) puts jewelry at 21.6%, above these ranges and best read as the omnichannel high end for a broader merchant mix rather than corroboration of the DTC-brand midpoints. No single vendor publishes a standalone jewelry-category rate broken out by fine vs. fashion, which is why a range is the honest presentation.

Narvar data sets the post-holiday seasonality. Narvar found January 2026 returns climbing nearly twice as fast as holiday sales, with earlier benchmark work placing roughly a third of holiday returns in January. The monthly index in the seasonality chart is directional, not a set of exact statistics. Source: Narvar, when does return season really start.

Return-fraud and cost figures come from dated industry reporting. The lab-grown diamond swap case is documented in National Jeweler's July 2024 report on a New York dealer charged over roughly $460,000 of swapped stones. Per-return cost ($10 to $65) and the ~48% full-price resale figure come from published reverse-logistics estimates.

Limitations. No vendor publishes a jewelry-only return rate as a standalone audited number, so the category ranges are synthesized across samples with different methodologies. The BRLT 6% reflects one premium, online-first profile and should be read as a floor, not a market average. In-store rates are sourced estimates derived from the well-documented finding that online returns run roughly double in-store. The seasonality index is directional only.

Frequently asked questions

what's the average return rate for jewelry sold online?

Plan for 8% to 15% for fine jewelry, with a working midpoint of 10% to 12%. Fashion and costume jewelry runs higher, 12% to 20%, because it is lower-priced and more impulse-driven. The cleanest public benchmark, Brilliant Earth, runs around 6% with a strict 30-day window and high-consideration buyers.

how does jewelry compare to apparel for return rates?

Jewelry returns less often but each return hurts more. Apparel sits in the high 20s on platform benchmarks; jewelry is lower on a unit basis but carries far higher AOV, so the dollar impact per return is among the highest in retail.

why do jewelry returns spike in january?

Holiday gifting. Engagement rings, bridal, and gift sets bought in December get returned or exchanged in January when the recipient knows the size and taste. Narvar found January returns climb nearly twice as fast as holiday sales, and roughly a third of holiday returns land that month.

what return window should i offer as a jewelry brand, 30 or 60 days?

Most major retailers cluster at 30 days for refunds with a longer exchange or resizing window, and that is the right default. A longer window lifts purchase rate, but a 30-day refund plus a 60-day exchange or resize captures most of that lift without leaving you exposed to late returns and fraud.

how do i reduce ring returns without hurting conversion?

Attack sizing first, since it is the single largest ring return driver. Free or complimentary resizing, a clear sizing guide, and at-home or AR try-on tools cut returns without removing the safety net that makes people buy. Pulling free returns outright usually drops conversion more than it saves.

how much does a jewelry return actually cost my business?

Roughly $10 to $65 all-in per item for reverse logistics, labor, and payment processing, plus markdown risk because only about 48% of returned jewelry resells at full price. On a $300 order at a 10% return rate, that erodes a meaningful slice of contribution margin on every order you ship.

should a jewelry brand offer store credit only or full cash refunds?

Offer cash refunds within the window but design an exchange-first flow, because 30% to 50% of would-be refunds can convert to an exchange or store credit when the flow nudges that way. Final-sale carve-outs for engraved, personalized, and resized items are standard and reasonable.

how do lab-grown diamond swaps work and how do i protect against them at returns?

The fraud substitutes a lab-grown stone of matching carat weight, at a fraction of a natural stone's value, in place of the original. The documented mechanism, illustrated by a 2024 NYC case involving $460,000 of stones, is a swap during appraisal or weighing. That same substitution risk applies when accepting high-value customer returns. Protect against it by matching the certificate number, verifying the laser inscription, and inspecting the stone against the original order record before refunding above a value threshold.