Jewelry brands get pitched case studies constantly. A 38x ROI here, a 62x ROI there, an “email should be 25% of revenue” claim dropped into a sales deck like it’s a law of physics. Before you let any of those numbers influence a budget decision, you need to know what they actually measured, over what window, and whether the brand publishing them had any incentive to round up.
This isn’t a debunking exercise. The results below are real and several of them are genuinely instructive. But jewelry is a category with its own math: 12-36 month purchase cycles, a campaign placed order rate that sits at the bottom of every cross-category benchmark report, and a last-click attribution model that structurally undercounts email’s influence on a purchase someone researched for eight months before buying. A case study that ignores that context isn’t lying, but it isn’t giving you a number you can actually use to evaluate your own program either.
What you’ll learn
- Why jewelry email results have to be read against jewelry-specific benchmarks, not generic eCommerce averages
- A three-question filter for judging any case study claim before you trust it
- Breakdowns of six real jewelry-brand results, each with what the numbers prove and what they don’t
- The pattern that actually repeats across every credible result, and why it isn’t “better creative”
- What a realistic performance target looks like for your own jewelry program
Why jewelry email results skew differently
Jewelry is a category where the standard eCommerce email playbook produces misleading benchmarks if you apply them without adjustment. Average campaign placed order rate across jewelry sits at roughly 0.08%, the lowest of any major retail category, against an all-category average closer to 0.16%. Click rates on jewelry campaigns run around 1.60%, so people are opening and reading. They’re just not buying off a single promotional email, because a $2,000 engagement ring or a $400 pendant isn’t an impulse click.
Ninety-day repeat purchase rate for luxury and jewelry brands runs 9-11%, with top performers near 11%. That’s not a weak number in context. It’s a category where the actual purchase cycle runs 12 to 36 months, which means 90-day RPR is measuring almost nothing about whether your retention program is working. Second-purchase AOV and referral rate tell you far more, because they reflect what happens when someone finally does come back or send a friend your way, rather than penalizing you for a customer who simply isn’t due for another purchase yet.
This matters directly for case study evaluation. If a case study reports a lift in 90-day repeat rate for a jewelry brand, ask what baseline it’s lifting from and whether that metric was ever the right one to chase. If it reports flow revenue against campaign revenue, know going in that flows in this category typically convert at roughly 13x the rate of campaigns (2.11% average flow placed order rate versus 0.16% for campaigns), so a flow-heavy result isn’t surprising, it’s structural.
How to read the numbers: a three-question filter
Before any case study number changes how you plan your own program, run it through three questions. This is the same logic behind Retention Side’s diagnostic sequence for underperforming programs: deliverability first, then list health, then flow coverage, then creative and offer, in that order, because fixing a downstream layer without addressing the upstream problem produces temporary improvement at best. The same ordering discipline applies to reading someone else’s results.
Question 1: Who published it, and what’s their incentive? Klaviyo’s own customer stories carry platform attribution, meaning the numbers come from Klaviyo’s dashboard reporting on Klaviyo-driven activity. Agency-hosted case studies (the kind published by the agency that did the work) are the agency’s own reporting, self-selected and self-interpreted. Neither is automatically wrong, but they answer different questions. Platform stories tell you what a well-built program can do inside a platform’s own numbers. Agency stories tell you what a particular agency wants a prospect to believe about their engagement.
Question 2: What kind of metric is it? “38x ROI” and “+52% attributed revenue” and “+80% open rate” are three completely different measurement types, and case studies frequently blend them without telling you which one actually reflects the health of the retention system versus which one is an easy number to make look impressive. An ROI multiple compares spend to attributed revenue and can look enormous on a small media spend. A revenue share (like “37% of revenue from Klaviyo”) tells you concentration, not growth. A flow conversion lift tells you something specific and useful about a single automation. Know which one you’re looking at before you compare it to your own numbers.
Question 3: What’s the category context? A 0.16% campaign placed order rate might be mediocre for apparel and excellent for jewelry. Judge every claim against jewelry-correct baselines, not the cross-category average, and check whether the case study’s window (some of these span 2020 to 2025) reflects current platform behavior or a snapshot from a different competitive and deliverability environment.
Case study breakdowns
Catbird: relationship-led education replacing the in-store consultation
Catbird reported +60% owned email revenue year over year, +40% revenue from existing customers, and +34% new online customers, built around Klaviyo segmentation developed with agency partner Imagination Media. The strategic logic: jewelry is an emotional, deliberative purchase, and fine jewelry brands historically relied on in-store staff to answer the questions that talk someone into a purchase. Catbird’s email program replaced that consultative layer with product education content delivered by segment.
What it proves: education-first email content can do real work in a category where the sales conversation traditionally happened face to face, and segmentation lets you deliver the right education to the right audience instead of blasting everyone the same content.
What it doesn’t prove: this data window is from around 2020-21, several platform and privacy-related shifts ago (iOS 14.5, Apple Mail Privacy Protection, and Gmail/Yahoo sender requirements have all landed since). Treat the percentage lifts as directional evidence of a strategy that works, not as a number you should expect to replicate today.
Dana Rebecca Designs: flows as the core revenue engine
Over an August 2021 to August 2022 window, Dana Rebecca Designs attributed 37% of total revenue to Klaviyo, with 22% of revenue coming specifically from abandoned cart recovery and a 51% average flow open rate. The brand split abandoned cart messaging by first-time versus repeat buyer and by product category, and ran A/B tests differentiating voice for self-purchasers versus gift-givers.
What it proves: flows, not campaigns, are where fine jewelry revenue concentrates, which lines up with the category-wide pattern of flows converting far above campaign averages. Segmenting abandoned cart by buyer intent (self-purchase versus gift) is a jewelry-specific nuance that generic cart-recovery templates miss entirely.
What it doesn’t prove: a 37% revenue attribution figure depends heavily on the attribution model and window used, and a different measurement approach on the same program could produce a meaningfully different number. Don’t treat this as a portable percentage.
J&Co Jewellery: segmentation and testing discipline
J&Co reported 38x Klaviyo ROI in 2024, a 112% year-over-year increase in email click rate, and 300+ A/B tests run across the year after migrating from Mailchimp. The brand gated exclusive offers to customers with three or more prior purchases and used browse-triggered, limited-time discount emails.
What it proves: testing volume and segmentation depth compound. Three hundred tests in a year isn’t a lucky campaign, it’s a program that treats every send as a hypothesis, and loyalty-gated offers protect margin better than blanket discounting while still rewarding your best customers.
What it doesn’t prove: an ROI multiple like 38x is highly sensitive to the denominator (platform spend, not total program cost), and platform-to-platform migrations (Mailchimp to Klaviyo) often show inflated early gains simply because the prior baseline was poorly optimized, not because Klaviyo itself is 38 times better.
Arms Of Eve: data-driven personalization at scale
Arms Of Eve reported 62x Klaviyo ROI and +52% year-over-year Klaviyo-attributed revenue in 2024, with a 39.8% increase in flow conversion value. The program used a personalized welcome journey, granular engagement segments (0-15 days, 16-30 days, and so on) cross-referenced with product-interest segments like earrings versus necklaces, and a regional heat map to inform send times.
What it proves: stacking two segmentation dimensions, engagement recency and product interest, produces measurably better targeting than either dimension alone. This is a genuinely transferable tactic for any jewelry brand with enough SKU-category spread to segment by product type.
What it doesn’t prove: the ROI multiple again reflects platform spend, not full program cost, and a single-year snapshot doesn’t tell you whether the gains are durable or a one-time correction from an underbuilt starting point.
Boodles: the luxury audit gap
Boodles, a luxury jewelry brand, saw email open rates increase by 80% after a full Klaviyo audit conducted by agency Unified. The audit found missing flows, weak segmentation, and no connection between online customer actions and the sales team. The rebuild added flows like “Drop a Hint,” wishlist reminders, and an engagement-ring research journey, with webhooks alerting the sales team when a customer took a high-intent action online.
What it proves: this is the clearest illustration in this set of the gap between “installed” and “built.” A brand can have Klaviyo running and still be missing the flows that matter most for how people actually shop for luxury jewelry, and connecting online behavior to a human sales team is a distinctly luxury-category tactic that mass-market brands don’t need but high-consideration purchases benefit from enormously.
What it doesn’t prove: the agency’s own claim that “email should typically contribute 25% of revenue” is an unverified assertion from the case study publisher, not a benchmark backed by independent data, and the published case study page contains an incomplete quote section. Treat the 80% open rate lift as credible (it’s a specific, measurable, believable number for a program moving from “barely built” to “properly built”) but don’t adopt the 25% revenue figure as a target.
Shinola: deliverability-first segmentation at BFCM
Shinola, a watches, jewelry, and lifestyle brand, reported Klaviyo-attributed BFCM revenue up 27% year over year, revenue per email recipient up 48%, and flow revenue up 40%, working with partner Zettler Digital. The work prioritized deliverability before scaling send volume, then layered segmentation by product category and engagement level, with outlet offers targeted specifically to sale-prone segments rather than the full list.
What it proves: protecting inbox placement before increasing volume is the correct sequencing, especially heading into the highest-volume sending period of the year, and isolating discount-motivated segments so full-price buyers aren’t trained to wait for a sale is a sound practice for any jewelry brand trying to protect margin.
What it doesn’t prove: label this one honestly as a mixed-category result. Shinola sells watches and lifestyle goods alongside jewelry, so the BFCM lift isn’t a pure jewelry-category data point, even though the deliverability-first sequencing lesson still applies.
The pattern that actually repeats
Read across all six cases and the same four things show up in every credible positive result: deliverability protected before volume scales, segmentation built on more than one dimension (recency plus product interest, or purchase count plus behavior), flow coverage that goes beyond the basic welcome and abandoned cart defaults, and relationship or education-focused sends that don’t rely on discounting.
None of these brands won because of a clever subject line or a beautifully designed email. They won because the underlying structure, list health, segmentation logic, and flow coverage, was built correctly first. That’s the same four-pillar sequence (deliverability, list growth, flows, campaigns) that determines whether a jewelry email program actually performs or just runs. Creative quality matters, but it’s the layer that determines how well an already-sound structure performs, not a substitute for the structure itself.
What these results don’t prove
Take the whole set together and a few limitations apply across the board, not just to individual cases. Attribution windows vary from 2020 through 2025, and none of them account for how much harder last-click attribution makes it to credit email for a purchase that happened after an 8-month research cycle. Agency-reported numbers (Boodles, Porter Lyons, Chupi, Design Gold Jewelry) reflect the agency’s own measurement choices, not independently audited figures, and should be read as directional evidence rather than verified benchmarks.
Watch specifically for relative-percentage inflation. Design Gold Jewelry’s case study reports open rates rising from roughly a 1% baseline, and describing that recovery as a “3,700% increase” is mathematically accurate but practically misleading if you don’t see the starting number. A jump from a 1% to a 37% open rate is a real, valuable, deliverability-driven recovery. Presented only as a multiplier, it sounds like a different (and less believable) achievement than what actually happened.
What a realistic jewelry email target looks like
Instead of chasing someone else’s headline number, hold your own program against jewelry-correct benchmarks. Campaign placed order rate around 0.08% and campaign click rate around 1.60% are your category baseline, not a red flag. If your flows are converting anywhere near the category averages, abandoned cart around 3.33%, welcome series around 2.32%, browse abandonment around 0.95%, post-purchase around 0.54%, your automation layer is doing its job even if campaign revenue looks thin by comparison.
The cross-category gap is easiest to see in the benchmark table itself:
On the automation side, the gap between flow types is just as clear:
Stop measuring success on 90-day repeat purchase rate alone. With a 12-36 month purchase cycle, second-purchase AOV and referral rate are the indicators that actually reflect whether your retention system is working. And pay closer attention to first-to-second purchase conversion (typically 18-30% across eCommerce, though jewelry’s cycle length shifts the timeline, not necessarily the rate) than to any single campaign send, because that transition is where most jewelry brands lose customers permanently rather than temporarily.
Applied six times in a row, that filter is easier to follow as a decision path than as prose:
From case study evidence to your own system
Every result above that held up came from the same underlying structure: deliverability protected first, list quality managed deliberately, flow coverage built past the defaults, and campaigns layered on top rather than carrying the whole program alone. That’s not a coincidence and it’s not unique to any one agency’s methodology. It’s what a retention system looks like when it’s actually built instead of just installed.
Retention Side works with eCommerce brands, including jewelry and luxury, above $300K/month that have acquisition working and need the retention side of the business built to match. Klaviyo email is the entry point, but the same four-pillar structure, deliverability, list growth, flows, campaigns, extends into SMS, push, loyalty, and other channels based on what your specific audience actually responds to. The diagnostic order matters as much as the tactics: we look at deliverability, then list health, then flow coverage and trigger reliability, then creative and offer quality, because fixing the wrong layer first wastes time and produces results that don’t hold. If you want to know where your own jewelry email program actually stands against the benchmarks above, contact Retention Side for a structured email program assessment.
The case studies in this article are proof of concept, not a promise of your outcome. What they consistently demonstrate is that the brands getting real results built the underlying system first and let the metrics follow, rather than chasing a headline number and hoping the structure would catch up. That’s the distinction worth carrying into your own next decision, not the multiplier in someone else’s case study.