Most jewelry ad advice is opinion dressed up as best practice. Run video, it's what the algorithm wants. Post carousels, people swipe. Boost your best post. You've heard all of it, usually from someone who has never had to defend a media budget to a store owner.
This is different. What follows is a year of real numbers from a jewelry account we manage. Every format, every placement, broken down by what actually pulled clicks and what quietly wasted budget. We run the account, so the data is ours to read. The client is anonymized under our NDA, which means we report in click rate and cost per click instead of dollar totals, and we round our own numbers rather than hand you false precision. The patterns are the point, and the patterns are clear.
What we measured, and how to read it
Over the trailing twelve months, this account ran north of 3% CTR at about $0.30 a click, with a CPM around $10. For a considered, high-ticket category, a click rate above 3% is strong. It runs well above what we typically see on the accounts we manage, and most of the people clicking weren't impulse buyers, they were people researching a five-figure decision.
Quick translation, since these numbers get thrown around loosely: CTR is how often a view turns into a click, CPC is what each of those clicks costs you, and CPM is what a thousand views costs. That's the whole vocabulary you need for this teardown.
Two honest caveats before the findings. We group ads by the account's own naming convention, so the format comparisons are as clean as the labels we applied when the ads went live, and we report direction over decimal-point precision. And because the client is under NDA, we round our own numbers and frame everything by relative performance rather than dollar totals. The comparisons hold up. The dollar figures stay private.
If you want the wide-angle version of this, our monthly jewelry retailer advertising report tracks what the whole market is running. This post is one account, up close.
Pattern 1: Product over production
Here's the finding that surprises most owners. The polished video didn't win.
Group the account's ads by format and the static and catalog buckets both clear video on click rate, comfortably. When someone is weighing a serious purchase, a clean look at the actual piece pulls more clicks than a produced spot does. They want to see the thing, not a mood.
The single clearest proof is one plain product ad. It ran above 4% at roughly a quarter a click, served at scale. No film crew, no concept. Just the piece, shot well, beating everything else in the account on both attention and cost.
That doesn't make video useless. It moved the most volume by far, and it earned that volume without the click rate falling apart. Video's job in this account was reach. The product shots did the converting work. So before you book a film crew for your next campaign, remember that the piece itself, shot clearly, is the ad.
Pattern 2: Static is the cheapest click at scale
The click rate is only half the story. The static bucket also bought the cheapest clicks. Static product images came in noticeably cheaper per click than both video and catalog. That gap compounds across thousands of clicks.
And the best static above wasn't a fluke. At roughly a quarter a click, run at scale, it set the efficiency floor for the whole account, from nothing more than a plain product photo. Clean photography is the cheapest scaled click you can buy here. Everything else gets measured against it.
The good news for most retailers is that this is the one asset you already have, and probably underuse. You don't need a studio. You need a clean background, good light, and a piece worth looking at. We wrote a full walkthrough on how to photograph jewelry for your website without hiring anyone, and the same shots feed straight into your ads.
Pattern 3: Lead with a tight, curated set
Inside the catalog ads, the tightly curated multi-piece sets out-clicked the long tail of one-off single-product variants. The strongest catalog unit in the account was a small Valentine's grid, not a single ring on its own.
The lesson isn't quantity, it's curation. A wall of forty pieces is a decision to avoid. A short, themed set of the right few gives someone something to want without burying them. Curate hard, build the set around an occasion, and let your full inventory live on the collection page where the customer is already committed to browsing.
Pattern 4: Reels scales, Stories leaks
Placement is where a lot of budget quietly disappears. Here's where the money went.
Placement | CTR | CPC | Cost to serve |
|---|---|---|---|
Instagram Reels | ~3.4% | ~$0.25 | lowest |
Instagram Feed | ~3.7% | ~$0.35 | mid |
Instagram Stories | ~2.5% | ~$0.46 | high |
Reels did the heavy lifting. The most reach of any placement, a click rate around 3.4%, about a quarter a click, and the cheapest impressions in the account. Feed posted the highest click rate, close to 3.7%, but couldn't match Reels on volume or cost.
Stories is where the leak was. A click rate around 2.5% at nearly $0.50 a click, almost double the cost of Reels for a lower click rate. That's heavy spend at the weakest efficiency of the big three. And across every meaningful placement, Instagram out-earned Facebook. Facebook Reels lagged worst of all, under a 2% click rate.
The takeaway for an owner is simple. A vertical 9x16 cut built for Reels is the single highest-leverage asset you can produce right now. If you make one thing this quarter, make that.
> *This is the proof. The playbook for building and running these campaigns is in our Meta ads guide for luxury retailers.*
Pattern 5: Stop boosting posts
The worst performer in the account was a boosted post. It limped in under half a percent at nearly $2 a click, many times the cost per click of a real campaign for a fraction of the engagement. Not close.
The boost button feels like advertising. You see a post doing well, you tap boost, you put $50 behind it. But boosting optimizes for engagement, likes, comments, the things that make a post look busy. It does not optimize for clicks to your site or feet through your door. The data here says it underperforms every proper campaign format you could run instead. If you're boosting, you're paying a premium for the wrong outcome.
Pattern 6: Build creative around the gifting calendar
The single best ad in the account wasn't a clever new format. It was timing.
A Valentine's catalog push hit nearly 7% at under $0.20 a click, close to double our usual rate. That's the strongest efficient click rate of any ad we ran all year, and it happened because the creative landed in the window when people were already shopping for exactly that reason.
Your buying calendar is a creative calendar. Valentine's, anniversary season, the holidays, every occasion that drives someone into your category is a chance to put the right piece in front of the right buyer at the moment they're already leaning in. The catch is lead time. The ads that win these windows are planned weeks ahead, not thrown together the week of. Those famous brand campaigns built the names, and you can see the best watch ads ever made for how. This is what converts now.
The catch: creative only matters if you can trace it to the register
A near-7% Valentine's click rate is a great number. It is also useless on its own.
Here's the problem nobody selling you ad services wants to say out loud. A click is not a sale. You can run the best creative in your market, win every placement, and still have no idea which of those clicks walked into your showroom, sat down, and bought a ring. Click rate tells you what gets attention. It tells you nothing about what made money.
That's the gap we close. Creative is the front half of the work. The back half is connecting that click to a showroom visit, an appointment, and a sale you can see in your numbers. Without that, you're optimizing toward the wrong target, chasing cheap clicks instead of real revenue. We break the whole problem down in our guide to revenue attribution for luxury retailers, because the teardown above is only half useful until you can tie it back to the register.
Get your account torn down
If you're running Meta ads and can't say which creative drove a showroom visit, you're guessing with real money. We'll pull your account, tear down what's working and what's wasting spend, and connect it to revenue you can actually see. Book a creative and tracking review and we'll show you where your budget is going and where it should be.
OWNER INPUT NEEDED
These are the open items only Hagop can resolve before publish. None block the draft; all should be settled in owner review (Stage 3) or fact check (Stage 5).
- Slug vs. directory name mismatch. The brief's intended URL slug is `jewelry-ad-creative-teardown` (chosen for cannibalization control against the monthly report). The task created the working directory as `what-works-in-jewelry-ads`. The frontmatter uses the brief's slug. Confirm which slug ships. If `jewelry-ad-creative-teardown`, we should rename the directory to match before publish for consistency.
- Real creative examples. After the confidentiality pass the post no longer cites internal ad labels and describes the winners generically (the best static, the Valentine's catalog push). If you want to show actual creative as figures (the winning static, the Valentine's catalog ad, a Reels cut), confirm which images you're cleared to publish under the NDA and supply them. Right now the post shows no creative, only the (now rounded) performance data.
- The 20,946-study internal link. The brief's franchise plan links this teardown to the rescued `jewelry-ad-study` page. That page does NOT exist in the CMS yet (Part A work). Once it's live, add the cross-link ("competitor creative vs. our first-party creative, the two halves of the franchise"). Flagged in the outline's Internal Link Summary.
- CPM gaps in the placement table. RESOLVED by the confidentiality pass. The placement table no longer reports per-placement CPM (it now shows a qualitative "cost to serve" column instead), so the IG Feed / IG Stories CPM blanks no longer exist. Nothing to fill.
- Account-level CTR vs. category benchmark. RESOLVED in fact check. Section 2 no longer cites an external benchmark. The line now reads "It runs well above what we typically see on the accounts we manage," a first-party practitioner statement that asserts no unverifiable external figure. [VERIFY] tag removed.
OPEN [VERIFY] ITEMS
- RESOLVED: The "most retail benchmarks sit lower" line is reworded to a first-party statement (see #5 above). No external benchmark figure is asserted, so no source is required.
- All first-party figures (impressions, clicks, CTR, CPC, CPM, per-placement, the Valentine's and store_ad standouts) are from the account's live Meta data and are reportable as relative/non-dollar-total per NDA. No named client spend appears.
