Case Study · Fine Jewelry · Awin

We found exactly where a $2M jewelry program was leaking revenue.

A US fine-jewelry retailer running a mid-six-figure monthly program on Awin came to AME with revenue sliding and eight agencies' worth of dead ends behind them. The obvious stuff — slow payouts, a validation backlog — was real, but easy. The revenue was hiding somewhere less obvious: in the commercial terms with their biggest partners, and in a network setup that was quietly capping every sale. Here's what most brands never think to look at.

$166K
Monthly affiliate revenue at audit
549
Findings surfaced (38 critical)
79%
of revenue riding on just 3 partners
−25%
Revenue trend, month over month
The situation

A healthy-looking program losing ground every week

SectorFine & Fashion Jewellery
MarketUS
PlatformAwin
Engagement~8 months
Managed periodNov 2025 – Jun 2026
Registered publishers720
Product catalog (SKUs)158,000+
Prior agencies8
Revenue trend, MoM−25%

On paper the program had scale: 720 registered publishers, a six-figure monthly run rate, a full product catalog of 158,000+ SKUs. But revenue was down 25% month over month and nobody could say why. The brand assumed it needed more affiliates. It didn't. It needed the program it already had to actually function. It had also worked through eight agencies before coming to AME, and not one of them had found, let alone fixed, what was actually wrong.

We ran a read-only deep-dive audit across the entire Awin account — commercial terms, tracking, commission logic, product feed, publisher mix and offers — and produced 549 discrete findings, 38 of them critical. The pattern was clear immediately: this wasn't a growth problem, and it wasn't the obvious admin either. The real money was tied up in the deals with their biggest partners and in network setup that quietly capped every sale — the two things nobody had touched in eight agencies.

What was leaking

Where the money was really hiding

  • The top partners were under-rewarded and drifting. Every affiliate sat on a flat 5% default while the program publicly promised "up to 25%", so the highest-volume partners had no reason to lean in — and the single biggest had collapsed 90%+ in one month. Nobody had renegotiated a thing.
  • The network setup was silently capping every sale. The conversion tag was missing at checkout, commission matching was broken, and the 158,000-product feed was mis-categorised as "Fashion" and two months stale — losing and mis-attributing revenue before anyone could see it.
  • 79% of revenue rode on just 3 publishers, while ~111 "active" partners sent clicks and converted nothing. No leverage, no diversification, no plan for either.
  • And yes, the basics were overdue too — slow payouts, a pending backlog. Real, but easy. Table stakes, not where the revenue was.
What most brands miss

The levers nobody thinks to pull

  • Anyone can spot a slow payout. The upside was in renegotiating the top partners and fixing the network setup — the levers nobody thinks to pull.
  • Move those and revenue rises with zero new traffic and no new partners. You're capturing and re-pricing what's already there.
  • None of it is obvious from a dashboard. It takes someone who knows exactly where programs leak and where partners will move.
  • Eight agencies had walked straight past all of it.
Why eight agencies missed it

They were trained to run the program, not to question it

The last eight teams weren't lazy. Affiliate account management, almost everywhere, is trained around daily admin: approve these transactions, post this deal, send that newsletter. Almost nobody is trained to ask whether the program's foundations are even sound, or how to rebuild them when they aren't.

AME doesn't run a playbook. Every program is broken in its own way, so we read each one from scratch and get our hands dirty in the account until we find what's really holding it back — the things neither the brand nor the last agency ever thought to check.

Findings the last eight agencies missed

What we found that nobody else did

  1. Last-click attribution quietly handing credit — and commission — to the wrong partners.The dashboard looked fine; the routing didn't. Commission was leaving the program for publishers that hadn't actually driven the sale.
  2. A landing page flawed from the first click.Conversion was dying long before commission ever mattered — so every fix downstream was capped by a top-of-funnel problem nobody had opened.
  3. Commission logic that contradicted the public promise.The best partners were underpaid against the "up to 25%" the program advertised, and no one had flagged it.
  4. 43 affiliates promoting expired or unauthorised coupons.Codes that were long dead or were never theirs to run — quietly eroding margin and undercutting full-price sales.
  5. Tracking and feed faults silently capping the ceiling on everything above them.Every optimisation upstream was fighting a ceiling nobody could see, because the ceiling itself was invisible on the dashboard.

None of it shows up on a daily to-do list. All of it shows up in the revenue.

The approach

Fix first. Grow second.

We don't hand over a 40-page report and walk away. The audit becomes a prioritised task list — plain, sequenced, and owned. And the tasks that actually move revenue aren't the obvious admin. Here's where the money came from, in order:

Renegotiate your highest-value partners Renegotiate

Your biggest affiliates are also your most negotiable — but almost nobody reworks the deal. We reset commercial terms with the top partners (rate, placement, exposure, exclusivity), win back the ones who'd collapsed, and turn a passive payout into an active partnership. This is where the biggest single lift came from.

Rebuild the commission architecture to match the promise Fix

End the flat-5%-for-everyone default and finally deliver on the "up to 25%" the program advertised — rewarding by publisher type and basket value so the right behaviour actually pays, and your best partners have a reason to prioritise you.

Fix the network setup that caps every sale Fix

Reinstall the conversion tag at checkout, repair commission matching, and re-categorise the 158,000-product feed to Jewellery and refresh it daily — so every sale tracks, attributes and pays correctly instead of leaking out silently.

Break the reliance on three publishers Grow

Reactivate the ~111 dormant "active" partners and go after the hardest partners in affiliate to win — content publishers — so the program stops living or dying on three accounts (more on that below).

Clear the basics Hygiene

Yes, we also settle the overdue invoice, unfreeze validation and clear the pending backlog. Important and quick, but table stakes. We get it off the board early so the real work can run.

17
content publishers onboarded, with zero upfront fees. AME's system surfaced 60 target blogs, each with contact details and ready-to-send outreach. Seventeen came on board, fee-free. Ask any brand how often they land a content site without paying to be there — most are lucky to get one a month. We got 17.
The opportunity
Projected recovery · illustrative

What a fixed program of this size can unlock

Based on the audit findings, here's the kind of recovery available from renegotiating the top partners and fixing the network setup — before a single new publisher is even recruited.

~20%
Revenue lift, typical, over 6–8 weeks of fix work
~$1,700/day
Recovery target as fixes compound (~$51K/month)
$0
Extra ad spend or new traffic required

These recovery figures are illustrative projections built from the audit's findings and AME's typical outcomes when these specific issues are resolved on a program of this scale. They are not a guaranteed or booked result, and every program differs. The diagnostic figures above (revenue, pending value, payment time, findings count) are the actual state of the audited account.

The result · managed period

Eight months on: growing through the weaker season

We didn't hand the audit over and leave. We ran the programme for eight months, Nov 2025 to Jun 2026. It ended with the programme growing through its weaker season instead of shrinking through its strongest, the commercial terms rewritten, the tracking sound and the partner base diversified.

$1,297,000
Total sales driven
4,184 tracked transactions
$193,000
Monthly revenue at handback
June 2026 — a mid-season month
~8 mo
Audit to recovery
First fixes live Dec 2025
1,562,000
Qualified clicks delivered
Traffic into the brand site
214
Active publishers
Producing, not just recruited
30%
Active-partner rate
vs 10–20% industry norm

Monthly sales revenue — managed period (USD)

Seasonal, not smooth — three months of repair before the account compounds
$166k
$189k
$104k
$142k
$131k
$158k
$214k
$193k
NovDecJanFebMarAprMayJun
Discovery & repair phase
Month-on-month decline
Scaled
Trailing quarter +23% vs the opening quarter
Peak month $214k in May (Mother's Day)
Like for like Jun $193k vs Nov $166k, off-peak vs peak
Reading the curve
Nov – Dec
Audit and diagnosis. Peak gifting season lifted the account, but it was still underperforming its own seasonal norm — which is what made the decline so hard to see.
Jan
The post-holiday collapse every jewellery programme lives through, with fixes live but not yet compounding. The lowest month of the engagement.
Feb – Mar
Valentine's lifts February, then March falls back — but $131k in March still beats the pre-audit run rate in a weaker month.
May – Jun
Mother's Day peak at $214k, then June settles at $193k — 16% above the audit month, off-peak against peak.

Diagnostic figures come from the opening audit; revenue figures are actual tracked sales. Jewellery is strongly seasonal — read each month against its calendar position.

The work, in numbers

What the repair cleared, and what the growth added

The first eight weeks were spent clearing the structural faults above. Only once they were closed did a single hour go into recruitment — which is why the first quarter looks flat and the second does not.

Critical findings cleared in the first eight weeks
Flat-rate commission Missing conversion tag Broken commission matching Mis-categorised feed 43 expired coupons Payment backlog
549findings surfaced in the opening audit, 38 of them critical.
38%of monthly revenue was being mis-attributed before tracking was repaired.
43affiliates found promoting expired or unauthorised codes, all removed.
$246k
in creator-driven sales. 19% of all programme revenue, from 138 creators and content partners onboarded during the engagement, all on commission-only terms with no upfront fees. Creators were layered alongside the content publishers on the same fee-free basis, so the programme stopped living or dying on three accounts.

Where the revenue came from

Share of tracked sales through the managed period, by partner type

Content & editorial
28%
Sub-networks & technology
24%
Influencers & creators
19%
Voucher & deal
14%
Cashback & loyalty
9%
Search & direct
6%
The outcome

A programme that grows instead of leaking

The brand kept the audit, the task list and the documentation — the same materials we used to run it.

01

The repairs

Commission architecture, tracking, feed categorisation and commercial terms rebuilt and documented, so the fixes hold after the engagement ends.

02

The playbook

The prioritised task list, recruitment targets and activation sequences, written down so the in-house team can keep working it week to week.

03

Ongoing support

We stay on call after handback. Whenever the account manager needs guidance, we're there — the relationship doesn't end when the programme moves in-house.

What the client walked away with

  • Trailing-quarter revenue up 23%, against an opening quarter that included peak gifting season.
  • A June of $193k against a November of $166k — off-peak beating peak.
  • Top-three partner concentration cut from 79% to 41%.
  • 549 findings closed or scheduled, including all 38 critical items.
  • A programme growing month on month instead of declining 25%.
The bottom line

From −25% a month to off-peak beating peak in eight months — on the traffic they already had.

$1,297,000 in sales · 4,184 transactions · 1.6M clicks · 30% active partners

The turning point

Eight agencies in, and one audit changed everything

This brand had cycled through eight agencies looking for someone who could actually move the needle. After seeing nothing but AME's first audit — the findings, the prioritised tasks, the plan — they ended their incumbent agency on the spot and moved their program to us to run.

It was the first time anyone had shown them exactly what was broken and precisely how to fix it. That's the difference between a report and a task list, and it's why the audit sells the relationship, not the pitch.

The takeaway

Most brands don't need more affiliates. They need the partners and the setup they already have to work harder.

The revenue is usually already there — in commercial terms that were never renegotiated, and network setup that quietly caps every sale. That's the work AME does that the last eight agencies didn't: find the levers no one else pulls, work them in priority order, then grow from a foundation that actually holds.

You came here to grow your program. Let's start with the revenue it's already leaking.

I've spent 28 years in affiliate, and almost every program leaks in the same places. I'll audit yours and hand you a plain, prioritised task list: exactly what to fix first and grow next. Free, and yours to keep forever. No card, no catch. Brands who work the list week to week see up to a 20% lift in traffic in 6 to 8 weeks. Imagine where that puts you in six months.

Free forever · yours to keep whether you hire me or not · about two minutes to start

About this case study. It describes a real fine & fashion jewellery program audited and subsequently managed by Affiliate Marketing Express. The client is anonymized for confidentiality. Diagnostic figures are drawn from the account audit; recovery figures are illustrative projections, not a guaranteed outcome. Managed-period figures — revenue, transactions, clicks, publishers and partner mix — are drawn from tracked network reporting across the managed period. Individual results vary by program size, category and market.