A programme with excellent economics, being dismantled by its own admin.

The programme launched on Awin roughly two years before the audit and was running at a mid-six-figure monthly revenue level — approximately $166,000 per month at the point of audit. That headline figure masked a programme that had moved from rapid early growth into a sharp decline. Revenue had peaked a year earlier and was falling 25% month over month through what should have been its strongest season. Conversion had collapsed from the programme’s strongest historic months to sit at the absolute floor of the jewellery sector benchmark range (1.0–3.0%, AME Reference Libraries). This was not sector seasonality. The jewellery sector had not contracted by anything like this magnitude in the same window.

The most damaging structural problem was payment trust. Average payment time was measured in months rather than weeks — many multiples of the 30-day benchmark — and the account had exceeded its credit limit with overdue invoices outstanding. Transaction validation was explicitly paused, with a banner in the account stating that validation had been halted pending invoice payment. The Awin Index sat well below the 70% threshold that signals a healthy programme to prospective publishers. This payment behaviour poisoned every other growth lever available: cashback and loyalty publishers require reliable cash flow before they will commit inventory, and the programme’s several hundred dormant publishers were never going to re-engage while payouts ran months late.

Revenue concentration presented an existential risk. Three publishers accounted for 79% of revenue. The publisher mix was dangerously skewed towards ad network publishers, which contributed roughly half of all clicks while delivering the lowest conversion rates in the programme. Content, editorial, influencer and creator publishers were severely underrepresented — and in the top-20 revenue list, influencer and editorial partners were absent entirely. For a category as visual and as aspirational as jewellery, that is the single most consequential structural gap in the programme.

Commission structure was fragmented and expensive in both directions at once. The default rate sat at a flat 5% — below the 8% jewellery sector floor — which discouraged new publisher recruitment, while undocumented publisher-specific overrides pushed individual rates as high as the 25% the programme publicly advertised. The resulting effective programme-wide rate was materially above the 8–12% sector benchmark. Three category-specific commission groups captured zero transactions, because the tracking integration required to route sales into those groups had never been implemented. The groups existed; nothing had ever flowed through them.

Fraud and compliance exposure was significant. Eighteen publishers generated over 1,000 clicks each with zero transactions across the entire audit window, collectively consuming a double-digit percentage of all programme clicks. At the programme’s own conversion rate, legitimate traffic of that volume would have produced thousands of transactions; the probability of zero is effectively nil. A major card-linked browser extension partner, receiving the programme’s highest commission rate, had been confirmed through class-action litigation to overwrite affiliate cookies — meaning a substantial share of its attributed revenue was misattributed away from the publishers who had actually originated the sale. A separate ad network partner generating six figures in attributed revenue had a near-empty website, minimal verifiable presence, and had recorded ghost transactions at zero value.

Operationally the programme remained reactive. Publisher outreach had increased in the months before the audit and a triggered communication framework had been built, both genuine positives. But only four of the sixty-nine transacting publishers carried tags, Partner Discovery and the Opportunity Marketplace were both access-denied, the Campaign Tool was unavailable, and Product Reporting was switched off. There was no structured relationship management framework for the top publishers, no seasonal activation calendar, and no documented management cadence beyond auto-validation.

The good news, and the reason the recovery plan in this document is credible: the core economics were strong. Earnings per click ran several times the sector average and represented a compelling recruitment proposition that was, at the time of audit, completely invisible to prospective publishers. The 158,000-product feed was active and current. The site itself was established with strong trust signals. Seasonal alignment was solid, with clear Valentine’s, Mother’s Day and holiday peaks visible in the revenue data. Voucher attribution was correctly configured on every offer. These fundamentals meant the programme could be rebuilt into a high-performing channel — but only after the billing crisis was resolved, the commission structure rationalised, the fraud exposure addressed and the publisher base diversified.

2A. Metric scorecard

Figures shown as a rating rather than a number are commercially private to the client. They are reported here against the same benchmark the client document used, so the assessment is intact even where the underlying value is not published. Benchmarks are AME Reference Library values for the jewellery sub-sector and are not client data.

2B. Area scorecard

Exceptional earnings-per-click economics. The programme’s EPC exceeded the jewellery sector ceiling by a wide margin and lifetime conversion sat mid-sector within the 1.0–3.0% jewellery benchmark. This combination means publishers who did promote this brand earned significantly more per click than on competing jewellery programmes. AOV was solid for the sector. These economics were the programme’s single most powerful competitive asset and should have been featured prominently in every recruitment material, the programme profile and all publisher communications. To protect this strength the conversion decline had to be diagnosed urgently — if conversion kept falling, the EPC advantage would erode and take the programme’s primary recruitment differentiator with it.

Extensive and current product feed. The 158,000-product catalogue with same-day product changes and weekly imports provided a strong foundation for product-level affiliate marketing. Breadth covered diamonds, gemstones, lab-grown diamonds, engagement rings, fashion jewellery and a licensed designer collection. This depth is a genuine differentiator — many competing jewellery programmes operate with catalogues an order of magnitude smaller. To convert the asset into revenue, Product Reporting needed enabling (a single checkbox) so publishers could see product-level performance, and product-feed creatives needed deploying to leverage the catalogue for dynamic advertising.

Established brand with strong trust signals. The brand had operated for over two decades, held a strong rating with the national business bureau, and presented a professional commerce experience with free worldwide shipping, a lifetime guarantee and a 60-day returns policy. The site carried dedicated affiliate and coupon landing pages, both of which support the channel. Review profiles on curated platforms were strong. This credibility makes publisher recruitment significantly easier once the operational issues are resolved. The one risk to this strength was review polarisation on complaint-specific platforms, suggesting inconsistent fulfilment quality — worth monitoring, but not undermining the core proposition.

Voucher attribution correctly configured. All voucher offers had attribution enabled with commission unchanged across commission groups. This is the correct setup and avoids the common failure where voucher publishers claim unattributed sales or commission is silently modified by voucher usage. Code hygiene needed attention, but the attribution infrastructure itself was sound and protected content partners from last-click voucher overwrite — a critical safeguard given the programme’s stated goal of recruiting more content and editorial publishers.

Triggered communication framework established. The account carried eight triggered communications covering the publisher lifecycle from onboarding through performance milestones: clicks-but-no-sales outreach at two thresholds, first-sale congratulation, mid-tier performance recognition, and high-traffic optimisation contact. All were created on a single date, indicating a deliberate build-out rather than incremental accretion. The two most critical lifecycle triggers — the welcome email and the 15-day activation check-in — remained in draft, but the framework itself demonstrated intent and could be completed in an afternoon.

Strong seasonal alignment. Programme revenue clearly tracked jewellery sector seasonality — Valentine’s Day, Mother’s Day, wedding season and the Black Friday through Christmas period all showed the expected peaks. The programme’s historic peak month demonstrated its revenue ceiling when seasonal demand and publisher activity aligned. This natural alignment means seasonal activation campaigns amplify existing demand rather than having to create it, which materially reduces the investment required to drive a seasonal lift.

5.1 Overview

The programme carried 720 registered publishers, of which a minority were classified as active — below the 60% benchmark. Of those active publishers only around a third generated any transaction at all. Roughly 111 partners counted as “active” sent clicks and converted nothing. Sixty-nine publishers recorded sales in the year-to-date window. The programme was in contraction, with revenue down 25% month over month and conversion sitting at the jewellery sector floor.

5.2 Type distribution

The jewellery sector reference mix calls for 40–50% editorial and content, 20–25% ad networks and PPC, 15–20% cashback and extensions, and 10–15% coupon and discount (AME Reference Libraries — Optimal Publisher-Type Mix, Jewellery). Ad networks at eight of the top twenty publishers are severely over-indexed. Content, editorial, influencer and creator publishers — which together should form the programme’s backbone in this category — are effectively absent from the top performer list. The complete absence of influencer and creator publishers is the programme’s most significant structural gap given jewellery’s visual, aspirational purchase journey.

5.3 Concentration heat map

Concentration is critical at every tier. The programme was effectively a three-publisher operation with a long tail attached. The top ten generated roughly 90% of revenue, leaving the remaining transacting publishers to share about a tenth between them. This is significantly more concentrated than the programme’s own lifetime picture, which suggests the publisher base had narrowed further during the audit year as legacy content publishers disengaged and ad networks filled the gap they left.

5.4 Device performance

Smartphone dominated both click volume and transaction volume, confirming that mobile was the primary purchase device — not merely the research channel. This is a significant finding, because the conventional jewellery assumption is that consumers research on mobile and buy on desktop. That assumption did not hold here. Desktop nonetheless delivered a substantially higher AOV, indicating that the highest-value purchases — engagement rings, diamond pieces, premium collections — were still being completed on desktop. Desktop’s materially lower conversion rate suggests either friction in the desktop purchase flow or that desktop visitors sit higher in the funnel.

Tablet showed the highest conversion rate of any device on minimal volume, which may reflect the deliberate, considered purchase behaviour typical of tablet users further along the decision journey.

Cross-device measurement showed zero attributed transactions, indicating that cross-device tracking was either not enabled or not capturing the expected mobile-research-to-desktop-purchase path. Given the AOV gap between devices, enabling it would reveal whether high-value desktop purchases were initiated on mobile — a critical input to attribution accuracy and publisher crediting.

Implications for programme management. Creative and landing page strategy should be mobile-first given where the transactions actually occur. The desktop AOV premium means desktop-targeted campaigns should emphasise the high-value categories. Product-feed creatives should be optimised for mobile display. Publisher recruitment messaging should lead with the programme’s mobile conversion strength, which is a genuine differentiator in this category.

5.5 Pending approvals

Auto-approval was active on the account plan. At the time of audit the partnership list showed no pending applications in the queue. A small number of publishers had been declined and a handful had ended or left the programme. The approval rate sat above the 90% benchmark, indicating that approved publisher quality was not being compromised by the auto-approval setting — though the absence of manual review is precisely how the two largest click-fraud accounts were admitted without scrutiny in the first place.

The publisher mix was narrow, ad-network-heavy at the top, and structurally mismatched to the programme’s stated goals. The programme needed to recruit across content, editorial, influencer and comparison types while protecting its genuine cashback relationships. Mid-tier publishers in positions five to twenty should be the primary growth target — large enough to generate meaningful revenue, small enough that losing one is not catastrophic. The several hundred dormant publishers represented a real reactivation opportunity, but the payment crisis had to be resolved first or any reactivation campaign would fail on contact.

Partner names are replaced with type-and-rank labels. Every analytical column — role, evidence, cost, incrementality, action and next step — is the client document’s own assessment, unchanged. The decision framework (protect and grow / review commercially / monitor / deprioritise, remove, monitor) is the standard AME partner disposition set.

The programme had no structured publisher relationship management framework of any kind. There was no top-20 publisher plan despite the largest single partner generating over a quarter of revenue and the top three accounting for 79% of sales. The most commercially important relationships in the programme carried no documented management protocol, no scheduled review cadence and no differentiated communication strategy.

Publisher Tags showed only four tagged publishers out of the sixty-nine with sales in the year-to-date window. That means sixty-five of sixty-nine transacting publishers were untagged, including every one of the top twenty revenue generators. Without tags the Communication Centre cannot segment outreach by publisher type, value tier or engagement level — so every communication is either broadcast to the entire base or hand-targeted to one publisher at a time.

Communication activity showed signs of engagement but remained tactically limited. The Communication Centre was accessible and showed roughly twenty manual communications in the month before audit, primarily single-publisher outreach focused on activation. One promotional newsletter went to all publishers. But this activity was tactical and reactive rather than strategic: no scheduled calls or meetings with top publishers, no quarterly business reviews, no media pack distribution, and no early-access or preview programme for high-value partners.

Recommended tiered management framework

Content freshness on top publisher sites could not be verified from within the network account — that requires an external review of each publisher’s brand-related pages. Competitor-exposure monitoring was not in place: there was no tracking of which publishers also promote competing jewellery programmes, which in a category this concentrated is a material blind spot. Meeting and call history requires external evidence from CRM or email records, which was not available from the platform.

Immediate actions. Tag all sixty-nine transacting publishers by type and tier. Apply bespoke commission rates to Platinum publishers to protect those relationships before renegotiating anything else. Create a Platinum contact schedule. Distribute the programme’s genuine differentiators — above all its earnings per click — to Gold-tier publishers to encourage increased promotion.

Partner Discovery and the Opportunity Marketplace were both access-denied. Direct navigation to either returned an access error. Both are expected to be available on the account’s plan tier, and a persistent dashboard banner warning — noting that no offers had been made on any publisher opportunities recently, and that thousands of opportunities were live — confirmed the tools should have been accessible. Access had to be restored through network support before any structured recruitment could begin.

With both primary recruitment tools inaccessible, the programme had no mechanism to identify, evaluate or invite new publishers. The only channel was the passive publisher invite link. Auto-approval was active, meaning any publisher who found the programme was approved automatically. That removes friction, but it also means fraudulent publishers join without scrutiny — which is exactly how the two largest click-fraud accounts entered the programme.

Recruitment gaps by publisher type

The network’s own intelligence tool had identified a handful of recommended partners at the time of the original audit, but no action had been taken on any of them.

The programme’s earnings per click was its strongest recruitment asset by a distance, and it was not featured in the programme profile or in any publisher-facing material. The 5% default sat below the 8% jewellery sector floor, making the programme uncompetitive in any publisher’s side-by-side comparison — before the Awin Index even signalled the payment risk.

Recruitment priority actions. First, contact network support to restore Partner Discovery and Opportunity Marketplace access. Second, raise the default commission to 8% to reach the sector floor. Third, feature the programme’s earnings per click prominently in the profile. Fourth, target five to eight content and editorial publishers across jewellery, bridal and gift-guide verticals. Fifth, target two to three influencer or creator partners. Sixth, use the Opportunity Marketplace once restored to post tenancy offers for the coming wedding season. Seventh, cap or review the eight ad network publishers driving high-click, low-conversion traffic before concentration worsens further.

9.1 Current state

The commission structure was simultaneously too low and too high. The 5% default sat below the jewellery sector floor of 8%, creating a barrier to recruitment — prospective publishers comparing programmes saw a below-market rate and moved on. Meanwhile undocumented publisher-specific overrides pushed individual rates up to the 25% the programme publicly advertised, producing an effective programme-wide rate above the sector ceiling.

The Commission Manager showed four commission groups all set at the same default rate, and twelve publisher rate tiers, with the large majority of publishers on the standard rate and a much smaller group on elevated rates. The category-specific groups captured zero transactions because the tracking integration to route sales by product category had never been implemented. No basket-value rules existed — the section was empty. No assist or multi-touch rules existed. The sole bonus rule awarded a flat sum at a low transaction threshold, with no tiering. The commission-by-assist section was empty, meaning upper-funnel publishers who introduce a customer but do not close the sale received no credit at all.

The publisher-level commission analysis revealed wide, undocumented variation. Several partners received rates two to five times the default. None of those elevated rates carried documented justification tied to performance criteria, incrementality evidence, or a negotiated agreement with an expiry date. This is the single clearest example in the audit of the gap between what the programme advertised publicly — “up to 25%” — and how the rate was actually allocated: not by performance, but by whoever had negotiating leverage.

9.2 Recommended architecture

9.3 Budget impact

Moving the default from 5% to 8% increases commission cost on standard-rate publishers by roughly 60% in isolation — but since a meaningful group of publishers already received elevated rates, the actual programme-wide impact is considerably smaller. Simultaneously reducing top-tier rates to a 12% cap offsets the default increase. Net effect: the effective rate moves from above the sector ceiling down towards the 10–11% range, reducing total commission cost while improving recruitment competitiveness at the same time.

That is the counter-intuitive result worth stating plainly. Raising the headline rate the market sees, and lowering the rates nobody could justify, moves cost down and competitiveness up simultaneously. The basket-value rules add incremental cost only on high-value orders where margin is strongest. The category-specific rates require the tracking integration to be completed before they can take effect at all.

Across the 129 tasks the priority distribution was 6 critical, 16 urgent, 29 high, 61 medium and 17 low. The client document carries each task with fifteen columns; the fields are laid out here as a card because fifteen columns is unreadable on any screen. No field has been dropped in the transform.

Day 0–30: stabilise

The first thirty days stop the financial bleeding, remove the immediate fraud exposure and lay the operational foundation. Nothing in the growth plan works until this is done.

30-day success criteria. Pending backlog at zero. Validation resumed. Fraudulent publishers suspended. Default commission at 8%. Auto-validation at 30 days. Terms complete. All publishers tagged. Wedding season campaign active. Welcome email rewritten. Both lifecycle triggers live.

Day 31–60: tune

60-day success criteria. Active publisher rate above 60%. Ad network share of the top 20 declining. Effective commission rate approaching benchmark. Newsletter cadence established. Welcome pack published. All compliance investigations concluded.

Day 61–90: scale

90-day success criteria. Revenue trend stabilised or growing. Top-5 concentration below 65%. Active rate above 65%. Ad network share declining. Effective commission within benchmark. Validation at 30 days. Awin Index improving. Peak-season calendar in progress. At least two new publisher types generating revenue.

12.1 Standing cadence

12.2 Retail calendar moments

Jewellery seasonality creates clear activation windows that have to be planned forward rather than reacted to. The complete jewellery sector calendar (AME Reference Libraries) carries fourteen activation moments. At audit the programme was covering seven of them.

Immediate actions at audit. Father’s Day was four weeks away with no campaign planned. Wedding and engagement season was imminent with no preparation of any kind. Anniversary season briefing needed to begin within weeks. Every remaining moment had to be pre-scheduled against the briefing dates above.

No. This programme could not scale in the state it was found. Paused transaction validation, an exceeded credit limit and an Awin Index well below the healthy threshold created a hostile environment for publisher recruitment and retention. The commission structure simultaneously underpaid at the default and overpaid at the top without justification. Revenue concentration in three publishers meant any growth built on the existing base was inherently fragile — particularly when three of the top five were ad networks with unverified traffic sources. The programme had to stabilise before it could grow.

What held it back. Three structural issues, in order of severity. First, payment trust: until invoices were cleared, validation resumed and payment time moved towards 30 days, no operational improvement would reach its full effect. Cashback publishers will not join, dormant publishers will not reactivate, and the Index will keep suppressing applications. Second, commission economics: the 5% default actively repelled new publishers while undocumented overrides inflated cost without accountability — the programme was paying premium rates to publishers it could not even verify. Third, fraud and compliance exposure: a double-digit share of clicks converting nothing, a six-figure publisher with no verifiable presence, and a top-rate publisher confirmed to overwrite cookies. These had to be addressed in parallel with the payment crisis, not sequentially after it.

The first five tasks, and why that order. One: clear the invoices and resume validation, because the programme may be suspended otherwise and every other improvement depends on it remaining live. Two: suspend the zero-transaction high-click publishers, because it is a one-day action that immediately removes the largest block of click waste and mechanically improves the conversion rate by taking fraudulent traffic out of the denominator. Three: raise the default to 8% and cap elevated rates at 12%, because the structure blocks recruitment and inflates cost at the same time, and fixing both sides unlocks every subsequent growth initiative. Four: diagnose the conversion collapse, because it is the primary driver of the revenue decline and no amount of recruitment compensates for a site that is not converting. Five: tag every publisher and activate the two draft lifecycle triggers, because these are half-hour tasks that immediately restore segmentation and complete the lifecycle framework, which everything downstream depends on.

What the brand should not do yet. No investment in premium publisher placements, paid media through the affiliate channel, or significant creative production until the payment crisis is resolved and commission is rationalised. Spending on recruitment before fixing payment trust buys churn at full price. Launching seasonal campaigns before clearing the backlog sends publishers contradictory signals. Building product-feed creatives before enabling Product Reporting means nobody can measure whether they worked.

What to review in 30 days. Has validation resumed? Has the Index improved? Is the pending queue holding at zero? Has the conversion decline stabilised? What were the outcomes of the two fraud investigations? Where has the extension partner rate renegotiation reached? Has Partner Discovery and Opportunity Marketplace access been restored?

What requires external evidence before a final conclusion. Analytics funnel data is needed to diagnose the conversion collapse definitively — a tracking audit identifies network-side issues, but if the problem is on-site, only funnel data reveals it. Traffic source verification is needed from the six ad network and sub-network partners who collectively represent the majority of revenue and clicks, none of whom had verified sources. The brand’s finance team must confirm margin tolerance to set the ceiling for commission restructuring. Updated brand guidelines are needed to establish a current compliance standard.