Creator Affiliate Marketing Case Study in Pakistan

The Client

A Rawalpindi-based D2C hair-care brand — sulfate-free shampoos, scalp serums, and reformulated herbal hair oils across a tight 34-SKU catalog — selling nationwide through its own Shopify storefront. The brand was founded in 2022 by a formulation pharmacist who still signed off on every claim the company made publicly, ran a team of 24, and had built roughly PKR 21M in monthly online revenue. Around 66% of orders completed as cash-on-delivery, the default trust mechanism for Pakistani beauty and personal-care buying, and average order value sat near PKR 3,150, with routine-building bundles doing most of the lifting.

Acquisition was Meta-dependent to a fault: about PKR 5.8M a month in spend carrying roughly two-thirds of orders at a blended ROAS that had slid from 3.1x to 2.4x over two quarters. The founder’s instinct — shared by most Pakistani D2C operators in 2026 — was that creators worked, because customers kept mentioning them. The company’s records disagreed: a year of macro-creator activity had produced engagement spikes, a warehouse of screenshots, and not one attributable rupee.

When the brand approached WeProms Digital for influencer marketing campaign management, the brief was blunt. Creator spend had to stop being a brand-awareness line item priced on follower counts and start being a performance channel priced on revenue — with the attribution to prove it, and without wrecking a contribution margin the founder could recite from memory.

The Problem

Four issues defined the starting position:

• Flat-fee creator spend with no measurement. Six macro-creator posts over twelve months at an average of PKR 300k each — PKR 1.8M total — with no codes, no tracked links, no contracts, and no usage rights. Whether the posts sold product was a matter of internal folklore.
• Declining paid efficiency. Blended Meta ROAS had fallen from 3.1x to 2.4x as frequency on core audiences passed 5.0 and the same studio creative wore out. The team was paying more per order every month with no new creative engine feeding the account.
• A channel that was invisible where it mattered. An exit-of-funnel survey suggested roughly 9% of buyers had creator exposure on their path to purchase, while every dashboard the brand owned showed zero. A channel you cannot measure is a channel you cannot scale, defend, or cut.
• A margin fear with no model behind it. The founder assumed commissions would erode margin. Nobody had done the arithmetic — commission percentage against first-order contribution margin — to find the actual ceiling, so the conversation had stalled for two quarters.

The diagnosis was straightforward: the brand did not have a creator problem, it had a pricing and measurement problem. Creator spend was being bought like outdoor advertising and then judged like performance marketing.

Phase 1 — Creator Spend Audit and Program Economics (Weeks 1-2)

Phase 1 settled two questions before any creator was contacted: what had the flat-fee spend actually produced, and what could the brand afford to pay on a performance basis.

Post-mortem on the flat-fee year. We pulled the engagement and revenue-timeline data around each of the six macro posts. Every post produced a traffic spike to the storefront within 72 hours and a step-change in branded search — and no measurable order lift that survived the week. The macro creators’ audiences were broad-lifestyle, not concern-led, and their content format (polished single-image posts) generated admiration rather than purchase intent. The spend had bought awareness the brand could not capture.

Survey-based baseline. Over two weeks we added a one-question post-purchase survey — “where did you first hear about us?” — collecting 640 responses. Roughly 9% named a creator, about a third of those naming creators the brand had never worked with. The honest conclusion: creator influence was real but diffuse, the flat-fee posts were a small slice of it, and tracked code redemptions would understate the channel. We froze both numbers as the baseline so the program would be judged against a known gap between survey truth and trackable truth.

The unit-economics ceiling. With the founder, we built the commission model: average creator-code order value of PKR 3,280 against a 58% contribution margin, meaning roughly PKR 1,900 of first-order contribution per order. Commissions of 10–15% of revenue cost PKR 330–490 per order — leaving positive contribution on first orders alone, before any repeat purchase or content-asset value. The fear dissolved once the arithmetic was on one page: the brand could pay up to 15% and still earn more per creator-order than it earned on its fatiguing paid social.

Tier and attribution design. We designed a three-tier, commission-only structure with product seeding as the only upfront cost:

Tier	Follower range	Compensation	Program role
Nano	2k–15k	Product + 15% commission	High-trust, concern-specific conversion
Micro	15k–75k	Product + 12% commission	Workhorse content and steady redemptions
Mid	75k–250k	Product + 8% commission; retainer only after two quarters of proof	Reach spikes for launches

Attribution was designed around how Pakistani buyers actually convert: unique per-creator discount codes (the buyer gets an incentive, the brand gets a clean attribution key that survives screenshots, DMs, and cash-on-delivery checkout), shortened tracked links as the secondary signal, and a strict dedup rule — code wins over link — feeding a single creator revenue report dashboard. Commissions paid on first orders only, one redemption per customer per code.

Phase 2 — Affiliate Scaffolding and First Cohort (Weeks 3-5)

With the economics and attribution settled, Phase 2 built the machinery and the first roster.

Infrastructure. We configured the storefront’s discount engine for per-creator codes with single-use-per-customer limits and connected a referral app that gave each creator a live personal dashboard — orders, revenue, and pending commission visible in real time. This sounds like a convenience feature; in practice it was a motivation engine. Creators who can see their earnings ticking up post again without being asked.

Recruitment. Rather than an open call, which in Pakistan reliably attracts free-product hunters, we hand-recruited from a 120-name shortlist built on audience-overlap and engagement quality: nano and micro creators across Rawalpindi, Islamabad, Lahore, and Karachi whose content already sat inside the three concern pillars we had mapped — monsoon-season hair fall, dandruff and scalp health, and heat-and-styling damage. Personalized outreach referencing specific posts converted 46 creators in the first cohort, weighted roughly 70% nano and micro.

Briefs with claim discipline. Each creator received a seeding box containing a full routine (cost PKR 8,500 per box, PKR 391k for the cohort), a one-page brief covering texture, scent, and routine placement — and an explicit instruction to avoid therapeutic claims. Hair-fall in particular is a space where Pakistani creators drift into medical territory; the brief kept the framing cosmetic (“part of a scalp-care routine”) because a brand founded by a pharmacist could not afford DRAP-adjacent claim risk carried by third-party posters. A 20-minute product walkthrough call preceded every first post, because creators who could answer ingredient questions in their own DMs converted visibly better than those who referred buyers to the brand.

Cadence and support. Posting was staggered so tagged mentions stayed continuously active rather than spiking and dying, and a single WhatsApp support group handled creator questions inside business hours. By the end of Phase 2, all 46 creators were live, 38 pieces of content were in market, and redemptions were compounding in the dashboard — PKR 1.4M of attributed revenue across the cohort’s first five weeks, the majority from nano creators whose audiences trusted them enough to actually use a code.

Phase 3 — Optimize, Prune, and Amplify (Weeks 4-8)

Phase 3 is where the program stopped being an experiment and started being a channel.

The leaderboard and the prune. Every Monday, creators received a ranked redemption leaderboard. Four weeks in, the data was unambiguous: 11 of 46 creators had generated 78% of attributed revenue — the power law every experienced creator program expects. The difference here was that pruning cost nothing. Under commission-only terms, the 17 creators with zero redemptions were simply retired from active seeding and briefs, with no sunk fee to write off and the door left open. The roster concentrated itself around the 29 proven producers.

The commission-tier test. We split the micro tier into two cohorts for four weeks: 12% versus 15% commission. The 15% cohort posted 41% more often and delivered 58% more attributed revenue per creator, and the orders still cleared a 43% contribution margin after commission. The test settled the founder’s original fear with data: at this AOV and margin, paying creators more made the brand more, not less. The 15% rate was adopted for micro producers.

Whitelisting proven content. Every onboarding contract included 90-day paid usage rights at no additional fee — negotiated upfront, when it costs a courtesy, rather than after a post performs, when it costs leverage. In week six we licensed the nine strongest organic assets and ran them as whitelisted ads against matched audiences alongside the brand’s studio creative:

Creative source	CTR	CPA (first order)	ROAS at scale
Studio creative (baseline)	1.1%	PKR 2,940	2.2x
Whitelisted creator assets (top quartile)	1.9%	PKR 1,870	3.4x

The creator assets outperformed on every line — unpolished, concern-specific, delivered by a face the audience recognized from organic feed. Roughly PKR 0.58M of month-three program cost was licensing and amplification, and it was the cheapest incremental ROAS in the brand’s entire paid account. Brands that want more of this output without the recruiting overhead can commission dedicated UGC production alongside the program; in this engagement, the affiliate roster supplied more usable creative than the brand could amplify.

The incrementality check. The uncomfortable question about affiliate programs is whether codes merely re-label demand that would have arrived anyway. In week seven we paused all Karachi-based creator codes for fourteen days while keeping Lahore and twin-city codes live. Karachi redemptions fell 64% against the prior fortnight while total Karachi revenue moved within normal weekly variance — indicating the large majority of redemptions were incremental purchases the codes had prompted, with perhaps a tenth representing buyers who would have found the brand regardless. We logged that share honestly as a program cost of doing business, and the geo-pause became a quarterly verification ritual rather than a one-off.

Seasonality and code hygiene. The timing happened to align with the monsoon corridor, when humidity-driven hair-fall concern surges in Pakistani search and social conversation; the concern-led briefs meant the roster’s content met rising demand rather than fighting it. Three codes leaked onto coupon aggregators in week five; all three were rotated and single-use limits absorbed most of the abuse, keeping leak waste under 2% of attributed revenue for the period.

Phase 4 — Reconcile ROAS and Scale (Weeks 8-12)

The final phase made the number trustworthy and then made the system repeatable.

The ROAS reconciliation. Month three closed with PKR 5.6M of code-and-link attributed revenue against PKR 1.40M of program cost — commissions of PKR 0.67M at a blended effective rate near 12%, ongoing seeding for the expanded cohort at PKR 0.15M, and licensing plus amplification at PKR 0.58M. That is the 4.0x month-three run rate, and it is deliberately conservative: the whitelisted ads’ own platform-attributed revenue stays in the paid account’s numbers, while the licensing spend that enabled those ads is carried as program cost. Blended across the full ninety days, the program returned 3.9x on PKR 12.4M of attributed revenue — and the attributed revenue figure is the deduped tracked number, not the survey-inflated one. Total online revenue grew from PKR 21M to PKR 23.2M a month over the same window, so the creator channel’s growth was additive to a paid-social baseline that had been flat.

Order-level economics. At a creator-code average order value of PKR 3,280, month three carried roughly 1,700 attributed orders against PKR 1.40M of cost — PKR 824 per attributed order, down from an estimated PKR 1,385 in the flat-fee era (the older figure is itself generous, being survey-estimated rather than tracked). Every rupee of creator cost in the new model followed a confirmed order.

Scaling the roster the way networks scale. Productive creators recruit other creators. We formalized it — a flat 3% of referred-creator revenue for twelve months, capped — and the roster grew from 46 to 63 without a single additional cold-outreach hour. Nano and micro recruits in Karachi and Lahore filled geographic gaps the first cohort had left open.

A quiet retention signal. Creator-code buyers repeated at 24% within the measurement window against a site average near 19% — buyers who arrived via a trusted routine recommendation behaved like referred customers, not ad clicks. We flagged it for the brand’s retention work rather than claiming it as a program result, but it improved the channel’s honest lifetime-value math.

Institutionalizing the loop. The weekly leaderboard, the monthly prune, the quarterly tier review, and the quarterly geo-pause became standing operating procedure documented in a two-page program handbook, so the system survived contact with the brand’s own team rather than depending on ours.

Final Results at 90 Days

Metric	Before	After (90 days)	Change
Program ROAS (month-three run rate)	Not measurable	4.0x	New
Blended 90-day program ROAS	Not measurable	3.9x	New
Tracked creator-attributed revenue / month	~PKR 0.35M (estimated)	PKR 5.6M	≈16x
Creator share of online revenue	~2% (tracked)	24%	+22 pts
Cost per attributed order	~PKR 1,385 (estimated)	PKR 824	−41%
Whitelisted creator ads ROAS	2.2x (studio baseline)	3.4x	+55%
Active creator roster	6 flat-fee, untracked	63 onboarded, 11 driving 78%	New channel
Total online revenue / month	PKR 21M	PKR 23.2M	+10%

The arithmetic behind the headline: 1,700 attributed orders at a PKR 3,280 code AOV against PKR 1.40M of commissions, seeding, licensing, and amplification — cost that only existed because the orders did. These figures are illustrative outcomes built from common patterns WeProms sees across Pakistani D2C beauty and personal-care brands, not an audited third-party statement. They exist so a growth team can sanity-check what a commission-only creator program should plausibly return at this margin and order value.

What Made This Work

1. Pricing was the strategy. Commission-only terms aligned every rupee of creator cost with a confirmed order, which changed the program’s failure economics: a non-producing creator cost a seeding box and nothing else. The power law stopped being a risk and became the operating model.
2. The commission ceiling was derived, not negotiated. Building the 15% ceiling from contribution margin before recruitment ended the founder’s margin fear with arithmetic, and later allowed the 12%-versus-15% test to be run confidently — the higher rate won on data.
3. Attribution preceded amplification. Codes with dedup rules and survey triangulation were live before the first post, so pruning, tier changes, and licensing decisions were all made against tracked numbers. The two-truth baseline — survey says 9%, codes prove 2% — kept the program honest about what it could claim.
4. Paid amplified only proven organic winners. Whitelisting spent against assets that had already redeemed, which is why creator creative returned 3.4x in an account where studio creative had worn down to 2.2x.
5. The content met seasonal demand. Concern-led briefs aligned with the monsoon hair-fall corridor, so the roster posted into rising demand instead of manufacturing it from a standing start.

What Teams Can Apply

For Pakistani D2C brands weighing a creator affiliate program:

1. Derive the commission ceiling before you talk to creators. Contribution margin per order divided by average order value sets the maximum commission that keeps first orders profitable. Run that number first; it converts the margin debate into a spreadsheet row.
2. Recruit for trust density, not reach. A cohort of thirty nano and micro creators inside your buyer demographic will outperform one macro post at the same cost — and under commission-only terms, the ones who do not produce cost you a seeding box each.
3. Make attribution a discount code, not a link. In a market where buyers ask creators questions in DMs, screenshot recommendations to family groups, and check out cash-on-delivery, a code attributes cleanly everywhere a link gets stripped or ignored. Enforce one redemption per customer and rotate anything that leaks to aggregator sites.
4. Negotiate usage rights at onboarding. A 90-day whitelisting clause costs a sentence when nobody knows what the content will do, and a negotiation after it performs. It is the cheapest paid-creative arbitrage available to a D2C brand.
5. Verify incrementality before you believe the number. Pause codes in one city for two weeks and watch both redemptions and total revenue. If redemptions collapse while revenue holds, the program is creating demand; if the reverse, you are paying commission on orders you already had.

WeProms Digital has run this commission-first creator framework across Pakistani D2C brands in beauty, fashion, wellness, and food. The commission tiers, concern pillars, and amplification share shift with each catalog and margin profile — the sequence of economics, attribution, cohort, prune, and amplify is what stays constant.