Loyalty Referral Programme Case Study in Pakistan

The Client

An Islamabad-based D2C cosmetics and skincare brand formulating professional-grade makeup and skincare for South Asian skin tones — foundations in warm and olive undertones, long-wear liquid lipsticks, niacinamide and vitamin C serums, and a daily moisturizer-and-SPF range. The catalog sat at roughly 180 SKUs, average order value was about PKR 4,800, and monthly revenue had reached around PKR 12M. The brand sold direct through their Shopify storefront and through a network of about 80 partner salons and beauty parlors across Islamabad, Lahore, and Karachi that carried the retail range on their shelves and used the professional products in-service.

Acquisition was working. The team spent roughly PKR 1.8M a month on Meta, supplemented by a disciplined influencer programme that drove strong launch-day spikes around each shade drop and skincare release. First-purchase conversion was healthy, the product earned genuine affection, and the brand enjoyed the kind of unsolicited Instagram tags, WhatsApp forwards, and friend-to-friend recommendations that most beauty brands pay heavily to manufacture.

The problem was that almost none of that affinity was being captured or compounded. Buyers purchased once and were never given a structured reason to return, to refer, or to consolidate their beauty spend with the brand rather than dispersing it across competitors. And the salon network — 80 partners who physically held the product, applied it on clients, and recommended it daily — was generating referrals for free that no one tracked, rewarded, or scaled.

When they came to WeProms Digital, the brief was twofold: turn one-time buyers into returning members, and turn the brand’s already-live word of mouth — both customer and salon — into a measurable, rewarded acquisition channel. We scope this kind of work in our loyalty and referral programme setup service; the build below is how it was applied here.

The Problem

Five issues were suppressing retention and leaving referral revenue on the table:

1. Repeat purchase rate stuck at 23%. Beauty is a consumable, replenishable category — foundation runs out in a few months, serums and cleansers in a month or two, lipstick shades get repurchased seasonally. There was every natural reason for buyers to return, yet nothing in the experience prompted or rewarded it. No replenishment reminder, no “you’re running low” trigger, no incentive to come back to this brand rather than browse a competitor’s drop.

2. No loyalty programme at all. No points, no tiers, no perks, no reason to consolidate spend. A buyer who loved the foundation and would happily have bought their cleanser and moisturizer from the same brand instead spread those purchases across three competitors, because none of them — including this one — gave her a reason to concentrate.

3. Word of mouth was live but uncaptured. Customers were already recommending the brand — in WhatsApp groups, in Instagram comments, in salon chairs. But there was no referral mechanic to capture, attribute, or reward it. The brand was paying Meta roughly PKR 1,200 to acquire a customer that an existing buyer would have sent for a small fraction of that, if only the path existed.

4. The 80-salon partner network was untapped as a referral channel. Partner salons carried the retail range and recommended it to clients every day, but informally. There was no tracked referral code, no reward for the salon, and no way to know which salons drove which orders. The network was the brand’s highest-trust, highest-conversion acquisition surface, and it was operating entirely offline and unmeasured.

5. Acquisition cost climbing while LTV stayed flat. Customer acquisition cost had risen about 17% year on year — Meta’s increasing cost-per-thousand and category competition — while 12-month customer lifetime value had gone flat. Unit economics were slowly eroding, and the only durable fix was to raise the denominator: make existing customers worth more, and acquire new ones more cheaply through referral.

Compounding the above, the team had no cohort visibility into which buyer segments actually returned. They could see average repeat rate, but not how a skincare-only buyer behaved versus a makeup buyer, or a salon-referred buyer versus a Meta-acquired one. You cannot design a retention programme for segments you cannot see.

Phase 1 — Diagnosis and Data Foundation (Weeks 1-2)

A loyalty and referral programme fails when the underlying customer data is shallow. The first two weeks were about understanding who actually returns and why, then building the data foundation the programme would run on.

Cohort and LTV analysis. We pulled two years of order history and built repeat-behaviour cohorts by first-purchase category (skincare, makeup, mixed), by acquisition source (Meta, influencer, organic, salon), and by first-order AOV tier. The analysis surfaced what gut feeling could not: skincare-first buyers returned at nearly double the rate of makeup-first buyers (replenishment rhythm), salon-referred buyers had the highest 90-day repeat rate of any source, and the top 12% of spenders were quietly driving a disproportionate share of revenue. These three findings shaped the entire programme — tiers, the salon track, and VIP treatment all traced back to them.

Platform decision. We selected a loyalty platform that integrated cleanly with both Shopify and Klaviyo (which the brand already used for email), supporting points, tiers, a referral give-get, and a partner/affiliate-style track for salons. Integration health and catalog sync were verified before any programme logic was built.

Profile enrichment. As in any retention work, we enriched each customer profile with the properties the programme logic would need: primary category purchased, first-order tier, acquisition source (including salon-attributed), purchase recency, and predicted next-order window for replenishment. This is the same customer segmentation and personalization discipline that underpins all lifecycle work — without segmentable profiles, tiers and rewards fire blind.

Deliverability baseline. Because loyalty and referral communications would run heavily through email (points updates, tier-ups, referral invites), we authenticated the sending domain and confirmed inbox placement so programme messages actually landed.

Phase 1 results (by end of week 2):

Diagnostic	Before	After foundation
Repeat-behaviour cohorts	Not tracked	By category, source, AOV tier
Salon-attributed orders	Untracked	Source captured at order
Segmentable profile properties	3	8 (category, tier, source, recency, etc.)
Loyalty platform integration	None	Shopify + Klaviyo verified
Inbox placement (seed test)	86%	95%

Phase 2 — Build Loyalty and Referral Architecture (Weeks 3-6)

With the data foundation in place, we designed the programme around the two forces that actually drive repeat purchase in beauty: replenishment and recommendation. Every mechanic traced back to one of those two.

Tiered points loyalty. We launched a points-for-actions structure rather than points-for-purchases-only. Buyers earned points for purchasing (a fixed ratio per rupee), but also for behaviours that compound retention — writing a product review, creating an account, following on Instagram, and completing their profile with skin type and shade. The points were redeemable as rupee-off at checkout. Three tiers — Bronze, Silver, Gold — were tiered by 12-month spend, each unlocking incremental perks: free shipping at Silver, early access to shade drops and a birthday gift at Gold. Tier thresholds were set against the cohort data from Phase 1, so each tier was achievable but meaningful.

Referral give-get. The referral mechanic was the core acquisition lever. An existing member shared a unique referral link; the friend received PKR 500 off their first order over a floor value, and the referrer earned points (redeemable as rupee-off) once the friend’s order shipped. The give-get was deliberately asymmetric in currency — the friend gets a tangible rupee discount (which matters for first-purchase COD hesitation), while the referrer earns points that pull them back for their next purchase. This converts one acquisition into two retention events: the new customer, and the referrer returning to redeem.

Replenishment automation. Because skincare and foundation are predictable consumables, we built replenishment logic tied to the predicted next-order window from Phase 1. Roughly when a buyer’s serum or foundation was likely running low, they received a replenishment nudge — product-specific, with a points bonus for reordering within the window. This turned the category’s natural repurchase rhythm into a structured retention event rather than leaving it to chance.

Salon partner referral track. This was the most category-and-market-specific build, and it tied directly to the brand’s presence in beauty parlors. Each of the 80 partner salons received a unique tracked referral code. When a salon client bought online using that code, the order was attributed to the salon and the salon earned a commission or product credit. The salons received simple enablement — a digital kit with their code, shareable WhatsApp assets, and a quarterly statement of their referral earnings. This turned an offline, invisible recommendation flow into a tracked, rewarded, scalable acquisition channel run by the highest-trust advocates the brand had.

On-site and email integration. The loyalty panel was embedded on the storefront (points balance, tier status, rewards available, referral link) and connected to Klaviyo so that points updates, tier-ups, referral milestones, and replenishment nudges all flowed through email. A returning buyer saw their tier and points the moment they logged in — a constant, visible reason to consolidate spend.

Phase 2 results (by end of week 6):

Programme element	Status	Early signal (first 3 weeks live)
Points loyalty + 3 tiers	Live	Enrolment climbing post-purchase
Referral give-get	Live	First referred orders attributed
Replenishment automation	Live	Skincare reorder rate climbing
Salon partner track (80 codes)	Live	First salon-attributed orders tracked
On-site panel + Klaviyo sync	Live	Tier-up emails driving return visits

Phase 3 — Optimize, Segment, and Launch VIP (Weeks 5-10)

Coverage was in place; this phase was about performance, segmentation, and disproportionate investment in the best customers. As in the build phase, we began optimising each element the moment it had enough volume to read.

VIP treatment for Gold and the top decile. The Phase 1 cohort data showed the top 12% of spenders drove a disproportionate share of revenue. We built distinct VIP treatment for this group above and beyond Gold tier: a dedicated concierge WhatsApp line, early access to every shade drop 48 hours before public launch, a quarterly gratis product tied to their purchase history, and a higher-touch post-purchase experience. Concentrating effort where LTV already lives is where disproportionate retention gains come from — flat effort across all buyers produces flat results.

Tier-segmented campaigns. With tiers and segmentable profiles in place, every campaign now targeted a defined group. Bronze members received upgrade-to-Silver incentives. Silver members received category-expansion prompts (a skincare buyer introduced to the matching makeup range). Gold members received exclusive drops and replenishment. Broadcast volume dropped, but revenue per send rose sharply because the right products reached the right buyer at the right tier.

Referral reward testing. We tested the referral structure: friend discount depth (PKR 300 versus PKR 500 versus PKR 750), referrer reward type (points versus flat rupee credit), and the share-channel mix (WhatsApp versus Instagram versus email). Each test ran to a pre-set sample size before a winner was locked. The PKR 500 friend discount with points-for-referrer consistently won — meaningful enough to overcome first-purchase COD hesitation, without eroding margin. Referral conversion rate climbed steadily across the phase.

Replenishment tuning. We refined replenishment timing by product type — a 30-day window for cleansers and serums, 60-90 days for foundation and moisturizer — and added a points bonus that decayed if the buyer waited beyond the window. This sharpened the replenishment conversion rate and pulled reorder timing forward, which directly lifted the 90-day repeat figure.

Birthday and milestone rewards. Capturing birthday at profile completion let us trigger a birthday-month reward (a points bonus plus a small gratis gift for Gold members). These touchpoints cost little and generated some of the highest-engagement sends of the programme, because they felt personal rather than transactional.

Salon partner enablement and leaderboard. We provided partners with a simple quarterly leaderboard showing top-referring salons, and a tier of salon rewards for hitting referral thresholds. This introduced a light competitive dynamic that materially increased active participation among the partner network — salons that had recommended passively began recommending actively once the tracking and rewards made it worthwhile.

Phase 3 results (by end of week 10):

Metric	Start of phase	End of week 10
Member repeat purchase rate (90-day)	28%	36%
Referral revenue share	7%	12%
Programme enrolment (post-purchase)	31%	44%
Salon-attributed orders (monthly)	140	310
VIP tier revenue concentration	31%	36%

Phase 4 — Measure and Compound (Weeks 10-16)

The final phase turned the programme into something durable and measurable at the LTV level — the metric the brand actually cared about. Repeat rate and referral revenue had already moved substantially; the task now was to prove the LTV lift, lock in the iteration rhythm, and keep the programme from drifting into a discount treadmill.

Retention and referral dashboard. We built a dashboard tracking repeat purchase rate by cohort and tier, 12-month customer LTV by segment, referral revenue share and cost-per-acquired-customer via referral versus paid, salon-attributed order volume and revenue, points redemption rate, and net programme margin (loyalty revenue minus reward cost). This replaced gut-feel retention reporting with a defensible view of which tiers, segments, and partners drove the LTV lift. The 31% LTV figure and 14% referral share are what this dashboard made visible and trackable over time.

Reward-cost discipline. A loyalty programme can flatter revenue while quietly eroding margin through reward liability. We tracked reward cost as a share of loyalty revenue separately, so the brand saw net retention contribution, not just gross. Points expiry and a sustainable earn ratio were tuned to keep reward liability bounded while preserving the perceived value that drove repeat behaviour.

Iteration cadence. We established a monthly retention and referral review: repeat-rate movement by tier, referral revenue trend, salon partner activity, redemption rate, LTV trend, and one new test. This cadence is what kept LTV climbing past the build phase rather than settling — the programme iterated monthly against the dashboard rather than running static.

Referral as a durable acquisition channel. By the end of the phase, referral had matured from an experiment into a reliable, cheaper acquisition channel. Referred customers arrived with higher trust, converted at a higher rate, and — critically — themselves referred others at a higher rate than paid-acquired customers. This is the compounding effect: a programme that lowers CAC while raising LTV simultaneously, which is the rare change that improves unit economics in both directions.

By the 120-day mark, the programme had done what acquisition spending alone never could: member repeat purchase rate had climbed from 23% to 39%, referral revenue had grown from 3% to 14% of total, and 12-month customer lifetime value was up 31% — all while blended customer acquisition cost fell as referral took pressure off paid.

Final Results at 120 Days

Metric	Before	At 120 days	Change
Member repeat purchase rate (90-day)	23%	39%	+16 pts
Referral revenue share	3%	14%	+11 pts
12-month customer LTV	Baseline	+31%	Compounding
Programme enrolment (post-purchase)	None	47%	New channel
VIP tier revenue concentration	31%	38%	+7 pts
Salon-attributed orders (monthly)	Untracked	360+	New channel
Blended customer acquisition cost	Baseline	-14%	Improving

Each result traces to a specific phase: cohort analysis and profile enrichment in Phase 1 enabled the tiered loyalty, referral give-get, and salon track in Phase 2; VIP treatment and segmented campaigns in Phase 3 drove the repeat-rate acceleration; and the retention dashboard and iteration cadence in Phase 4 made the gains durable, measurable, and margin-aware.

What Made This Work

1. Built on the two real drivers of beauty repeat. Replenishment and recommendation are why beauty buyers return. Every mechanic — replenishment automation, referral give-get, points-for-reviews — traced back to one of those two forces. A loyalty programme that ignores the category’s actual repeat drivers is just a generic discount programme in disguise.

2. Captured word of mouth that was already happening. The brand’s customers were already recommending it. The referral give-get did not create advocacy — it captured, attributed, and rewarded advocacy that was leaking away unmeasured, converting it into the cheapest acquisition channel the brand had.

3. Turned the salon network into a tracked channel. The 80 partner salons were the brand’s highest-trust advocates, operating entirely offline. Giving them tracked codes, simple enablement, and rewards transformed a passive, invisible recommendation flow into a measured, scalable, rewarded acquisition surface — and one that converted far better than paid.

4. Concentrated VIP effort on the top decile. The cohort data showed the top 12% of spenders drove a disproportionate share of revenue. Treating them distinctly — early access, concierge, gratis product — is where the outsized LTV lift came from. Flat treatment across all buyers produces flat results.

5. Tracked net retention, not gross loyalty revenue. Building the dashboard that separated reward cost from loyalty revenue, and cohort LTV from vanity loyalty metrics, is what made the programme defensible. It let the brand see that the 31% LTV lift was real net contribution, not margin-eroding reward spend dressed up as retention.

What Teams Can Apply

For Pakistani beauty and personal-care brands that want retention and referral to compound:

1. Design loyalty around your category’s real repeat driver. Beauty is replenishable and recommendation-led; other categories are seasonal or occasion-led. Copy the logic of your category, not someone else’s playbook, and every reward should map to a behaviour you actually want to encourage.

2. Capture your existing word of mouth with a structured referral give-get. If your customers already recommend you, you are paying full acquisition cost for customers a friend would have sent for less. Build the path, attribute it, and reward both sides.

3. Treat your B2B partners as a referral channel. Salons, clinics, stylists, and retailers who carry or use your product are your highest-trust advocates. Give them tracked codes and rewards, and an offline recommendation flow becomes a measured acquisition channel.

4. Concentrate VIP treatment on the top decile. Not every buyer deserves equal effort. Spend your retention budget disproportionately on the customers who already drive a disproportionate share of revenue — that is where LTV lift concentrates.

5. Track cohort LTV and net retention, not gross loyalty revenue. Points issued and rewards redeemed are vanity unless you can see net contribution by cohort and tier. Build the dashboard that shows whether the programme is compounding the business or quietly eroding margin.

WeProms Digital has applied this loyalty and referral framework across Pakistani beauty, skincare, personal-care, and salon-retailed brands. The specific tiers, reward structures, and partner mechanics change with each catalog and channel mix — but the driver-aware, referral-capturing, partner-leveraging, LTV-tracked approach stays the same.