Loyalty Programme Case Study Pakistan

The Client

A Lahore-based D2C specialty coffee roaster and roasted-snacks brand, founded in 2022, selling freshly roasted coffee beans, ground coffee, nut butters, roasted almonds and cashews, and trail mixes through its own Shopify storefront, a nascent subscription option, and a small network of retail stockists in Lahore and Islamabad. The founders had built genuine brand affection — a loyal core reordered every two to three weeks — but the business around that core was fragile. Average order value sat near PKR 2,400, monthly orders hovered around 3,200, and the roaster was a textbook recurrence category: coffee runs out, snacks get finished, and the natural repurchase cycle is two to four weeks.

The brand had grown on Meta paid social, where discount-coded acquisition campaigns delivered cheap first orders. The trouble was what happened after the first order, which was usually nothing. They engaged WeProms because repeat purchase had stalled, CAC was creeping up as Meta costs rose across Pakistan, and a referral line printed on every coffee bag (“tell a friend, get a reward”) had never been tracked. This page is an illustrative composite built from the patterns WeProms sees across Pakistani D2C food brands — anonymized and aggregated, not a report on a single named client.

The market context made retention the right lever rather than acquisition. Specialty coffee and premium snacking had expanded quickly in Lahore and Islamabad across the prior three years, with a growing base of urban buyers who would reorder when the beans ran out, but the category had also commoditized at the low end as cheaper roasted-coffee sellers entered on price. Discount-led Meta acquisition was still affordable per first order, but every one-time discount shopper who failed to return represented acquisition spend that never amortized across a second purchase. The economics only worked if the brand converted a meaningful share of those first orders into a second one inside the natural repurchase window — which is precisely what the existing setup was structured not to do.

The Problem

Acquisition was working; retention was not. A cohort analysis built from Shopify order data and Klaviyo profiles produced a clear set of blockers:

• Repeat rate well below category potential. The 90-day repeat purchase rate sat at 18%, against a realistic coffee-category ceiling near 30% or higher. Most buyers made one order and disappeared.
• Revenue concentrated in one-time discount shoppers. 62% of revenue came from customers acquired through Meta discount codes who never returned at full price. They had been trained to wait for the next coupon.
• No retention infrastructure. There was no loyalty programme, no member accounts, no points balance, no rewards. The store treated every visitor as a stranger on every visit.
• A flat, unsegmented email list. The 14,000-contact list received the same weekly newsletter regardless of purchase history. No tier-based, behaviour-based or win-back flows existed.
• Untracked referral. Packaging mentioned a referral reward, but no codes, links or attribution existed. The brand had no idea whether word-of-mouth was working or how much it contributed.
• Subscription underused. A subscription option was live on the store but only 4% of customers had adopted it, partly because subscribing carried no advantage over one-time purchase.

The strategic question was not whether to add points — it was how to convert a discount-dependent acquisition engine into a margin-healthy retention engine without collapsing short-term revenue.

Phase 1 — Diagnosis and Cleanup (Weeks 1-2)

We opened with a cohort and unit-economics audit rather than a tool selection exercise, because the right programme design depends entirely on how the customer base actually behaves. Using Shopify order history and Klaviyo profiles, we segmented customers by order count, recency and discount sensitivity, and measured three things: the 90-day repeat rate by acquisition cohort, the time-to-second-purchase distribution, and the discount-rate sensitivity of repeat buyers.

The most important finding was the repurchase window. Among customers who did buy a second time, roughly 70% did so within 45 days of their first order — and if they had not returned by day 45, the probability of a second purchase fell off a cliff. That 45-day window became the spine of the entire programme design: every flow, nudge and tier threshold would be built to act inside it.

Cohort insight	Finding	Implication
90-day repeat rate	18% overall	Large headroom to category ceiling
Time-to-second-purchase	70% within 45 days	Intervene hard inside the 45-day window
Discount-driven share	62% of revenue	Wean off coupons via tier perks
Subscription adoption	4% of customers	Tie subscription into loyalty benefits

Cleanup ran alongside the analysis. We deduplicated the email list, suppressed the segment of serial coupon-only buyers from discount-led campaigns, and selected a loyalty platform that integrated natively with both Shopify and Klaviyo so that points, tier and referral events could trigger lifecycle messages without a custom integration project. We also confirmed the unit economics: at a PKR 2,400 AOV and healthy gross margins typical of roasted coffee and snacks, the brand could fund a meaningful points earn rate and a double-sided referral reward without eroding margin.

Phase 2 — Build and Restructure (Weeks 3-5)

Phase 2 designed and built the programme. The core decision was tiering rather than a flat points programme, because coffee and snacks are a status-adjacent, identity-driven category where progression and recognition pull harder than a straight discount.

Tier design

We built three tiers, named in the brand’s own language, each tied to a rolling 12-month spend threshold:

Tier	Threshold (12-month spend)	Earn rate	Signature perks
Sipper (entry)	PKR 0	1 point per PKR 20	Welcome reward, birthday bonus
Regular	PKR 30,000+	1.25 points per PKR 20	Free shipping, early access to new roasts
Reserve (top ~10%)	PKR 75,000+	1.5 points per PKR 20	Concierge WhatsApp line, limited-edition micro-lots, priority dispatch

The deliberate mechanics were status (the Reserve tier was invitation-feeling, capped near the top 10% of customers), early access (new roasts released to Regular and Reserve 48 hours before everyone else), and a concierge WhatsApp line for the top tier — a perk that costs almost nothing but reads as premium in a market where WhatsApp is the default communication channel.

Points, redemption and referral

Customers earned one point per PKR 20 spent, with bonus points for account creation, leaving a verified review, following on Instagram and completing a purchase within the 45-day window. Redemption was kept simple: PKR 100 off at 200 points, a free 250g bag at 500 points, and rotating seasonal rewards. We introduced a 12-month points-expiry rule to create urgency — points that do not expire do not drive redemption, and unredeemed points are a liability on the balance sheet, not a marketing asset.

The referral mechanic was rebuilt as a tracked, double-sided reward: the referring customer received PKR 300 store credit when their friend placed a first order, and the friend received PKR 300 off their first order. Each customer received a personal referral link from their account page. We sized the reward deliberately — PKR 300 each side was large enough to motivate sharing but small enough to preserve margin at the brand’s AOV.

Lifecycle integration

The programme was wired into Klaviyo so that loyalty events triggered lifecycle messages automatically: a welcome email with the first-points earn, a tier-up celebration with the unlocked perks, a “you’re PKR X away from the next tier” nudge sent when a customer crossed 80% of a threshold, a points-expiry reminder 30 days before expiry, and a win-back flow for customers who crossed day 40 without a second purchase. Member accounts and the points balance were surfaced on the storefront so customers could see their tier and progress on every visit.

Phase 3 — Optimize and Scale (Weeks 4-8)

We launched the programme with a coordinated campaign across email, WhatsApp and the storefront, presenting the loyalty programme to the existing 14,000-contact list as a re-engagement event rather than a new feature. Existing customers were back-awarded points for past purchases up to a capped lookback window, which immediately placed a meaningful slice of the base into the Regular tier and gave them a reason to return now rather than later. The launch sequence itself ran as a three-touch flow: an announcement email explaining the tiers and the back-awarded points balance, a WhatsApp message 48 hours later to the segment that had not opened the email (which in Pakistan reliably outperformed the email on click-through), and a final reminder timed to a payday window when discretionary spend peaks. A storefront banner and a post-purchase points-earned notification on the order confirmation page closed the loop for buyers who arrived mid-campaign. The back-award mechanic was the single highest-converting element of the launch — telling a lapsed customer they already had PKR 200 in points waiting produced an immediate reorder spike that no discount-led campaign could have matched at that margin.

Three optimization loops ran through Phase 3. First, the tier-progression nudges: customers who crossed 80% of the Regular threshold received a targeted message, and a measurable bump in second purchases followed each send. Second, points expiry: the first expiry reminders drove a redemption spike as customers returned to spend points they would otherwise lose, converting stored liability into revenue. Third, referral tuning: we tested the reward amount and found the PKR 300 double-sided reward outperformed both a smaller reward and a single-sided alternative, because the friend-side discount was what actually closed the referred first order.

Programme metric	Pre-launch	Week 8
Member accounts created	0	8,900
Repeat customers in active tiers	—	41% of base
Referral links shared	untracked	2,100
Points redemption rate	n/a	23% of earned points

Subscription and loyalty were deliberately linked. Subscribers earned points automatically on each renewal and reached the higher tiers faster, which reframed subscription from “the same product, delivered” into “the same product, delivered, with status” — and lifted subscription adoption as a side effect of the loyalty programme rather than a separate project.

Phase 4 — Measure and Compound (Weeks 8-12)

Phase 4 turned the launch into a durable system by institutionalizing the cohort measurement that drove the design. We re-ran the 90-day repeat-rate cohort for customers who had experienced at least 90 days of the programme, compared it to the pre-launch baseline, and built a live dashboard tracking repeat rate, LTV by cohort, tier distribution, referral attribution and email-and-WhatsApp revenue share.

The compounding mechanics did the rest. Each new repeat customer entered a tier, earned points, and became more likely to return again — the classic loyalty flywheel. As the Reserve tier grew, its members began to self-identify in support conversations and on social media, which in turn generated organic word-of-mouth on top of the tracked referral flow. By the end of the quarter the programme was no longer a project; it was infrastructure that touched acquisition (referral), retention (tiers and points) and revenue share (lifecycle messages). We set a monthly governance cadence with the brand to review tier distribution for unhealthy concentration, check that referral rewards were still margin-positive at the current AOV, and retire any flow whose engagement had decayed below threshold — the discipline that keeps a loyalty programme compounding rather than coasting.

Final Results

Measured across the 90-day cohort following launch, with LTV compounding tracked over six months:

Metric	Before	After	Change
90-day repeat purchase rate	18%	34%	+16 points
Repeat-customer LTV	baseline	+41%	—
Referral share of new customers	3%	11%	+8 points
Subscription adoption	4%	13%	+9 points
Email and WhatsApp revenue share	9%	24%	+15 points
Discount-driven revenue share	62%	38%	-24 points

Revenue context. The shift from 62% discount-driven revenue to 38% is arguably the most important number on the page, because it represents revenue moving from a low-margin, coupon-dependent channel into a higher-margin, relationship-driven one. Combined with a 41% lift in repeat-customer LTV and subscription adoption more than tripling, the programme changed the unit economics of the business. These figures are illustrative — a buyer should treat them as a realistic outcome shape to sanity-check fit, not an audited third-party result.

What Made This Work

1. Cohort diagnosis before programme design. The 45-day repurchase window was discovered in Phase 1 and shaped every decision after. A loyalty programme designed without that insight would have set tier thresholds and flow timings against the wrong clock.

2. Tiers over flat points. In an identity-driven category, progression and recognition pulled harder than discounts alone. The Reserve tier’s concierge WhatsApp line and early access to micro-lots cost almost nothing and drove disproportionate engagement from the top 10%.

3. Points expiry as urgency. Expiring points in 12 months converted stored liability into redemption-driven revenue and gave the expiry-reminder flow real pulling power. Perpetual points would have underperformed.

4. Double-sided referral. The friend-side discount was the lever. Referral only compounds when the referred friend actually converts, and the matched PKR 300 offer closed that first order reliably.

5. Lifecycle integration. Wiring loyalty events into Klaviyo — tier-ups, threshold nudges, expiry reminders, win-backs — turned a static programme into an active system that reached customers inside the 45-day window.

6. Subscription folded into loyalty. Linking subscription to faster tier progression reframed subscription as a status benefit and lifted adoption without a separate subscription-marketing effort.

What Teams Can Apply

For Pakistani D2C food and beverage brands:

• Find your repurchase window before you build. Run a cohort analysis on time-to-second-purchase and design every tier threshold, nudge and expiry around that window. A programme built on the wrong clock under-delivers no matter how good the rewards are.
• Choose tiers for status categories, flat points for commodity ones. Coffee, snacks, beauty and supplements reward tiering because buyers identify with the brand. Commodity categories usually do better with simple points and frictionless redemption.
• Build referral into the post-purchase moment, not just the packaging. A line on the bag is untracked and unoptimized. A personal referral link surfaced right after delivery, with a double-sided reward, turns word-of-mouth into an attributed acquisition channel.
• Integrate loyalty data into your email and WhatsApp lifecycle. Tier-ups, threshold nudges and points-expiry reminders are where most of the repeat-rate lift actually comes from. The programme is the data layer; the flows are the engine.
• Reward subscription inside loyalty. Let subscribers earn faster and reach higher tiers sooner. Subscription and loyalty solve the same problem — repeat purchase — and should reinforce each other rather than compete for the same budget.

For the underlying service, our loyalty and referral programme setup covers the full scope, the D2C food brand industry hub frames the demand patterns this work plugs into, and if churn is the sharper problem, subscription retention and churn management is the adjacent lever to pull alongside loyalty.