Klaviyo Email Marketing Case Study in Pakistan

The Client

A Lahore-based semi-fine jewelry ecommerce store selling gold-plated and sterling silver pieces — everyday rings, necklaces, earrings, and bridal sets — through their own Shopify storefront. Their average order value sat around PKR 18,000, monthly revenue was roughly PKR 16M, and they had built a clean, well-photographed brand that converted well on first visit. Acquisition was not their problem. Paid social on Instagram and Meta was bringing in first-time buyers at a respectable clip.

The problem was what happened after the first purchase. Most buyers vanished. The store’s repeat purchase rate — the share of first-time buyers who came back for a second order within 90 days — was stuck at 18%. For a category where customers build collections over time (a ring today, matching earrings next quarter, a gift for an anniversary), that number meant the brand was leaving most of its lifetime value on the table.

The store already paid for Klaviyo, the platform we specialise in for retention work — see our Klaviyo email marketing service — but they were using barely a fraction of it. Four basic flows were live and mostly unedited since launch. Every campaign went to the full list. There was no segmentation by jewelry type, no occasion logic around the wedding season and Eid peaks that dominate Pakistani jewelry demand, no post-purchase journey, and no attempt to bring lapsed buyers back. The team had the right tool and the wrong operating model.

When they came to WeProms Digital, the brief was specific: make retention work. Acquisition costs had climbed about 19% year on year, and 12-month customer lifetime value had gone flat. The only sustainable way to improve unit economics was to increase how often existing customers came back — and how much they spent when they did.

The Problem

Five issues were suppressing repeat purchases and LTV:

1. Klaviyo installed but dormant. The account existed, but only four default flows were active and none had been meaningfully edited. Welcome, abandoned cart, and two generic abandoned-checkout variants were sending unbranded, untimed messages. The platform’s real power — segmentation, conditional flow logic, predictive analytics — was completely unused.

2. No segmentation at all. Every campaign and flow treated a customer who bought a PKR 4,000 pair of studs identically to one who spent PKR 80,000 on a bridal set. There was no separation by category (rings, necklaces, earrings, bridal), by spend tier, by gifting versus self-purchase, or by where a customer was in their lifecycle. You cannot drive repeat purchases if you cannot tell your segments apart.

3. Jewelry treated as a one-off category. The entire programme assumed each customer buys once. There was no logic for the realities of jewelry: that customers build collections, that they buy for occasions, that they return for gifts, and that a bridal buyer is a different human from an everyday-wear buyer with completely different follow-up needs.

4. No occasion engine. Pakistani jewelry demand is sharply seasonal — wedding season from late autumn through early spring, Eid spikes, Valentine’s Day, and the December gift window. The store had no anniversary, birthday, or seasonal trigger flows to capture the natural rhythm of why people buy jewelry. Every one of those occasions is a repeat-purchase opportunity that was being missed entirely.

5. Cart abandonment poorly handled for a high-consideration category. Jewelry purchases involve deliberation — sizing, metal choice, COD trust, returns. The default cart flow was a single generic reminder that did nothing to answer the real questions that cause jewelry carts to stall. Recovery sat at just 6%.

The combined effect: email contributed only 9% of revenue, repeat purchase rate was 18%, and LTV was flat. The fix was not more subscribers or more campaigns — it was rebuilding how the store used Klaviyo to turn one-time buyers into returning collectors.

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

Retention work fails when the underlying data is shallow. The first two weeks were about making Klaviyo actually know who each customer is, so that every flow and segment we built later had real signal to act on.

Account audit and flow triage. We audited the live Klaviyo account: reviewed the four existing flows, the integration health, the profile properties being captured, and the campaign history. The existing flows were either redundant (two near-identical abandoned-checkout flows competing) or broken (a welcome flow sending to the wrong trigger). We paused the duplicates and flagged the rest for rebuild in Phase 2.

Profile enrichment. This was the highest-leverage work of the phase. By default, Klaviyo knew each customer’s email, order count, and total spend — and almost nothing else. We enriched every profile with the properties that drive jewelry segmentation: primary jewelry category purchased (rings, necklaces, earrings, bridal), first-order AOV tier, gifting signal (shipping address different from billing, or gift-note usage), and purchase recency. This turned a flat customer record into a segmentable one.

List hygiene and deliverability. We ran engagement-tier segmentation across the 27,000-subscriber list, sunset the long-inactive tail with a re-engagement sequence, and verified domain authentication (SPF, DKIM, DMARC). Inbox placement on monitored seed lists moved from 88% to 96% by the end of the fortnight. A clean, authenticated list is the prerequisite for any retention programme — segmentation is worthless if the messages never land.

Tracking verification. We confirmed the Klaviyo Shopify integration was capturing every relevant event — viewed product, started checkout, placed order, refunded — and that the catalog feed was syncing so that category-based flows and product recommendations would work. We also verified that historical order data had backfilled correctly, since predictive features depend on a complete order history.

Phase 1 results (by end of week 2):

Diagnostic	Before	After cleanup
Customer profile properties	3	8 (category, tier, gifting, recency, etc.)
Inbox placement (seed test)	88%	96%
Live flows (de-duplicated)	4 (2 redundant)	2 held for rebuild
Historical orders backfilled	Partial	Complete
Engagement-tier segmentation	None	Active / warming / cold defined

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

With rich profiles and a clean list, we rebuilt the flow architecture around how jewelry buyers actually behave. The goal was coverage of every repeat-purchase moment, driven by the enriched data rather than by generic timing.

Rebuilt welcome series (4 emails). The old welcome flow fired on the wrong trigger. We rebuilt it around the newsletter sign-up event with a first-order incentive delivered immediately, a craftsmanship-and-trust email (materials, plating longevity, COD and returns policy — the questions Pakistani jewelry buyers ask before paying PKR 18,000 online), a bestseller showcase segmented by the category the subscriber browsed, and a final urgency nudge. Trust-building content matters far more in high-consideration jewelry than in low-ticket categories, and the rebuilt series reflected that.

High-consideration cart abandonment (3 emails + 1 SMS). Jewelry carts stall for specific, answerable reasons. We rebuilt the cart flow as a sequence that addressed each: an immediate reminder, a next-day message answering sizing, metal, and COD questions, and a final incentive — with a consented SMS nudge for carts above a high-value threshold. This is the same abandoned cart and browse recovery discipline we apply across ecommerce, tuned for the longer jewelry decision window. Recovery moved from 6% to 11% within the first two weeks of the new flow going live.

Post-purchase care and cross-sell journey (3 emails). After a jewelry purchase, customers need care guidance (how to keep plating from tarnishing, how to store pieces) and a natural cross-sell moment. The post-purchase flow delivered care instructions specific to the product bought, a day-14 satisfaction check, and a day-30 introduction to a complementary piece — a ring buyer shown matching earrings, a necklace buyer shown the coordinating bracelet. This flow became the single largest driver of the second-purchase lift.

Occasion engine. This was the most category-specific build. We launched anniversary and birthday flows (capturing dates at sign-up and post-purchase), a wedding-season flow timed to the Pakistani wedding calendar, and seasonal Eid and gift-window flows. Each occasion flow was segmented by relationship to the brand — self-buyers versus gift-buyers received different creative — and carried curated, occasion-appropriate collections rather than generic catalogs.

VIP segmentation. Using the spend-tier property from Phase 1, we isolated the top 10% of customers and built a distinct VIP journey: early access to new collections, a dedicated VIP-only flow, and a higher-touch post-purchase experience. In a category where a small share of customers drive a large share of revenue, treating the top decile differently is where disproportionate LTV comes from.

Phase 2 results (by end of week 5):

Flow	Status	Early signal (first 2 weeks live)
Welcome (rebuilt)	Live	First-purchase conversion up, refund rate down on trust content
Cart abandonment (rebuilt)	Live	Recovery 6% to 11%
Post-purchase care and cross-sell	Live	Second-purchase rate climbing
Occasion flows (4)	Live	First anniversary and birthday redemptions
VIP journey	Live	Top-decile repeat rate materially higher

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

Coverage was in place; this phase was about performance and the predictive logic that turns a good retention programme into a compounding one. As in the build phase, we began optimising each flow the moment it had enough volume to read.

Predictive next-order timing. Klaviyo’s predictive analytics estimate when each customer is likely to order next based on their history. We activated this to drive a “we thought you’d like to see” flow that reached customers at their predicted reorder window with a curated, category-appropriate selection, rather than on a fixed 30-day cadence. For a collectible category like jewelry, timing the nudge to when a buyer is actually ready beats blanket cadence.

Collection-completion logic. Jewelry buyers often intend to build a set over time. We built flows that recognised when a customer owned one piece from a collection and introduced the coordinating pieces — effectively a replenishment equivalent for a non-consumable category. This widened the number of categories each customer bought from and lifted average repeat-purchase order value.

Category- and tier-based segmentation across campaigns. With the Phase 1 properties in place, every campaign now targeted a defined segment: bridal buyers, everyday-wear buyers, gift-buyers, VIPs, lapsed high-spenders. Broadcast volume dropped, but revenue per send rose sharply because the right products reached the right buyers.

Structured A/B testing. We tested the elements that move repeat revenue: subject lines on the post-purchase cross-sell, discount depth in cart recovery, send timing on the predictive flow, and which coordinating piece to feature in collection-completion. Each test ran to a pre-set sample size before a winner was locked. This lifted cart recovery from 11% to 14% and sharpened the cross-sell conversion rate over the phase.

Win-back with tier awareness. Lapsed buyers entered a win-back sequence, but the incentive and tone varied by their original spend tier — a former VIP received a different, higher-touch win-back than a single low-AOV buyer. Tier-aware win-back recovered lapsed buyers at 13% versus the 5% baseline, and it brought back the higher-value lapsed customers first, which is where the LTV recovery concentrated.

Phase 3 results (by end of week 8):

Metric	Start of phase	End of week 8
Repeat purchase rate (90-day)	22%	28%
Cart-abandonment recovery	11%	14%
Win-back reactivation	8%	13%
Klaviyo revenue share	14%	21%
Email and SMS revenue (monthly)	PKR 2.3M	PKR 3.4M

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

The final phase turned the rebuilt retention engine into something durable and measurable at the LTV level — the metric the store actually cared about. Repeat rate had already moved substantially; the task now was to prove the LTV lift, lock in the iteration rhythm, and protect the gains.

LTV and retention dashboard. We built a dashboard that tracked repeat purchase rate by cohort, 12-month customer LTV by segment, flow-attributed revenue, and segment-level return. This replaced gut-feel retention reporting with a defensible view of which segments and flows were driving the LTV lift. The 34% LTV figure is what this dashboard made visible and trackable over time.

Cohort visibility. By tracking repeat behaviour by first-purchase cohort, the store could see that bridal buyers and everyday-wear buyers had very different return curves — and allocate creative and flow effort accordingly. This kind of cohort insight is what separates a retention programme from a set of flows.

Iteration cadence. We established a monthly retention review: repeat-rate movement by segment, LTV trend, flow performance against target, and one new test. This cadence is what kept LTV climbing past the build phase rather than settling.

Loyalty integration. We connected the flow logic to a lightweight loyalty structure so that VIP and post-purchase flows referenced tier status and rewards, giving customers a reason to consolidate their jewelry spend with the store rather than dispersing it across competitors.

By the 120-day mark, the programme had done what acquisition spending alone never could: repeat purchase rate had climbed from 18% to 31%, 12-month customer LTV was up 34%, and Klaviyo — which had been contributing 9% of revenue — now accounted for 24% of monthly revenue through flows and segmented campaigns.

Final Results at 120 Days

Metric	Before	At 120 days	Change
Repeat purchase rate (90-day)	18%	31%	+13 pts
12-month customer LTV	Baseline	+34%	Compounding
Klaviyo revenue share	9%	24%	+15 pts
Email and SMS revenue (monthly)	PKR 1.4M	PKR 3.8M	+171%
Cart-abandonment recovery	6%	14%	+8 pts
Win-back reactivation	5%	13%	+8 pts
Post-purchase cross-sell conversion	Not tracked	9% of buyers	New channel
Inbox placement (seed test)	88%	97%	+9 pts

Each result traces to a specific phase: deliverability and profile enrichment in Phase 1 enabled the rebuilt flows in Phase 2, the predictive and win-back logic in Phase 3 drove the repeat-rate acceleration, and the LTV dashboard and iteration cadence in Phase 4 made the gains durable and measurable.

What Made This Work

1. Profile data before flows. The decisive move was enriching each customer profile with category, spend tier, gifting signal, and recency before building anything. Segmentation is the entire game in retention, and segmentation is impossible without those properties. Flows built on flat data would have produced flat results.

2. Jewelry treated as collectible, not one-off. The whole programme was rebuilt around the reality that jewelry buyers return — for collections, for gifts, for occasions. Collection-completion flows, occasion logic, and predictive next-order timing all flowed from that single framing shift.

3. Occasion engine matched Pakistani demand rhythm. Wedding season, Eid, anniversaries, and the gift window are when Pakistani jewelry buyers open their wallets. Building flows around that calendar captured repeat purchases that a generic monthly cadence would always miss.

4. Tier-aware treatment of the best customers. The top decile behaves differently and should be treated differently — in post-purchase, in VIP journeys, and in win-back. Concentrating effort where LTV already lives is what produced the outsized LTV lift.

5. LTV as the headline, not revenue share. Revenue share is a vanity number without the cohort and LTV tracking behind it. Building the dashboard that proved the 34% LTV lift is what made the programme defensible and let the store shift acquisition budget with confidence.

What Teams Can Apply

For Pakistani ecommerce brands in high-consideration categories that want retention to compound:

1. Enrich your customer profiles before you build flows. If you cannot segment by category, spend tier, and behaviour, your flows are operating blind. This is the single most underused lever in Klaviyo.

2. Build around your category’s real purchase rhythm. Jewelry is collectible and occasion-driven; other categories are consumable or seasonal. Copy the logic of your category, not someone else’s playbook.

3. Capture and trigger on occasions. Anniversaries, birthdays, and seasonal peaks are free repeat-purchase opportunities. Most stores never ask for the dates and never send the trigger.

4. Make win-back tier-aware. Not every lapsed customer deserves the same effort. Spend your win-back budget reclaiming the high-value lapsed buyers first.

5. Track LTV by cohort, not just revenue. Revenue share tells you email is working; cohort LTV tells you the business is compounding. Build the dashboard that shows the difference.

WeProms Digital has applied this Klaviyo retention framework across Pakistani ecommerce brands in jewelry, fashion, watches, and accessories — you can see more of the vertical context in our work for digital marketing for jewelry brands. The specific segments, occasions, and flow logic change with each catalog, but the data-first, category-aware, LTV-tracked approach stays the same.