Head of Product · Liminal · Temasek
Turning affiliate
commerce into an incentive system.
Growth driven by embedded product loops — cashback, referrals, and tiering — validated through early guerrilla acquisition and scaled via organic SEO and paid at <$2 CAC.
01 — The problem
Incumbents are hard to beat.
Affiliate commerce is fragmented — different networks, different rates, opaque tracking, and rewards that decay, reset, or go unclaimed. Platforms like ShopBack have proven users will change shopping behaviour for better rewards. But they've also built strong distribution and habit, making switching behaviour hard even when a better system exists.
The risk wasn't product. The risk was whether users would actually switch away from what they already knew.
Key risks we had to validate
Will users switch from established incumbents if rewards are meaningfully better?
Do crypto-native rewards create enough incremental value over traditional cashback?
Will affiliate platforms eventually be disrupted by AI agents — cutting out the user-facing layer entirely?
02 — The insight
Affiliate routing is a yield optimisation problem.
Instead of static cashback, transactions can be dynamically routed across affiliate networks to maximise returns — then converted into programmable, on-chain rewards. The user gets more. The system gets smarter. And unlike points that expire, on-chain crypto rewards — USDC, BTC, and partner tokens — are real, transferable value.
This reframing changes the product entirely. Laguna isn't a cashback app with a crypto layer on top. It's a transaction routing system with a user-facing rewards interface — built on affiliate infrastructure, with on-chain distribution at the back end.
GMV → affiliate commission → on-chain crypto rewards. Multi-token, chain-agnostic distribution.
03 — Product & growth system design
Growth embedded in product mechanics.
Laguna was designed so that every transaction simultaneously generates revenue and drives growth. The core challenge wasn't building a better product — it was changing user behaviour away from incumbents. The answer was to make switching economically rational and compound that advantage over time. This allowed Laguna to scale without linear increases in acquisition spend.
Loop 1 — Cashback → Repeat usage
Every purchase returns real, liquid value — USDC, BTC, or partner tokens — not points that expire. This anchors retention to financial benefit rather than UI or brand loyalty, reinforcing habit with each transaction, increasing purchase frequency (35% of users make a repeat purchase within 30 days), and reducing reliance on paid reacquisition.
Loop 2 — Referral → Earnings amplification
Users earn a percentage of their referees' cashback indefinitely — not a one-time bonus. This converts referrals from a single acquisition event into a compounding revenue stream, increasing LTV per acquired user (referrals contribute ~27% of new users), and shifts CAC from an upfront cost to a revenue share model — you only pay when users generate value.
Loop 3 — Tiering → Behavioural reinforcement
Cashback upsize rates increase with activity tier (Bronze → Silver → Gold). Higher spend unlocks better rates, which incentivises higher spend — a self-reinforcing loop that aligns user behaviour with platform monetisation and increasing user lock-in over time (~2–4% of users reach Gold tier, contributing ~8–10% of activity — Gold users generate ~2–3× more value than baseline).
Unit economics
LTV:CAC ratio of 1.5–2.0× within the first year — a direct result of compounding product loops reducing reacquisition cost while increasing per-user value over time.
These loops mean that at scale, growth, retention, and monetisation compound together — eliminating the need for independent spend on acquisition, retention, and monetisation.
04 — Growth execution & validation
Guerrilla first. Paid last.
While Laguna's long-term growth is driven by embedded product loops, early traction required validating user switching behaviour through scrappy acquisition. The core question: would users actually leave incumbents if the rewards were meaningfully better?
Phase 1
Guerrilla
Posted in forums and communities. Watched for unprompted mentions of Laguna. Went directly into crypto chats to observe organic reactions and test whether the value proposition landed.
Phase 2
Community-led
Activated Web3 communities directly — drawing on my experience at Seedly helping MCC cards like Revolut build local community traction. Same playbook, different vertical.
Phase 3
Organic
Shifted to reviews and SEO optimisation once we had proof of switching behaviour. Built content loops that created sustainable, compounding user acquisition without paid spend.
Phase 4
Paid at scale
Once the funnel was proven, moved to paid acquisition and optimised CAC to under $2 — well below the LTV of an active shopper. Data-driven iteration on channels and creative.
$2
CAC at scale
53%
Peak shopper CVR
~20%
M6 retention
$233
Peak ARPPU
04 — The system
Built as a full-stack system, not a feature set.
Laguna was sequenced deliberately — starting with the simplest cashback flow, then layering in wallets, on-chain rewards, tokenomics, and analytics. Each layer made the system harder to replicate and more valuable to the user.
Merchant marketplace
100+ global merchant integrations across travel (SIA, Emirates, Agoda, Trip.com), retail (Nike, iHerb, Shopee, Lazada, Taobao), and more. Managed affiliate tracking, order reconciliation, and payout accuracy end-to-end.
Admin portal
Internal operations dashboard for managing merchants, tracking affiliate payouts, and monitoring system health. Over 50 Tableau dashboards tracking LTV, CAC, retention cohorts, merchant performance, and token distribution.
Multi-token wallet & rewards engine
Web2-friendly wallet supporting USDC, BTC, ETH, ANIME, MOCA, APE and more. Chain-agnostic reward distribution — GMV converts to affiliate commission, then to on-chain crypto payouts. Miniapps deployed on Solana and Base.
Analytics stack
50+ Tableau dashboards tracking LTV, CAC, retention cohorts, persona segmentation, merchant performance, conversion funnels, token distribution, and retention curves. Data-driven iteration on every growth lever.
05 — Partners & ecosystem
Partners & ecosystem
Integrated Laguna into 10+ Web3 ecosystems as distribution and credibility channels. Managed 100+ merchant partnerships end-to-end — deal negotiation, integration, tracking, and payout reconciliation.
Web3 ecosystem partners
Merchant partners
06 — What this becomes
The next risk: agents.
AI agents acting on behalf of users will eventually handle discovery, comparison, and purchase — cutting out the user-facing affiliate layer entirely. This is the existential risk to platforms like ShopBack, and to Laguna as it currently exists.
The de-risking strategy: evolve the routing and incentive infrastructure into the execution layer for agent-driven commerce. The same system that optimises cashback for a human today can be used by an agent to transact automatically tomorrow.
Routing and incentives stop being UX features. They become core infrastructure for intent-based transactions.
The evolution
User browses → Laguna routes transaction → USDC rewards distributed on-chain
Agent detects intent → Laguna routing executes transaction → rewards distributed without user input