Case studies across digital commerce, omni-channel and data.

Challenge

Omni-channel programs were mature in North America but largely absent across Asia-Pacific, and the region had no dedicated owner. Digital revenue was targeted to grow ~40% a year, but the capability to enable it — store-fulfilled inventory, mobile payments, unified guest journeys — did not yet exist regionally.

Approach

• Built the region's first omni-channel roadmap, prioritising 50+ initiative candidates into a costed, multi-year investment case across twelve markets
• Delivered the marquee programs first: a mobile-POS rollout and a store-fulfilled shipping model, proving the model with revenue before scaling
• Established ways of working between the regional digital team and the global omni and retail-technology organisations

Results

Takeaway: A function that didn't exist became a delivering capability with a costed multi-year agenda — led by shipping revenue-generating programs, not by writing strategy decks.

Challenge

Several Southeast Asian markets sat below the revenue threshold for the group's standard commerce platform, and neither cross-border models nor dedicated e-commerce distribution centres were viable at their scale — leaving them with no direct e-commerce channel despite strong retail presence.

Approach

• Conceived a "small-market omni" model — brand.com storefronts fulfilled entirely from in-country retail-store inventory via flexible order routing, removing the need for a dedicated e-commerce DC
• Ran an end-to-end operating-partner RFP across four vendors, scored on ~1,300 requirement line items
• Executed per-country launches across ten functional workstreams each, with local payment methods and compliance

Results

Takeaway: A design constraint became the solution — store-fulfilled omni made direct e-commerce financially viable in markets a standard platform build could never justify.

Challenge

The direct-to-consumer channel existed to acquire high-value consumers early and maximise lifetime value, but its economics trailed other routes to market — and the business needed both a new-market entry and a path to channel-level profitability.

Approach

• Owned the D2C brand.com P&L line by line — from gross sales through to channel contribution — with driver-based revenue plans per market reconciled monthly against analytics actuals
• Authored market-level growth playbooks and orchestrated a new-country market entry end to end
• Built the CRM and telesales retention engine and merged consumer databases into a single addressable base

Results

Takeaway: A channel that had trailed on economics was turned profitable — with a new market launched and a merged first-party data asset to sustain it.

Challenge

Loyalty and CRM were fragmented across markets, with no consistent segmentation, no automated lifecycle journeys, and limited visibility into which members actually drove revenue.

Approach

• Relaunched loyalty as a tiered VIP club with digitised in-store and off-store member acquisition
• Stood up automated lifecycle journeys — welcome, voucher-reminder, cross-sell, birthday, renewal and win-back — on a marketing-automation platform
• Introduced value segmentation on spend and purchase latency, with differentiated propositions and reporting per segment

Results

Takeaway: A relaunched loyalty programme and automated CRM engine turned an anonymous customer base into a segmented, measurable revenue driver.

Challenge

Survey feedback arrived as hundreds of messy free-text responses that took hours to tag manually, produced overly broad buckets, and usually ended up on a slide rather than driving assortment decisions.

Approach

• Built a privacy-first analysis pipeline running locally — no customer data left the system
• Used AI to extract short product-intent phrases from open text, then auto-clustered them into actionable themes
• Compared themes against the current assortment to flag gaps, prioritised by frequency, seasonality and marketing potential

Results

Takeaway: Applied AI didn't replace strategy — it created the space to listen at scale, securely, and act on what customers actually asked for.

Challenge

Thousands of SKUs carried flat, rule-bound tags and vague, inconsistent descriptions with no keyword alignment — throttling search, merchandising and personalisation, and blocking organic-traffic growth.

Approach

• Built a local, API-powered pipeline that never sent data to a third-party platform
• Layered contextual, cultural tags on top of rule-based ones — dietary, cuisine, usage, festive and premium signals
• Generated structured, SEO-ready descriptions and metadata to a fixed, on-brand template

Results

Takeaway: Work that once needed a team of merchandisers and writers became one person, one machine and a controlled AI pipeline — with the data never leaving the system.

Challenge

Forecasting stretched last year's numbers forward with a rough promotional uplift — ignoring seasonal shifts, treating festive peaks as random spikes, and consuming hours in spreadsheets while still producing stock-outs and overstocks.

Approach

• Structured data at the monthly level with promotion and festive flags — light to manage, rich enough to capture demand shifts
• Used AI as a method coach and Python for local execution, keeping data private
• Fitted a seasonal model per SKU with separate festive and promotional adjustments

Results

Takeaway: Forecasting moved from guesswork to clarity by making it reflect real retail cycles — complex to simple, slow to fast, everyday to festive-ready.

Challenge

Following a major tax-regime change, the distribution network — seven plants and 50+ depots serving thousands of ship-to points — needed to be redesigned for cost and service over a five-year horizon.

Approach

• Built a greenfield network-optimisation model on specialist supply-chain software
• Disaggregated a five-year demand forecast down to city level and validated service requirements by zone and season
• Modelled depot networks with and without a super-stockist tier to test consolidation

Results

Takeaway: A leaner, optimally-located depot network cut logistics cost materially while holding service — turning a regulatory disruption into a structural efficiency gain.

Challenge

A multi-brand, multi-market rollout had to stand up brand.com, its enterprise integrations, an operating-partner path and a CRM backbone — all concurrently, against an aggressive timeline, with no existing template to follow.

Approach

• Authored the regional omni-channel blueprint and ran ~12 concurrent business-requirement workstreams
• Drove 20+ enterprise platform integrations — ERP, POS, CRM, loyalty, WMS, logistics and analytics
• Authored and ran a market operations-outsourcing RFP with blended-margin P&L vendor modelling

Results

Takeaway: A blueprint plus disciplined multi-workstream delivery took a from-scratch regional program to first-market go-live — the template for the markets that followed.

Challenge

The brand.com estate ran on an ageing stack that couldn't scale to the corporate roadmap, at a higher total cost of ownership than a modern cloud platform would carry.

Approach

• Ran early UX discovery into a full prototype-based style guide, then migrated product, customer and order data to the new platform
• Integrated the storefront with the ERP and CRM, building custom extensions where business rules fell outside the platform
• Templatised the theme so multiple brand sites could spin up from a common base

Results

Takeaway: A templatised re-platform turned a costly legacy estate into a scalable base the whole brand portfolio could build on.

Challenge

Large annual trade spend was being deployed without a clear view of which promotions were actually profitable — lift, cannibalisation and halo effects were unquantified.

Approach

• Audited and analysed promotions across category, channel, outlet type and promo type
• Built a hierarchical model to isolate true lift and ROI and quantify cannibalisation and halo effects
• Turned the coefficients into a what-if simulator for designing forward promo calendars

Results

Takeaway: Separating true lift from cannibalisation exposed profitable promotions and spend leakage — and left a reusable tool for planning the next calendar.

Challenge

A digital business over-dependent on a single channel needed to diversify, modernise an ageing platform stack, reposition a premium banner in decline, and move onto a board-grade planning and profitability footing.

Approach

• Diversified into a 10+ channel portfolio and re-platformed both banners onto a modern commerce stack
• Redesigned the customer-service operating model and built an in-house LLM analytics tool for natural-language business intelligence
• Built the bottom-up annual plan to channel contribution and presented it to the board

Results

Takeaway: A single-channel-dependent business became a diversified, profitably-growing portfolio — re-platformed, leaner in service, and running on a board-grade plan with contribution growing faster than revenue.