01/Case
What the points actually return
Your ad asks one person to carry commercial acumen and loyalty value in the same sentence as adoption, engagement and NPS. That combination is what a client asks for once a programme is live and the business case has gone quiet.
A client says: our programme has plenty of members and nobody can tell us what the points are buying.
Lulu Retail, Abu Dhabi headquartered and listed on ADX, published these in its FY2025 results on 13 February 2026. The first two are what every GCC loyalty deck opens with and both measure identification rather than behaviour, and the third is what makes the reward budget matter.
The reward budget, mechanic by mechanic
Worked example. Illustrative figures except where marked published.
Two thirds of $7.9bn puts roughly $5.3bn of sales on cards. A grocery programme earning around 0.9% of member spend issues about $47.5m of rewards a year, which is a quarter of the published net profit. Split three ways, net of the margin each mechanic genuinely brings in at a 22% gross margin:
Net contribution per mechanic, millions of dollars a year. Left of the line loses money, right of it makes money.
Base earn, every basketspend $28.5m · margin $9.4m
−19.1
Category boosts, own brandspend $11.9m · margin $21.3m
+9.4
Lapsed win-back offersspend $7.1m · margin $18.4m
+11.3
Platform, CRM and opsfixed, runs either way
−6.2
With base earn unfunded
+$14.5m a year of contribution, against −$4.6m as the programme runs today
/Fund the change, not the deletion
Half the budget is paying for baskets that were already coming, and that is where the $14.5m sits. I would halve the earn rate rather than remove it. A rate that goes to zero reads to 8.4 million members as the programme being cancelled. Halve it, move the money behind boosts and win-back, and hold back a random slice of members at the old rate so the result is measured rather than argued.
The app rebuild is the first thing I would refuse
It is the most sellable opening project and the wrong one. Nothing in that ledger is negative because of the interface. It is negative because of the earn rate and because no holdout exists, and a design sprint in front of an eight week measurement buys a nicer screen on top of the same arithmetic. The app work comes after there is a number worth designing around.
What ships first, and how it gets judged
The earn-rate change, behind a holdout. Base earn halves for a randomly assigned 90% of members while 10% keep the current rate. Assignment stable per member, one points balance visible in the app, reversible in a single release. Read at week eight on margin per active member, and if the treated group is not at least flat, the rate goes back and the boosts case gets rebuilt.
Win-back targeting, which is where AI belongs here. Lapse-risk scoring on transaction history decides who gets an offer and how large it is. Not a chatbot in the app. Sia became an OpenAI Advanced Partner on 23 July 2026, and this is a scoring and orchestration problem of exactly that kind.
Enrolment at the store openings, for acquisition. Fifty GCC stores are planned for 2026 to 2028 (published). Opening week is the cheapest enrolment moment a grocer ever gets, and it happens at the till, which is where your own Deep Review solution already points. One owned metric: enrolments per opening and their 90 day retained rate, owned by product, not by the store.
The honest limits of this. It is an outside-in read on published figures. The real earn rate, breakage, redemption mix and margin by category are internal, and the three way split of the budget is my assumption. Two weeks sitting with a client's CRM and finance people, which is what Forward Deployed Success is for, would replace my numbers with theirs, and the boosts line is the one most likely to move. Adoption, loyalty participation, retention and margin per active member are the KPIs I would own; NPS I would carry as a guardrail, not a target.
02/Fit
Three things I would bring, and the one I would not pretend about
A consumer digital product with the commercial number attached to it
Your ad puts commercial acumen next to product vision. At Ifolor I owned the e-commerce ecosystem and strategy for a CHF 100M+ business and reported to C-level, which meant every roadmap call was also a margin call. Research, A/B testing and analytics, on Magento and GA4, across web and app.
Ifolor Group, Senior Product Manager and Lead E-Commerce, Oct 2024 to Jul 2025: +9% conversion, +15% checkout step rate, cross-functional team plus external agencies, own budget and KPIs.
The commercial half of loyalty, which is where these programmes fail
Points are trade promotion with a login. My first years were spent planning consumer and trade promotions, building the calculations behind them and negotiating with brand manufacturers, which is the same question the ledger above asks. At Swiss Post I size AI-driven business opportunities and turn them into a prioritised roadmap with KPIs, including the build-versus-buy and cost-versus-benefit calls.
Promena, Cruspi, Domaco, 2010 to 2016: own brand portfolio, promotion planning, pricing calculations, monthly sales reporting. Swiss Post, AI Project Lead since Jan 2026: sales time down 30%, proposal feedback time down 90% via a synthetic pitch-check avatar I built.
The honest read: no loyalty programme has my name on it
You ask for demonstrated Loyalty Programs experience and name Salesforce Loyalty Cloud as an advantage. I have neither. What I have is the product and commercial spine that sits underneath one: retention and conversion work on a real e-commerce P&L, the promotion economics I learned on the FMCG side, and client-facing delivery with senior stakeholders at UBS and Baloise while at Brixel. If the shortlist needs someone who has already run a rewards ecosystem end to end, I am not that person. If it needs someone who can tell a client what their programme returns and then ship the change, I would like the conversation.
03/CV
The stations
AI Project Lead, Business Development
AI-driven business models from opportunity sizing to a prioritised roadmap with KPIs, build-versus-buy calls, and delivery from concept to launch. Sales time down 30%, proposal feedback time down 90%.
Founder
A live cycling platform I built end to end, 331 hand-picked routes across 12 countries, with real users. Product, code and growth in one pair of hands.
In between jobs and own ventures
Built smedium (first clients, websites shipped, AI workflows automated), grew Pedal Peak, and went deep on AI tooling and prompt engineering. Also cycled in Togo and Benin.
Senior Product Manager, Lead E-Commerce
Owned the e-commerce ecosystem and strategy for a CHF 100M+ business, reporting to C-level. +9% conversion, +15% checkout step rate through research, A/B testing and analytics.
Lead Project Manager
Owned partnerships with financial institutions including UBS and Baloise, and was the bridge between senior client stakeholders and the internal product team.
Marketing & Growth Lead, Founding Team
Founding team of a health venture: two funding rounds closed, 10 locations, 23 people, 170+ customers, 1000+ client matches in year one on a hypothesis-and-data go-to-market.
Growth & Venture Builder
Go-to-market for several internal startups from validation to scale-up, with experimentation to lift conversion and cut acquisition cost.
Intrapreneur, Innovation
Market pilots from MVP to launch: Smide (now BOND Mobility), XpertCheck, Lizzy. Owned the MVPs, ran market experiments, briefed and managed agencies.
Earlier: Junior Trade Marketing Manager at Promena (2016), Junior Product Manager at Cruspi (2014 to 2015), Assistant Product Manager at Domaco (2010 to 2014), Accountant at Kuoni and AMAG (2008 to 2010), commercial apprenticeship at Bridgestone Switzerland (2005 to 2008). This is the retail and FMCG grounding the ad asks for, and where promotion economics became a habit.
04/About
Why this question and not another one
My first job in product marketing was working out whether a promotion would pay for itself. Volume forecast, what the trade would fund, what we would have sold anyway. If the sums were wrong the promotion still ran, and somebody found out three months later.
Sixteen years on, the mechanics moved into an app and got a points balance, and the question did not change at all. That is the part of product work I like most, and it is why the ledger above is the first thing I wanted to show you rather than a roadmap.
Outside work I ride a lot, which turned into Pedal Peak, a cycling platform I built and still run. It is where I learned to ship a consumer product on my own, and the reason I know what a week of real user feedback does to a plan.