20+ yrs
Data science & decisioning — banking and beyond
A data scientist by training — two master's degrees in econometrics and analytics — with over two decades building decisioning capability, most of it inside retail banking: CBA, NAB, Westpac, Suncorp, Emirates NBD. I build the customer decisioning layer that turns analytics and models into the next action a customer actually sees — so the investment lands in revenue, not on slide 12.
The consultancy delivers the framework and leaves. Eighteen months on, the platform is licensed, the data and marketing teams have stopped speaking, and the use cases that were live on a slide are stalled in a backlog nobody owns.
I've spent a career operating inside that gap as the person accountable for the thing actually working — not advising on it from the outside. That is a different kind of knowledge, and it's the only kind that survives year two.
Marketing won't act on decisions when it doesn't trust the data feeding the engine. No platform fixes that. It takes someone who has stood on both sides of the table at once.
When an offer fires matters more than the offer. Getting it right means holding customer behaviour, channel economics and risk appetite in your head simultaneously. Most programmes get one of the three.
Wave one launches, leadership declares victory, the team scatters. By month eighteen the engine runs stale models with no owner. That's governance, not technology — and it's the part nobody scopes.
Anyone can be trained to configure a platform. What can't be handed to a junior resource is knowing which decision to make, for whom, at what moment — and whether it's worth making at all. That starts with asking why, relentlessly, and framing a messy business problem as a decision flow precise enough to survive production. The platform just executes it.
That judgment didn't come from the technology. It came from sixty years of direct marketing — Lester Wunderman, who coined the term and made it measurable; Arthur Hughes, who built it around customer lifetime value; Peppers and Rogers, who made it one-to-one. The tools went real-time; the discipline didn't change. I learned it hands-on, under Malcolm Auld, before any of it was automated — which is why I can tell an executive why a decision works, not just how the system fires it.
My 2019 master's thesis on Next-Best-Action decisioning has now passed 2,000 downloads — more than 500 in the past twelve months alone, and still climbing. More than six years on, it's still being pulled into production environments across banking, insurance and financial services worldwide. Not cited in journals. Deployed.
Decision Management for Next Best Action Marketing. An econometric and architectural framework covering propensity modelling, action arbitration, channel attribution, and the organisational conditions real-time decisioning needs to survive at scale.
At Emirates NBD I took the real-time decisioning programme from two people and no live use cases to a 40-plus use-case capability across eight channels in under two years. Independent attribution put the programme's incremental revenue at a multiple of its run cost — the kind of return that justifies the capability, not just the licence.
I don't produce the deck. I help you build and own the customer decisioning capability the deck describes — working with a small number of banks at a time, typically three or four.
A short, vendor-agnostic review of where your decisioning actually breaks — data trust, trigger timing, governance, ownership. You get a clear read on why the spend isn't converting and exactly what to fix first.
I stay alongside your team through the build and into year two — the architecture, the arbitration logic, and the governance that keeps the engine alive after the launch headlines fade. The part that usually has no owner.
You've bought the platform and the models, and you're still not seeing the decisions or the revenue they were supposed to produce.
You're looking for a pair of hands to configure to someone else's design. That's a resourcing decision, and it isn't what I do.