JD Case Study · Citizens

Head of Decisioning Analytics & Optimization

A "Head of" role owning decisioning analytics and the optimization framework for a card portfolio — segmentation, statistical models, test-and-control, account-level profitability. This is exactly the decision-management and optimization engine I built (funnel, experimentation, LTV), now at portfolio scale.

US · multiple sites · View the job description →

Jun 22, 2026

What they’re actually buying

Build the decision-optimization layer that lifts portfolio ROI. Model the decisions — who to target, approve, and how to treat them — prove them with test-and-control, and wire the whole thing into an optimization framework that improves campaign and channel returns. Optimize the funnel, not one campaign at a time.

How I’d own it — first 90 days

Days 1–30

Map the decisions

Inventory the decision points (targeting, approval, treatment, retention) and their KPIs. Baseline account-level profitability by segment, and meet the Marketing, Risk, and Analytics partners.

Days 31–60

Stand up the optimization diagnostic

Ship a decisioning dashboard that shows the targeting → approval → activation → profitability chain by segment, and pinpoints where the framework is leaving ROI on the table.

Days 61–90

Run the first optimization experiment

Design a test/control on a decisioning change with profitability guardrails, deliver the business case in dollars, and stand up the experimentation governance and optimization cadence.

Signature analysis: the portfolio decision funnel

Eligible population 100 indexed to 100
Targeted 54
Approved / accepted 29
Activated 21
Profitable at 12 months 17 the only stage that matters

The optimization insight: the stages aren’t independent. Tightening approval to cut losses can starve activation and profitability; loosening targeting floods the funnel with low-value accounts. The job is optimizing the chain jointly against account-level profitability — which is what a real decisioning framework does, and a campaign-by-campaign view never will.

Illustrative figures — a portfolio decision funnel optimized end-to-end against profitability, not stage by stage.

Requirement → proof

What the JD asks forWhat I’ve already shipped
Decisioning analytics & optimization framework A full decision funnel diagnostic that locates and sizes the ROI leak.
Test & control, statistical / segmentation models An A/B + incrementality readout with guardrails and a sample-size calculator.
Account-level profitability / LTV An LTV / unit-economics model where the loss rate is the central optimization lever.
SQL / SAS, advanced analytical tools Show-the-SQL blocks, channel-aware, warehouse-ready.
Card / consumer-credit portfolio domain Competitive spotlights across the consumer-credit and BNPL field.

The same engine

The engine behind this portfolio is a decision-optimization framework: a funnel modeled end-to-end, experiments with profitability guardrails, and an LTV model that prices every trade-off. I built it for a lease portfolio; on a card book the levers are targeting, approval, and treatment — same machine, same job.

See the engine →

Decisioning is optimizing the whole funnel against profit, not tuning one campaign — and I’ve shipped exactly that, in public. Let’s talk about your ROI frontier. — Paul Brown