Customer Success · Commercial Operations · Founder-Side Execution

I turn messy customer problems into clear decisions, better systems, and commercial momentum.

I’ve led customer and account operations across global markets, built operating systems from scratch, and turned complex customer context into practical actions teams can actually use.

7 → 15

Immediate team scaled while supporting Spotify Advertising.

60 → 95%

Team quality improved after introducing clearer operating standards.

20 → 40

Weekly case baseline per person increased without sacrificing quality.

1,000+

Active advertisers across the broader portfolio I worked across.

Selected work

Six ways I think, build and operate.

Live products, strategic case studies, a commercial proposal and a rapid decision-system build. Each project shows a different part of the same operating instinct: find the real problem, make it legible, then build a practical way forward.

02 Working V1 · Founder QA

Make It Add Up

Life planning through financial reality

A local-first planning product that works backward from the life someone wants into the numbers and actions required.

ProductSystems ThinkingUX
Open live product ↗
03 Rapid build

Bolter

Campaign intervention decision system

A fast proof-of-work build translating Spotify-style campaign judgment into explicit checks, human gates and deterministic routing.

Decision DesignQARapid Prototyping
Open live build ↗
04 Strategic customer success case

Braze × Gojek

At-risk account rescue through lifecycle personalization

An executive-business-review style recovery plan that starts with customer behavior and business yield, not a feature dump.

Account RescueRenewalLifecycle Strategy
Full case ↗PDF ↓
05 Anonymized CRM case

European recruiting company

CRM operating model & at-risk search recovery

A 12-month CRM operating model plus a working HubSpot proof of concept designed to prevent delivery and relationship leakage.

CRMHubSpotChange Management
Full case ↗PDF ↓
06 Confidential proposal

Confidential premium retail brand

CRM & retention transformation strategy

A founder-facing operating proposal focused on reliable measurement, clearer ownership and selective scaling rather than more activity.

RetentionOperating ModelCommercial Strategy
Full case ↗PDF ↓

Operating experience

Built in the work, not just on slides.

I’m most useful where customer judgment, commercial ownership and operating discipline need to coexist.

Spotify Advertising · via Teleperformance

Global Lead, Account Coordination & Customer Success

2023–2025

Supported customer and account operations across NA, EMEA, LATAM and ANZ, moving from frontline account work into global leadership. The role sat between client needs, campaign performance, internal execution and commercial opportunity.

Customer

Account health, escalations, reporting, proactive outreach, RFP support and client-facing recommendations.

Commercial

Retention, growth signals, dormant-account recovery and expansion opportunities alongside Client Partners.

Operating system

QA standards, reporting logic, playbooks, prioritization and scalable workflows for a growing global team.

How I work

Diagnose the real problem.
Make the logic visible.
Build only what earns the right to scale.

My background crosses customer success, account management, operations, consulting and founder-side work. The common thread is turning incomplete information into a clear operating choice, then staying close enough to execution to see whether it actually works.

01 · Start with context 02 · Separate signal from noise 03 · Keep human judgment where it matters 04 · Measure before scaling

Contact

Need someone who can own the customer problem and improve the system around it?

Selected work includes live prototypes, strategic case studies and pre-engagement proposals. No confidential employer or customer data is published.

01 · Live product prototype

Narelle

A Preparation OS for Customer Success, designed around the gap between having customer information and actually being ready for the conversation.

Open live demo ↗
The problem

Customer teams often have the data but still rebuild context before every important moment.

Narelle treats renewal reviews, QBRs, risk conversations and handovers as moments that require preparation, not just record retrieval.

What I built

A guided preparation layer.

The V1 demo organizes customers, customer moments and a preparation workspace so the user can move from “who needs attention?” to “what do I need to know and decide?”

Operating principle

Preparation should reduce context reconstruction.

The product does not try to replace the CRM. It sits closer to the decision moment, helping the user pull relevant context together before a consequential customer conversation.

Scope honesty

Scripted V1, intentionally bounded.

The public version is a scripted product demo with simulation data, not a production CRM or live AI system. The point is the product logic, workflow and decision design.

“Having customer information is not the same as being prepared.”

02 · Working V1 · Founder QA

Make It Add Up

A local-first life and money planning product that works backward from the life a person or household wants into the monthly reality required to support it.

Open live product ↗
The product loop

Plan → Home → Act → Track → Invest → Explore

The system connects goals, obligations, reliable income, spending and investing rather than treating them as isolated calculators.

Design choice

Explainable over mysterious.

V1 is frontend-first and local-first, with browser storage, no bank connection and no opaque score. Calculations are meant to be understandable and pressure-testable.

Household reality

Solo or together without forcing one model.

Users can plan individually or as a household, see combined affordability and individual contribution, and stress-test a one-income fallback.

Current status

Feature-complete V1 under founder QA.

The working product is live while calculation red-teaming, founder QA and final polish continue. It should be viewed as a real working V1, not a finished financial product.

“The life you want. The numbers it takes.”

03 · Rapid build

Campaign Intervention Tracker

A rapid decision-system build translating real campaign-management judgment into explicit checks, bounded actions and human approvals.

Open live build ↗
Decision problem

Not every underperforming campaign needs the same intervention.

The workflow distinguishes whether to intervene, investigate, wait or escalate a growth opportunity based on pacing, performance, timeline, technical issues, client requests and dependencies.

Hardening

Make the verdict traceable.

The build was pressure-tested against a 14-case golden fixture set, with explicit checks, visible divergence and deterministic verdict logic rather than free-form “AI judgment.”

Guardrails

Human approval stays in the commercial loop.

Client sign-off, policy/stakeholder dependencies and restricted-industry states are treated as gates. The system surfaces the decision path instead of pretending it owns authority it does not have.

What it proves

Fast ramp without sacrificing reasoning quality.

The project was built to show how domain experience can be translated quickly into a usable decision workflow, then hardened through fixtures and reruns.

04 · Strategic customer success case

Braze × Gojek

An at-risk account rescue plan framed as an Executive Business Review: diagnose the business risk, define the customer behavior to change, then connect product capability to renewal value.

Diagnosis

The problem is behavioral before it is technical.

The case starts from low Braze adoption, a declining North Star, a six-month renewal horizon and executive scrutiny on cost versus a legacy platform.

North Star

% of monthly active users transacting on 2+ Gojek services within 30 days.

No numerical target was invented because the case did not provide a baseline. Supporting metrics explain whether personalization creates adoption and whether adoption becomes habit.

Recovery architecture

Grow the ecosystem + protect the ecosystem.

Program A moves single-service users toward a second service. Program B reinforces multi-service behavior so gains do not disappear. One operating model supports two lifecycle moments.

Renewal logic

Defend business yield, not feature count.

The recommendation is to test whether Gojek can realize more value from capabilities it already owns before paying the migration, rebuild and retraining cost of moving backward.

“Customer Success optimizes business outcomes, not feature usage.”

05 · Anonymized CRM case

CRM Operating Model

An anonymized CRM strategy and proof of concept for a European recruiting business, focused on reducing leakage across demand, delivery and relationship continuity.

Operating question

What should CRM actually improve?

Not more CRM usage. Less leakage across demand, delivery and relationship continuity.

System boundary

CRM coordinates the relationship. ATS owns search execution.

Only selected risk signals and owner actions move into CRM. Detailed candidate work remains in the recruitment system.

POC logic

Detect → validate → diagnose → recalibrate → recover.

The workflow keeps a human validation gate before customer-facing escalation and makes the recovery path explicit.

Validation discipline

Backtest, shadow, pilot, then scale.

Search Recovery Rate measures recovery while False Positive Rate protects user trust and prevents automation noise.

“The goal is not more automation. It is a clearer commercial operating rhythm.”

06 · Confidential proposal

Confidential Premium Retail

An anonymized founder-facing CRM and retention transformation proposal for a premium retail business in Indonesia. Client identity, people, fees and identifying operating details are intentionally excluded.

Diagnosis

The ingredients existed. The operating model was not yet reliable.

Customer touchpoints, retention activity and reporting needed clearer ownership, measurement and execution rhythm.

Recommended direction

Foundation → execution → evidence → selective scale.

Stabilize the fundamentals, improve a small number of important journeys, measure response, then scale only what earns it.

Four-month path

Validate · Build · Improve · Transfer.

The proposal moves from reality-checking and prioritization into implementation, evidence-led improvement and ownership handover.

Scope honesty

Proposal, not claimed implementation.

No implementation result is claimed. The public artifact demonstrates discovery, prioritization and operating-model design only.

“Make the foundation reliable → improve selected execution → measure what works → scale selectively → transfer ownership.”