Portfolio
I build the data and AI systems that sit underneath marketing decisions, not the campaigns themselves. Over the years, I have moved from business intelligence and web analytics through customer data platforms and attribution modelling to the AI products I design and ship today. That first chapter alone included a BI standardisation programme that delivered an estimated ten million dollars in efficiency gains. The thread connecting all of it is the same one: translating messy, disconnected business data into something a team can actually act on.
I work as a practitioner, not purely an advisor. Most of what I recommend, I have already built, tested, and broken at least once, across the range below.
What I Do
Data Strategy & Governance
Frameworks, policies, and operating models that keep data trustworthy, compliant, and aligned to business goals.
Customer Data Platforms & Identity Resolution
Unifying scattered customer records into a single, durable profile for accurate segmentation and activation.
Marketing Analytics, Attribution & Measurement
Connecting spend to outcomes through multi-touch attribution and incrementality testing, without leaning on last-click data alone.
SEO & Generative Search (GEO)
Technical SEO and the shift to AI-driven discovery, from Search Console pipelines that outgrow standard reporting to earning visibility inside AI-generated answers.
Conversion Rate Optimisation
Rigorous experimentation and funnel diagnostics that turn more visitors into customers.
AI Product Development & Automation
AI agents, retrieval-augmented systems, low-code automation.
Cloud Data Warehousing & Data Engineering
Scalable pipelines and warehouse architecture that make reliable data available where it's needed.
Data Science & Predictive Modelling
Statistical and machine learning techniques, including RFM scoring, clustering, and churn and lifetime value models, that turn raw transactional data into actionable segments and forecasts.
Project & Product Management
End-to-end delivery of data and AI programmes, from roadmap and requirements through vendor coordination, stakeholder alignment, and launch, so technical work ships as a finished product rather than an open-ended build.
AI Products I Have Built
Alongside client delivery, I design, build, and ship my own AI products. Both of the following are self-initiated, built from scratch, and in active use today.
AI Solution Advisor
What it is
A tool that matches a business to the right tools for its specific situation, rather than handing over another generic top-ten list. A structured assessment captures a business's context, goals, and constraints, then a retrieval layer matches that brief against a curated knowledge base of over 200 named tools, each scored across more than 25 fields covering fit, cost, technical requirements, and regional considerations. The output is a detailed implementation proposal with a recommended tool stack, an architecture diagram, and a roadmap, plus a chat interface for follow-up questions grounded in the same knowledge base.
How it evolved
The tool went through three generations. Version one simply matched a use case to a ranked tool list. Version two added a full assessment brief and produced strategic recommendations rather than a bare list. Version three, the current platform, added proposal generation, template-based documentation, and a retrieval-augmented chat agent, turning it from an internal advisor into a client-facing product.
Where it stands
Not yet publicly launched, since it currently runs on personal API credentials, but available for a live demo and a free assessment on request.
Avenote: AI Content Studio
What it is
A platform that takes a single voice recording, document, or webpage and turns it into whichever of thirteen-plus content formats the user actually needs: articles, blog posts, mind maps, presentations, journals, newsletters, social posts, and more. One transformation engine, five capture methods, translation into over 33 languages, and a growing set of integrations into the tools people already use.
How it evolved
Avenote grew in five layers, each one built on the last rather than replacing it: capture, transformation (transcription, translation, summarisation), content generation across the full format library, intelligence features like interview mode and organisation, and finally integration and distribution through a browser bookmarklet, Zapier and Make connectors, and direct email delivery.
Who uses it
Product managers turning meeting notes into roadmap slides, podcasters and creators converting interviews into articles and social content, consultants turning voice notes into client-ready briefs, agencies scaling content production across languages, and researchers capturing and synthesising sources while they browse.
Where it stands
Live and in active use at avenote.co.
Selected Project Portfolio
Client names are withheld on this site. Details are available on request.
Conversational AI Assistant: Global QSR Brand
Challenge
Customers had questions about a global quick-service restaurant brand's products, services and operations, covering health, hygiene, ingredients, transparency, compliance, legality, diversity and inclusion. The answers lived on a website that had become a weak touchpoint for those questions.
What I Built
Led the design and delivery of a bilingual conversational AI assistant that answers these questions directly, grounded in the brand's approved information.
Result
The assistant is in production across seven markets with strong adoption, and the brand has retired the legacy website it replaced.
Customer Data Platform: Luxury Automotive Brand
Challenge
A luxury automotive brand was running its digital presence across ten-plus marketing and ecommerce websites, an OEM layer, and multiple dealerships, each with its own fragmented tech stack and uncoordinated media activity. There was no unified view of a customer across any of it.
What I Built
Led the implementation of a customer data platform (Tealium) that ingested and harmonised data from marketing sites, ecommerce properties, Salesforce CRM, and Adobe Analytics into unified user profiles. Developed close to a hundred behavioural, transactional, and relationship-based audience segments, powered by a proprietary lead scoring model, then activated them across Meta, Google DV360, CRM channels, and the brand's own websites using a dynamic creative optimisation strategy. The programme ran across ten markets in the region.
Result
In the markets where CDP-built audiences went live, lead generation rose 30 percent and acquisition cost fell 20 percent, alongside lower bounce rates and higher time on site and media engagement. In other markets, a separate conversion rate optimisation programme integrating GA4 and Adobe Target lifted leads 10 percent.
Marketing Attribution: Gaming & Entertainment Platform
Challenge
A regional gaming and entertainment platform was making media spend decisions on last-click data alone, with no reliable way to see which channels actually influenced a purchase versus which simply closed it.
What I Built
Gathered data from web analytics, CRM, advertising platforms, and sales records into a single dataset, then evaluated first touch, last touch, linear, time decay, and position-based attribution models against the brand's actual sales cycle. Implemented a custom position-based model that weighted first and last touchpoints most heavily while still crediting the middle of the journey.
Result
The distribution showed that while the last touch carried real weight, mid-funnel social activity was doing more work than last-click reporting gave it credit for. Predicted customer lifetime value rose 10 percent, and the resulting budget reallocation is estimated to have lifted overall marketing ROI by 22 percent.
Customer Segmentation at Scale: QSR Multinational
Challenge
A quick-service restaurant multinational had years of clean transactional data across a customer base in the tens of millions, and no segmentation strategy built on top of it, which meant every campaign treated the full customer base the same way.
What I Built
Extracted transactional data into BigQuery, engineered recency, frequency, and monetary scores for every customer, and ran a K-means clustering model in BigQuery ML to group customers by behavioural similarity. Integrated the resulting segments with downstream channels, including the CDP, CRM and in-app personalisation, so each segment received tailored messaging and offers.
Result
Segment-led activation improved revenue 6 percent and customer lifetime value 12 percent. The model was handed over with a plan for ongoing monitoring, so the segments stay current as purchase behaviour shifts.
Organic Search Analysis via GSC API: B2B Software
Challenge
A software solutions provider could not get a clear picture of its own SEO performance. The Google Search Console interface caps reporting at 1,000 rows, restricts most views to a single dimension at a time, and its native connector into Data Studio was slow and, in places, produced numbers that did not reconcile with the same metric pulled through a different connector.
What I Built
Built a pipeline through the Search Console API directly into BigQuery and Looker Studio instead, which lifted the row ceiling to 50,000 per property and allowed keyword and page-level data to be combined in ways the standard interface never permitted. With that visibility, refined keyword-to-page mapping, rewrote meta tags and thin content on underperforming pages, and identified specific pages and keywords worth targeting for backlink outreach.
Result
Within six months of the optimisation programme going live, organic sessions grew by double digits, the number of pages ranking on page one for their target keywords increased materially, and the backlink outreach shortlist converted into several high-value referring domains, all tracked through the same pipeline that replaced the 1,000-row interface.
Automated Customer Journeys: Healthcare & Consumer Goods MNC
Challenge
A global healthcare and consumer goods client was running customer communications as manual, one-off sends, with no systematic way to trigger the right message at the right moment in a customer's journey.
What I Built
Implemented and optimised automated, trigger-based customer journeys on Salesforce Marketing Cloud, instrumented with Google Analytics and Google Data Studio so that journey performance could be measured and refined rather than left to run untouched after launch.
Result
Trigger-based journeys tripled user engagement.
Digital Transformation & Ecommerce Growth: Automotive Dealership Group
Challenge
A Dubai-based dealership group representing two global car brands needed to modernise a digital operation spread across the OEM relationship, the dealership's own departments, and inconsistent performance marketing, with digitisation and ecommerce still nascent in the region's automotive retail sector at the time.
What I Built
Led the dealership's broader digital transformation programme, took direct ownership of ecommerce and performance marketing execution across channels, and drove adoption of performance-based, CRM and CRO-led analysis across both the dealership's own departments and the wider group, rather than leaving digital as a siloed function.
Result
The mobile app audience tripled within a year, lead generation grew 3x and cost per acquisition fell 60 percent. Ecommerce sales of the group's premium car brands grew 3x, in a business that was already a regional pioneer in automotive ecommerce.
Enterprise BI Standardisation: Global Airline
Challenge
Multiple overlapping BI processes, tools, and reports across departments with no common standard, causing overlaps, redundant tools, and delayed delivery.
What I Built
Formed and led a 15-person Business Intelligence Competency Centre to standardise BI tooling and reporting across the business, then led a migration from Oracle and MS SQL Server to open-source MySQL.
Result
The standardisation programme delivered ten million dollars in cost savings and a 15 percent efficiency gain, and the database migration saved a further five million.
Productised Services
Working inside a global agency network gave me the chance to turn recurring engagements into repeatable, productised offerings rather than one-off consulting work each time.
Data Audit & Solutions Consultancy
A structured entry point covering progressive profiling, advanced segmentation, and cloud data warehouse or data lake design, built to meet a client wherever their data maturity actually is rather than where a standard proposal assumes it is.
CDP Accelerator Programme
A six-stage delivery framework — collect, standardise, transform and enrich, integrate, activate, then roll out and iterate — that turns a customer data platform deployment from an open-ended technology project into a defined, repeatable engagement.
Measurement Management Framework
A staged approach to measurement maturity: audit, plan, setup, interactive dashboards, and ongoing management, built so that measurement infrastructure keeps working after the initial project ends rather than decaying the moment the consultants leave.
Automation Adoption Framework
A structured path for small and mid-sized businesses to adopt automation without overreaching: chatbots for customer support, onboarding, and training; enhanced automations for recommender systems, business processes, and system integrations; and custom applications ranging from enhanced dashboards to full standalone apps.
Education & Credentials
Speaking & Mentoring
Digital Confex 2022
Presented on the future of data-driven marketing to an audience of over 1,000 digital professionals and entrepreneurs.
Workshops & Mentoring
Delivered workshops and mentoring sessions on digital optimisation for business growth to professionals, university students, and startup entrepreneurs, in association with Westford University and Eaton Business School.