The most unexpected truths are found at the intersections.
A marketing and growth leader who works where data and creativity meet. Twenty-five years in the craft, and across the last fifteen, applying emerging technology and now AI to one method: reading the unprompted signal, finding the unexpected truth (found at the intersection of surprising and familiar; the difference between a head nod and a head turn), and building the systems and stories that turn it into growth.
Background
Honed agency-side across Ogilvy, BBH, JWT, BBDO, R/GA and M&C Saatchi, then put to work in-house as a CMO and founder, building brands and growth engines directly. The brand-building spans nearly every category worth working in: technology (Microsoft), energy (ExxonMobil), telecoms (Verizon), CPG and QSR (Nestlé, PepsiCo, Starbucks, Pizza Hut), and financial services, including investment marketing for T. Rowe Price. The constant is persuasion across radically different audiences, from household decision-makers to multimillion-dollar enterprise technology buyers to the world’s wealthiest collectors. A psychology training still shapes how I read the room.
Different readers care about different things. Choose a lens and the work re-sorts to speak to you.
Selected work, in sequence
The problem. At the nadir of the 2009 crisis, T. Rowe Price needed to keep frightened investors from panic-selling and argue for disciplined active management to serious, skeptical readers of the financial press.
The approach. JWT’s celebrated “Ink” platform had established that the firm’s value lay beyond the surface numbers, in rigorous in-the-field research. I led the strategy carrying that television idea into a Wall Street Journal print campaign, translating a fluid, emotional film concept into a stark, evidence-forward “Truth” series for readers who wanted proof, not reassurance.
Ran in The Wall Street Journal and other premium financial titles at the depth of the crisis.
Extended a campaign widely regarded as a landmark in financial-services advertising.
Brand & Persuasion
The problem. Microsoft's enterprise buyers make multimillion-dollar technology bets and distrust advertising; straight promotion erodes credibility with exactly the people it's meant to persuade.
The approach. Reframe the brand from advertiser to publisher. The agency became an ongoing information network of 40+ developers, analysts and enthusiasts running a real-time loop and producing genuine editorial (interviews, live chats, analyst roundtables) distributed into the communities the audience already trusted.
300+ posts published; 1.4M+ total page views; one of the most popular Microsoft blogs.
around 9 million user engagements across the network over seven months.
Recognised with a Bronze Integrated Lion (Cannes Lions) and a One Show Interactive Merit (2010).
An early example of native content and real-time reactive messaging, years before either was standard.
Brand & PersuasionData & Systems
The problem. M&C Saatchi's CRM agency (LIDA, out of London) needed a New York presence from scratch: no office, no US clients, no team.
The approach. Partnered with the CEO to build the office end-to-end (team, positioning and new-business engine) and won its first major client, a global luxury wine & spirits group. Sourced an external data-science firm (Davis & Partners) and pioneered its AI/NLP suite for CRM: enriching profiles with social data and using embeddings to power more complete, personalised campaigns.
NY office built from $0 to roughly $10M in revenue with the anchor account.
Team of around 40.
A data-led CRM offering: machine learning, social signal and embeddings on an enterprise martech stack.
Data & SystemsOperator & Outcomes
The problem. Get a tiny pool of the world's wealthiest collectors to bid higher than ever on one work, while those buyers were kept anonymous, even from the agency. No CRM, no direct outreach; the usual luxury playbook was off the table.
The approach. Invert the model to one-to-many-to-few: build mass cultural hype so owning the piece became the ultimate status signal to the few who could buy it. Mined the body of contradictory commentary about Rabbit and made contradiction itself the medium, selling the outrage, not the object.
1.5B+ impressions (around 1.1B social); engagement ran 200% over CTR benchmark.
Against that backdrop of attention, Rabbit sold for $91,075,000, a world record for a work by a living artist.
Recognised with an Effie for business effectiveness.
Brand & PersuasionOperator & Outcomes
The problem. A genuinely powerful AI/NLP asset was being sold narrowly, as an implementation shop. The capability far outstripped how it was taken to market.
The approach. Repositioned the firm from implementer to AI-enabled marketing-solutions company and productised its engine into two market-facing offerings, owning concept and go-to-market for each.
Media Asset Analyzer: near-real-time sentiment-and-impact measurement for major media events; deployed for a major faith-based organisation.
Fuel.AI: a multilingual trend-scoring engine, taken to market with Leo Burnett (2021).
Both products ultimately sold as licensed IP; retained as a board advisor.
Data & SystemsOperator & Outcomes
The problem. Intent HQ had landed its first major contract but was scaling without a spine: no clear positioning, no defensible differentiation, and a sprawl of products that muddied the story for customers and investors.
The approach. Defined the North Star positioning and rebuilt the brand end-to-end; rationalised the portfolio into a focused two-pronged strategy and shaped the case for a third, enabled by the Anagog merger; partnered with the CEO on investor relations through fundraising.
roughly 3× revenue growth over two years (internal figure), and continued backing from existing investors.
Named AI Company of the Year at the British Data Awards 2024 (321 nominations, judged by a panel of CDOs).
Featured on the FT1000 (fastest-growing European companies) in 2022 and 2023.
The problem. Customer-experience measurement leans on surveys: people answering when asked, inside the frame the question sets. The richer signal is what people say unprompted, but it’s unstructured, noisy, and hard to compare across a journey.
The approach. I built CxFI in two halves. Frictions: an NLP/ML pipeline reads 2+ years of unprompted public conversation, isolates the discrete experiences along a customer journey (from 31 micro-steps to 14 macro-experiences), and scores each by net sentiment weighted by reach, normalised into a comparable index. Fources: four experience-design levers (See, Say, Think, Do) translate the emotion discovered at each step into a counteracting design response. Measure the friction with data; design the fix.
Deployed for household-name brands across QSR and CPG/beverage (anonymised).
Delivered root-cause friction insight clients described as something they’d always wanted but had no other way to obtain.
Reads the truths people share when no one is asking, unlike survey-based CX indices.
Data & SystemsBrand & Persuasion
The problem. A commerce arm was being spun out of its parent and needed to become a standalone brand, with its own identity, positioning and go-to-market, and a sharp reason to exist in a post-cookie market.
The approach. Identified the positioning (real, consented, receipt-derived purchase data as the replacement for the third-party cookie) and created the brand "Actual," the name carrying the argument. Architected the brand fast, built the GTM and marketing engines, and executed AI-natively: on-brand tooling, AI-generated reporting, and a unified-customer-intelligence architecture across the funnel.
A complete standalone brand, GTM engine and marketing engine, AI-native from day one.
Live in-market as "the Permissioned Purchase Data Platform," with named partners including DoorDash, Instacart, Amazon, General Mills, Ibotta and Upside.
The business has scaled materially during my tenure.
A flagship point of view: AI as an orchestration layer over a unified customer-intelligence base, deploying the optimal motion per customer, per moment. Its discipline, knowing when not to act, speaks directly to premium, regulated businesses.
A peer-published article whose real value is the arc it starts: the 2011 listen-and-respond loop is a manual ancestor of today's AI orchestration layer, evidence of foresight consistent across fifteen years.
BrandLens (a positioning framework built on the who / what / how / why of a brand, actionable and the basis for modern GTM), See-Say-Think-Do (engagement design), the 5-D Productization Model (turning capability into monetizable offerings), "The 2X4" connected-age model, and the "Unexpected Truths" methodology. Together they form a connected, still-growing body of proprietary frameworks.
Why these
Different problems, different decades. One way of working underneath: listen to the unprompted signal, find the unexpected truth, then design and orchestrate the response, privacy-forward and productization-minded throughout.
