Where the ideas come from

Under Elevate 28, WPP is moving from a holding company towards an operating company. My remit was the part of that shift that decides whether it works in practice: how the work itself gets planned, staffed, priced and measured across four operating units and four regions, using WPP Open.

WPP Open was already a platform for making work. My team extended it into one for making, managing and monitoring work — briefing, project creation, bottom-up planning and pricing from templates, resource and capacity management, demand against capacity, assignments and time recording, connected rather than bolted together. Project profitability and resource utilisation could be calculated as the time was entered rather than reconstructed weeks later.

T&P was the first agency to run it. Project management, resource management and timesheets went live there, and teams tracked forecast burn against plan.

A network of around 1,900 people across 45 offices on four continents. I joined with three connected ambitions: put AI into the working practice of the whole organisation rather than a lab at the edge of it, end the "black box" clients still experience at large agency groups, and sell and deliver business outcomes instead of time and materials.

The AI work is the one that landed. We trained the entire user base — not a pilot group, the organisation — to use AI in the work they do every day, and gave teams a safe environment to experiment in and deploy from. The point was never a centre of excellence producing demonstrations; it was changing what an ordinary Tuesday looks like for someone making client work.

Once people can use the tools, client ideas start arriving that need building. The Snickers campaign with José Mourinho is the example I point to: send a WhatsApp message describing a friend's mistake, receive a personalised video of an AI Mourinho teasing them for it. The idea came from the Mars client team and the innovation team; my job was helping them get it into the public's hands — guiding the specialist software partner, and holding the line on safety layers, scale and a campaign timetable that does not move. The One Show entry reports 444 million earned-media impressions and average engagement of over two minutes.

The outcome-based model was designed and ready to test with a client when the business was fully acquired by WPP. That trial never happened, so I published the blueprint instead — it became The Outcome-Based Revolution.

Power of One is Publicis Groupe's answer to a question every large group faces: how do you put the best people in front of a client, wherever in the organisation they happen to sit? Rather than sell a client a single agency, the group brings strategy, creative, production, media, data and technology together as one team — drawn from a network of roughly 100,000 people — and presents it to the client as one proposition.

My job was to build the platform that made that possible in practice rather than in principle. Roar let a single project combine people across agencies, countries and disciplines, and handled the commercial machinery underneath so the team and the client did not have to think about where in the Groupe anybody in the project was coming from.

I helped Publicis Groupe rise from third to first in its sector by market capitalisation by delivering the Power of One model through technology and improving how its agencies worked together. The company has ranked first by market capitalisation since 2023; my contribution was one part of a company-wide effort.

The team I built to do it grew to nearly 200 people. I identified and interviewed the core hires myself, then backed those leaders to recruit the rest.

A new CFO, who had watched technology take apart the industry they were coming from, asked me where technology fitted into the future of an agency.

I noticed agencies had enjoyed margins high enough that retrospective accounting was good enough. Once those margins tightened, decisions had to be made during the work, not reviewed after it.

So we built it. Real-time profitability and utilisation, with forecasting, on a platform flexible enough to bend to how BBH actually wanted to operate. It became BBH Pro, the predecessor to Roar. I also built a tool that locked people out of their computers until their timesheets were done — a feature that guaranteed to get data timely for the forecasts to mean anything.

What changed was not only the numbers. Management stopped spending effort on small, low-priority work and put it into projects that could build the agency's reputation.

Twice I have been the customer who ended up helping shape the product. With Clarizen — now Planview AdaptiveWork — I worked with the company on both the technology and the proposition: which capabilities genuinely mattered to enterprise customers, and how to explain them to a market that had not yet asked for them.

When Publicis decided to move to Adobe Workfront, my team assessed it against everything we had already built and catalogued precisely where it fell short of what an organisation of that size and shape needed. Rather than work around the differences, we formed a joint team with Adobe's Workfront organisation, handed over the requirements, designed the solutions together and agreed a delivery roadmap. Adobe released iteratively; my teams tested, validated and fed back.

Adobe and Publicis presented the result at Adobe Summit 2026: Roar Workfront running at around 60,000 users across more than 800 agencies.

Between running my own software businesses and joining an agency, I worked as a digital strategy consultant to large institutions — including an engagement with the Chief Information Officer of the Bank of England, and a project for Boots.

The detail of both is confidential and stays that way. What's important is that the same thinking has been applied inside a central bank and a national retailer.

I founded Sibilo in 2002 to build a hosted e-commerce application - Shopify-shaped idea except years earlier. We built it and sold only six in the first year. With an office and a payroll we had to change direction.

What we changed to was DataSmart, built for the UK's incoming data-protection compliance requirements. It became a profitable enterprise product with customers including BP, Volvo and a large part of the public sector — at its peak, 15% of the UK public sector. We designed successive versions with the customers themselves, which is now called customer-centred product design and then was just how we worked.

The cash funded two things. We invested in startups — some worked, more did not, and I learned more from the failures. And we acquired a software development company in Turkey and grew it, which gave us the distributed delivery capacity for custom work: more than fifty enterprise solutions, sitting in the gap between what small suppliers would aspire towards and what the large integrators wouldn't work on.

We built the business to be credible at that end of the market: a Microsoft Partner with ISV competency, an Oracle Partner, certified to ISO 9001:2000, and a winner of a DTI Smart Award. None of that wins work on its own, but without it a company our size never reaches the table.

Two I still enjoy describing: a sports-betting platform for an Italian casino business, Stravinci, automated far enough that a single bookmaker could oversee what normally took a room of them; and an online booking platform for The Landmark Trust, whose inventory includes lighthouses, castles and houses on Lundy Island.

Woto combined the immediacy of social sharing with flexible website creation and simple publishing tools. By October 2014 it was reaching more than seven million monthly unique visitors in 117 countries, with over 80,000 social shares.

Backed by Ingenious Ventures, Galata Business Angels, BBH's Zag and Sibilo. It is closed now — a useful education in how quickly an audience can arrive, and how little that guarantees.