This week Leaders Loop published an editorial profile of me, a properly reported one, by their senior journalist Louise Harrison, who went looking well past the LinkedIn headline. I’m genuinely chuffed to be the subject of one of their leader profiles, and not only for the obvious reasons of vanity (though I’ll own a little of that too).
The version of the story most people meet: a stage, a headset, a slide. The profile went looking for the part before that. Photograph from a recent Product Leadership summit.
The strange part is what it does to you. Reading your own life narrated by someone else (the chip-shop counter in Surrey, the two crossings of the world, the long unglamorous apprenticeship in payroll, the venture, the stage) is disorienting in a useful way. Things you’d only ever felt as instinct suddenly have an outline. Louise found a throughline I’ve spent two decades clumsily gesturing at in talks without ever writing it down: better, not just faster.
So rather than let the profile do all the talking, I want to take its raw material and say the quiet part out loud. These are the four principles I’d actually defend, not as slogans, but as things I decided, somewhere along the way, usually the hard way.
1. Better beats faster, and it isn’t close
Most teams I’ve worked with have no shortage of speed, ideas, or output. What they’re short on is decision quality. The bottleneck is almost never “can we move faster”; it’s “are we moving in the right direction, for reasons we could defend in six months.”
Good decisions need context: goals, constraints, real customer insight, data, trade-offs, and history. Strip that context away and what you get isn’t progress. It’s confident motion in the wrong direction, delivered impressively early.
A clean plan and the right plan are not the same thing; the difference is usually a person, not a number.
This is also the spine of everything I’ve been saying about AI lately. Large language models are brilliant guessing machines: superb with ambiguity, genuinely dangerous when you mistake them for the deterministic engines that reliable execution still demands. The interesting work isn’t prompting harder. It’s building the harness: the structured, persistent context that lets a model critique a real decision instead of dispensing fluent advice about a generic one. AI should sharpen judgement, not replace it.
2. Change your mind when the context earns it
The profile makes much of my zig-zags: Melbourne at seventeen, back to England, a year into a degree at City, University of London, then back to Australia for good. It would be easy to read that as indecision. I’d argue it’s the opposite.
At more than one fork I weighed the tidy spreadsheet answer against the things a spreadsheet can’t hold, and chose the messier, more human option, usually family. Each time, the plan lost to the context. I don’t regret a single one of those reversals, and I’ve come to think the willingness to revisit a decision when the facts change (rather than defend it out of pride or sunk cost) is one of the most underrated leadership skills there is.
Roadmaps are where this instinct goes to die. Once something is “on the plan,” it accretes owners, momentum, and sunk cost, and the organisation quietly stops asking whether it’s still the right thing. Good leaders keep asking anyway.
3. Respect the system underneath
I started my career at ADP, in payroll. It is the least glamorous and most instructive place a young technologist can possibly land. Payroll is unforgiving: deadlines that cannot slip, numbers that must be exactly right, a world where “roughly” is just a longer way of spelling “wrong.”
I spent my twenties living in that deterministic half, and it permanently changed how I see products. There is machinery underneath every shiny interface, and it has to work whether or not anyone is watching. That respect for the boring, reliable substrate is exactly why I’m wary of teams that want to bolt probabilistic AI onto processes that actually need to be deterministic. Know which half of the problem you’re in. Most failures I’ve seen are really a category error about that one question.
Decide well, revisit honestly, respect the system underneath. That’s most of it.
4. Keep building, don’t just advise
The thing I’d least want to lose, as careers get more senior, is the habit of actually making things. It’s the easiest instinct to surrender and the most expensive one to lose.
So alongside the day job I keep shipping. I run mybizopz, a business-management platform aimed at the same kind of small and medium businesses I’ve spent years building for, this time in the UK. I coded my own site from scratch rather than buy a template. I run a home-server stack and a clutch of AI experiments that I treat less as products than as a way of finding out where the technology actually is, rather than where the keynote circuit says it is.
Operating and building, not only advising, is the part of the profile I’m proudest of, because it’s the part I have to keep choosing. The fastest way I know to have an honest opinion about a technology is to try to build something real with it and watch where it breaks.
Louise ended her piece with the chip shop: a long way behind me now, but the instinct I formed somewhere around that counter (decide well, change your mind honestly, respect the machinery, keep your hands dirty) is the part I never left there.
It’s a generous thing, to have someone else find the throughline in your work and hold it up so you can see it clearly. My thanks to the team at Leaders Loop for the profile. And if you want the long-form version of the story, with all the transcontinental detours intact, it’s over there.