From Urban Planning to Brand Strategy: The Non-Linear Journey of Evan Larbi

Evan Larbi studied Classics, discovered a late interest in buildings, took a masters that sat half in architecture and half in social science, qualified as an urban planner, worked for a futurist, took a couple of coding courses with no intention of becoming a coder, spent under a year inside a music hardware startup, and ended up as strategy director at a London digital product agency. He now runs Studio Junction, where he does brand and product strategy for mostly tech-driven companies. In this episode he talks to Rob Woodhead about why language sits at the centre of strategy rather than beside it, why AI output keeps landing in an undifferentiated middle, what it means to act as an external co-founder, and why he thinks brand work should be enjoyable.

Key takeaways

• Language is not the output of strategy, it is the mechanism of it, because a strategy nobody can articulate cannot be acted on.
• Companies across entirely different sectors arrive with the same underlying problem: who they are, what they stand for, and how to articulate their offer.
• AI output tends to land in a competent middle, which is a structural problem for brand work whose entire purpose is differentiation.
• AI is more useful as a counterpoint that surfaces lines of inquiry you had not considered than as a producer of finished work.
• An external strategist can be more honest than an internal one precisely because they have a stake in the outcome but not in the politics.
• Telling clients only what they want to hear is a failure of professional duty, not a form of good service.
• Non-linear careers accumulate transferable material, and the connection between the parts usually becomes visible only in retrospect.

Episode chapters

• 00:20  Two non-technical routes into tech
• 00:51  Classics, teaching, and a late interest in buildings
• 02:00  City Design and Social Science: bridging physical and social design
• 03:09  Doing the jobs at an architecture practice that were not the architecture
• 04:39  Discovering civic tech around the first iPhone
• 07:44  What a futurist actually does
• 10:57  Coding courses, freelancing, and a music hardware startup
• 13:13  Ostmodern, and the broadcasters moving online
• 15:14  What managing developers teaches you
• 18:05  Why running an agency in London is hard
• 19:10  From project manager to strategy director
• 23:45  Why strategy is really about language
• 25:38  Studio Junction and the problem every client shares
• 30:21  Being an external co-founder
• 34:54  Where AI lands for brand and narrative work
• 41:11  Designing for joy instead of transaction
• 45:38  Advice to a 23-year-old with too many options
• 48:58  If Studio Junction were a band

What does a brand strategist actually do?

Asks a lot of questions, mostly. Evan is disarming about the mechanics: the work is conversations, and the output is a set of decisions about what a company is and how it says so.

What he is looking for is what he calls a narrative system. Not a tagline and not a tone-of-voice document, but a core idea coherent enough that the language deriving from it holds up on a website, inside a product, in a sales conversation and on social media without anyone having to reinvent it each time. The test of a strategy is whether it ripples outwards consistently.

His clients have spanned education, deep tech, fintech, fashion and the third sector, and almost all of them are tech-driven in some form. What is striking is how little the sector changes the problem. The question of who we are, what we stand for, and how we explain our offer to the people we want to reach turns up in effectively the same shape every time.

The work, in his description, is untangling that complexity.

This is the same problem we run into at the start of most product work. You can see how we approach it and the products it has produced.

Why is language at the core of strategy?

Because a strategy that cannot be articulated cannot be executed.

Evan describes himself as a strategist and writer, and he is willing to be slightly unfair about it: he thinks it is difficult to do strategy well without understanding words. Not writing in a decorative sense, but the connections between words, the difference between voice and tone, and what a message actually means once it leaves the room.

The origin of this is worth noting. Before founding Studio Junction, he spent years at Ostmodern working on what the company sounded like, trying to bring the same rigour to language that the founders, who were designers, had always brought to design. The design side of digital had been taken seriously for a long time. The words had not.

Asked whether the Classics degree feeds into it, he is honest that the literature is long forgotten. What carried over was an interest in expression: how an idea gets expressed, and what a message means.

Can AI write brand strategy and positioning?

It can produce something competent, which for this particular job is the problem.

Evan’s observation is that AI output tends to arrive at a middle. Not bad, often perfectly serviceable, and near-identical to what everyone else is getting. It reads well for the first two minutes and then reveals itself as indistinguishable. For most writing tasks that is an acceptable trade. For brand and narrative work it is close to fatal, because the entire purpose of the exercise is differentiation, and differentiation by definition requires being different from the average.

He is not dismissive though, and he uses these tools himself. Where he finds them valuable is as a prod: producing a counterpoint, surfacing lines of inquiry he had not considered, giving him something to argue with. The distinction is between AI as a collaborator in thinking and AI as a producer of finished work.

On whether AI will eliminate agencies, his view is that pressure will fall hardest on undifferentiated ones. Agencies offering something genuinely distinctive, or a judgement that cannot be replicated, are in a different position, at least for now.

The same pattern shows up in code, where AI-generated output is frequently competent and still wrong in ways that only an expert notices. Our Vibe Code Audit exists for exactly that gap, and The Jekyll and Hyde of Vibe Coding covers it in more depth.

Should a consultant tell clients what they want to hear?

No, and Evan has held that view since early in his career, when a colleague argued the opposite and it struck him as obviously wrong.

His framing for the relationship is external co-founder. He works mostly with founders and senior leadership, on questions that sit above the day-to-day, which means conversations get to emotional territory quickly. People running companies often have things they find genuinely difficult and do not discuss with the people who work for them, and sometimes cannot discuss easily with a co-founder either.

The structural advantage is a particular kind of distance. He puts it as having skin in the game and no skin in the game at once. He wants the company to succeed, and he can be told to go away, and both of those facts make honesty cheaper for him than for anyone on the inside.

The practical version is unglamorous: telling someone the messaging makes no sense, or that they are trying to do too many things and need to focus on one. Anyone who has run a client relationship will recognise the difficulty, which is delivering that without it landing as an insult.

Why does joy matter in brand building?

Partly on principle and partly as a competitive position.

Evan’s view is straightforward: building a brand should be fun, and a strategy process that feels like wading through mud has gone wrong somewhere. He occasionally has to remind clients of proportion. The work matters, but nobody in the room is an oncologist.

The competitive argument is more interesting, and it connects back to AI. The experience of using these tools is overwhelmingly transactional. You ask, it answers. If that is what the default becomes, then experiences that are not transactional acquire a value they did not have before. He is deliberately interested in how you build joy into something, whether that is working with people, developing a strategy or producing a piece of writing.

He is careful not to claim certainty about where any of this goes. His comparison is computing in the 1950s: some people insisting it will do everything, others insisting it is a curiosity, and nobody actually knowing. The strategy he draws from that is to look for what AI does not provide, and work there.

How do you build a career without a linear path?

By not overthinking it, which is advice he acknowledges is close to non-advice.

Asked what he would tell his younger self, Evan describes the vertigo of being in your twenties with too many options and no way to evaluate them. He mentions a book making the counterintuitive point that having fewer options in middle age is a relief rather than a loss, because abundant choice is genuinely hard to handle.

His practical position is that tangents come back. Do something, and the dots connect later, even when at the time it looks like drift rather than direction.

In his own case the thread only became visible in retrospect: he was always the words person wherever he worked, and eventually built a company around that without ever having set it as an intention. Rob makes the same point from his own non-linear route, that changing direction does not waste the experience, because most of it transfers.

About the guest

Evan Larbi is the founder of Studio Junction, a brand and product strategy practice working mostly with tech-driven companies across education, deep tech, fintech, fashion and the third sector. He was previously strategy director at the digital product agency Ostmodern, and before that a qualified urban planner. He describes himself as a strategist and writer.

About the host

Rob Woodhead is a co-founder of Old St Labs, a London-based software agency building web and mobile products for startups and innovators. He and Evan are long-standing friends and both came into technology from non-technical backgrounds.

Frequently asked questions

What is brand strategy?

Brand strategy is the work of deciding what a company is, what it stands for, and how it articulates that to its audience. Evan Larbi describes the output as a narrative system: a core idea coherent enough that consistent language can be derived from it across a website, a product and everything else.

Can AI write brand positioning?

It can produce competent, generic positioning. The difficulty is that brand work exists to differentiate, and AI output tends to converge on the same middle ground everyone else receives, which makes it poorly suited to the core task.

How should strategists use AI?

As a counterpoint rather than a producer. Evan Larbi uses it to surface lines of inquiry he had not considered and to give himself something to argue with, while treating finished output as a starting point rather than a deliverable.

What is an external co-founder?

A consultant close enough to a founder to discuss genuinely difficult questions, while remaining outside the company. The distance allows a level of honesty that is harder internally, because they want the business to succeed but are not implicated in its politics.

What does a futurist do?

Futures research constructs plausible scenarios about how the future might unfold rather than predicting it. Organisations use it to understand how political, social or technological change could affect them, a discipline with roots in post-war efforts to rebuild societies.

‍