01 — The journey
From £40 to $100m
Nine years. It started with a door, a shop floor and a garage.
Knocking doors
Field sales for Amber Windows at seventeen, finding my own leads by cold canvassing. Where I learned to handle an objection and close.
The shop floor
Two and a half years at Home Bargains on tills, deliveries and displays. Pricing, merchandising and impulse buying, learned in person long before I applied any of it online.
£40 and a garage
Founded NGNE with £40 in my bank account while studying law full time. No influencers, no sponsors, no investment. Just a small budget and Facebook ads learned fast. £325,027 in verified sales, with a best month of £40,149.
CMO at DeHub
Chief Marketing Officer from launch. Zero to over $100m in market capitalisation, a $1.4m budget, 100+ people through seven managers, and the first AI infrastructure I ever put into production.
RAPZ & First Class
Co-founded a premium fight brand with UFC featherweight Lerone Murphy, and a TikTok creator agency that grew from zero into one of the biggest in the UK.
Opening the gym
We opened a boxing and MMA gym in Essex. Ring, fit out and brand. The brands, the athletes and the room they train in, all under one roof.
Building with AI
Agent systems, analytics platforms and automation running in production. Not demos. Things people depend on every day.
02 — Deep dive
DeHub
Chief Marketing Officer, 2021 to 2024. An open source, user owned entertainment platform. I joined at the start and ran marketing through launch.
Zero to $100m
DeHub was built as an all in one entertainment and lifestyle platform spanning streaming, gaming, browsing and creator funding, with the audience owning a share of the economy rather than simply subscribing to it. Users could vote on what got funded, renewed or cancelled, and earn from their own engagement instead of only paying for it.
It was open source, peer to peer and borderless by design, which made the marketing problem unusual. We were not selling a product with an obvious shelf. We were explaining a new model to a global audience, in public, at speed, with everything we said visible to everyone at once.
I owned a $1.4m budget, held trusted access to company treasury funds in excess of $6m as part of the executive team, and ran a function of over 100 people through seven managers covering customer service, developers, video, design and community. The platform passed $100m in market capitalisation and won Corporate LiveWire's 2023 Innovation and Excellence Award during my time as CMO.
Explaining a new model
Nobody arrives understanding a user owned economy. Most of the work was translation: taking something structurally unfamiliar and making it land in thirty seconds, across languages, time zones and wildly different levels of technical knowledge. Seven managers each owned a channel and a discipline. My job was setting the standard and making sure it held everywhere at once.
Support at global scale
A worldwide audience means questions never stop arriving, in every time zone, through every channel we had. The obvious answer was to hire a large support team and keep hiring as we grew. Instead I built the infrastructure below, and the company never had to hire for a large portion of that work.
03 — AI infrastructure
Systems, not demos
What we actually built, why the first version failed, and where the machine is required to stop.
Real voice, not model voice
The first version worked and nobody trusted it. It was accurate, it was fast, and it did not sound like a person, which users detect even when they cannot name why. Better prompting never fixed it. We transcribed hundreds of thousands of real conversations, support emails and chat logs, then rebuilt the system on how our own people had actually written, including the parts that were not polished. Trust came back because the texture was right.
Grounded, not guessing
Answers are assembled from our own material rather than generated from whatever the model happens to hold. That is the difference between a system that is confidently wrong and one that is usefully right, and it is the single most important design decision in anything customer facing.
Where it must stop
Every system I build has a defined edge. Low confidence, sensitive subject, anything financial or personal, and it hands to a human rather than improvising. On the mental health charity build the entire design was this question, because a confident wrong answer to somebody in crisis is far worse than no answer at all.
Every channel, always on
Email, internal communications, Telegram and on platform support, running continuously across time zones. Onboarding automation handled the step that was quietly losing us the most users, generating what people needed rather than asking them to understand it. The measure was never accuracy in testing. It was whether people kept using it. They did.
04 — Paid media
£1.4m executed
Across my own brands and client accounts. Judged on contribution margin, not on the metric that flatters.
Creative is the targeting
The best performing thing I have run is athlete footage, shot and cut by me, that does not look like an ad. People watch it, engagement compounds, and distribution gets cheaper. Media cost falls because the content earns reach rather than buying all of it.
Then it is volume and discipline. A dozen variants of a winning idea, changing only the hook and the first frame, tested cheap on TikTok before real budget goes anywhere near it. Winners scale. Everything else dies quickly and without sentiment.
Underneath it, tracking built properly: Meta Pixel and Conversions API, GA4, Tag Manager, server side, so I read contribution rather than last click. ROAS flatters thin margin products and hides the ones that can genuinely take spend. Contribution per order is the number that decides.
Creative engine
Concepts get written, shot, cut and tested cheap before real budget goes anywhere near them. Most die. That is the process working, not failing. The discipline is killing things quickly and without sentiment, then putting everything behind the few that live.
The full funnel
Every stage is a place people leave. I own all of them, which is why the ad and the page are built as one piece.
05 — Selected work
Four businesses
Two I own, one I was CMO of, one I co-founded.
£40, a garage
and a shed
The first product was a sauna suit. No influencers, no sponsors, no budget to speak of. I taught myself Facebook ads properly, because there was nothing else. It did £30,000.
Then it stopped. Traction died and I could not find the next thing for weeks. What eventually found it was a Ryan Garcia video that made me think of a cobra reflex bag. I went straight to manufacturers, started the relationships that I still hold six years later, and tested the ads before I owned a single unit.
The tests worked, so I put every penny I had into the first order. £6,500, the entire account, on 80 bags. They sold out. The next order was 500, limited only by what fitted in a double garage and a shelter I put up in the back garden. Eventually 1,000 units at a time.
£325,027 in total sales between August 2020 and November 2021, with a best month of £40,149. Screenshots below, straight from the dashboard. Everything I now do at scale, I learned here with no margin for error.
The receipts
Shopify analyticsStraight out of the Shopify dashboard. £325,027.53 in total sales between August 2020 and November 2021, and £40,149.54 in December 2020 alone, up 82% on the month before. All of it out of a garage.
Yes, that is my business partner Mike Hales on the call in the corner, mid-conversation, entirely unaware he would end up on a portfolio five years later. Consider him the timestamp.
RAPZ Sports
A premium fight equipment brand I co-own with UFC featherweight Lerone Murphy and Mike Hales. Everyone else competes on price. We went the other way and built something that holds a £200 tag.
I own the whole commercial result. Store, paid media, creative, tracking, suppliers, stock. I shoot and cut the ads myself, and I rebuilt the product page around customer psychology rather than design taste.
First Class
An award winning TikTok live streaming agency and approved TikTok vendor, co-founded from zero with Mike Hales. We started as a sub agency, outgrew the business above us and went out on our own. I built the website and the AI tools it runs on.
NGNE
My first brand. £40, a garage and a full time law degree. The name, the logo, the sourcing, the quality checks, the store, the photography, the video and the ads. No investment, no agency, no team, no influencers.
Dr Recovery
Built an AI patient assistant trained on hundreds of thousands of the practice's own transcribed conversations, so it answers in their voice rather than a generic one. Added automated AI video production and rebuilt the social presence.
TrueLight
A site, an automated back end and a help system for people reaching out in urgent need, built for a mental health charity. The most careful thing I have made, and the clearest example of designing where a system has to stop.
06 — Client work
12+ businesses
Marketing, websites, AI systems and sometimes logistics. Athletes, clinicians, hardware startups and agencies. Some launched from nothing, some rebuilt.
Lab Zero
Precision cooling hardware for temperature sensitive protocols. Brand, store and go to market for the ChillBox Pro, sold across the UK, Europe and the US.
labzero.uk Creator economyFirst Class
Award winning TikTok live streaming agency and approved TikTok vendor, close to 1,000 creators. Co-founded from zero with Mike Hales. I built the website and the AI tools it runs on.
firstclassagencytiktok.com MedicalDr Recovery
Dr Imran Khan PhD. AI patient assistant trained on hundreds of thousands of his own transcripts, automated video production and a rebuilt social presence. Standing start to six figures a month.
dr_recovery_uaeThe gym, Essex
We opened a boxing and MMA gym in Essex. Full ring, equipment fit out and the brand around it. Physical space is a different discipline to a website, and it feeds straight back into what we make.
UFC & MMA athletes
Brand, content and commercial work for professional fighters, including partners at RAPZ. Athlete content that outperforms paid placement because it does not look like an ad.
Perfume, skincare, appliances
Paid media, stores and creative across consumer categories. Six businesses run beyond my own, more than 30 commercial websites built.
Logistics & sourcing
Where the marketing is not the problem. Supplier sourcing, landed cost, stock planning and fulfilment for clients importing at volume.
07 — In the field
The work, live
World champions, UFC fighters, our own fight shows, partnership gyms and international conferences.
08 — Supply chain
Won in the factory
The half of the job that never appears in a campaign.
China
Over 1,000 product lines built from nothing, working directly with more than 30 manufacturers rather than through a sourcing agent. Some of those relationships are six years old.
I have travelled to Shijiazhuang in Hebei Province three times to meet suppliers on their own ground, and that is where the real work happens. On one negotiation, done in person and built on a friendship rather than an email thread, I brought unit cost down by over 70%. Nobody gives you that over a video call.
Price and minimum order quantities, samples until the product is right, tooling, lead times, quality checked on arrival rather than trusted from a spec sheet, and the true landed cost once shipping, duty and returns are counted.
09 — Film
On camera
Ad creative, behind the scenes, and the places the work actually happens.
Ad creative
Ran on Meta & TikTokBehind the scenes
RAPZ product shootThe spaces
Essex & Punjab10 — Outside the work
Harjot Gill Farm
I own a farm in Punjab. We grow vegetables for the village, and I get back out there when I can.
There is a before and after of the land in the film section above. This is the tractor.
It is the only thing I do that has nothing to do with a number, and it keeps everything else in proportion.
Let's talk
Open to senior marketing, commercial and AI roles. Ad account results, store dashboards and further detail available on request.