How My Business Startup Failed — And What I Learned

By August 2018, I was sitting across from three of my team members, about to deliver the worst news of my professional life. After months of watching our revenue plummet and paying salaries from my own dwindling savings, I had to let them go.

It was the culmination of one of the most brutal business failures I've ever experienced. But that failure became the most important lesson I've ever learned about building resilient businesses in the age of AI.

The Fragility of Success

In 2017, alongside my main YouTube project, I was running "Pretty Good Gaming," a YouTube gaming news channel that had exploded in popularity. After a key collaborator, Mike, joined the team, we found a formula that just worked.

For me, it went from a small side project to generating 2.6 million views in a single month. At our peak, we hit 6.8 million monthly views, had a team of five people, and were generating over £12k per month.

Almost overnight this side project became the main project and my main revenue earner. Life was good.

We were following the classic playbook: success means you scale, and scaling means you hire people.

The Comedown

Then, almost overnight, everything changed. In December 2017, it felt like YouTube flicked a switch. The algorithm no longer favoured our content and the views plummeted. By January, our revenue had dropped by 40%.

The success we built by adding more people to handle the manual tasks of research and scriptwriting was suddenly, terrifyingly, unsustainable.

The months that followed were incredibly stressful. The stable revenue was gone, but the payroll for five staff members was not. I was paying my team out of my own pocket, watching my personal finances drain away while trying to keep the ship afloat. The situation took an immense toll on my personal and family life.

By August, my wife and I had to face the inevitable. I made the hardest decision of my life and had to let three members of my team go. Soon after, my main partner on the channel also decided to move on.

I didn't blame any of them. The creative grind and financial pressure had worn us all down. This was the hard lesson in scalability: our model, which was entirely dependent on scaling human hours to scale output, was fundamentally broken. It couldn't withstand a shock.

A System, Not Just People

For years, I carried the weight of that failure. But in the time since, I embarked on a two-year deep dive into AI and automation, learning everything from prompt engineering to workflow automation tools like Zapier and n8n, to advanced AI assistants for research and content generation.

That's when the epiphany hit me. Looking back at Pretty Good Gaming through this new lens, I realized the very tasks that created our bottleneck—the constant research, the analysis of trending news, the writing of first-draft scripts—were all functions that could be brilliantly supported, if not fully handled, by specialized AI assistants.

I now imagine a different reality.

What if, instead of five people in a room, we had one or two creative leads armed with an automated system? What if we had AI tools like Perplexity for research, Claude for script drafts, and automated workflows that could scan sources, summarize key points, and generate data-rich briefs, all delivered into our inboxes before we even arrived at the office? Operating like a team of tireless junior researchers doing the heavy lifting, freeing up the humans to do what they do best: provide the analysis, the personality, and the final creative spark.

In this imagined reality, might things have turned out differently?

A Thought Experiment

Let me be specific about what this would have looked like in practice.

Our biggest operational bottleneck was the daily research cycle—scanning 20+ gaming news sources, identifying trending stories, fact-checking claims, and creating research briefs for our writers. This took two full-time researchers about 6 hours daily and cost us roughly £3,000 per month.

An AI-powered system could have automated 80% of this: web scraping tools gathering articles, AI models like Claude analyzing and summarizing key points, sentiment analysis identifying viral potential, and automated workflows delivering ranked story briefs to our creative team each morning. The entire process could run for less than £200 monthly in tool costs.

Instead of two researchers, we'd have needed one person to review AI-generated briefs and make editorial decisions—turning a £3,000 monthly fixed cost into a £200 variable cost plus one strategic role. When YouTube's algorithm changed, we could have weathered the revenue drop instead of drowning in fixed expenses.

I now wonder if my old business didn't fail because the idea was wrong. Maybe it failed because the operational model was fragile, and the tools to build a better one didn't exist for me yet.

A New Kind of Operational Leverage

Business leverage traditionally falls into two categories: financial leverage (debt) or human leverage (a big team). Both come with significant cost and risk, as I learned the hard way.

AI and automation have created a third, more powerful option: digital leverage.

This is the essence of my "Elevation over Replacement" philosophy. It's not about getting rid of people. It's about building a robust operational engine that doesn't require a new salary for every new unit of output.

It's about taking the input → process → output loop that drives every business and supercharging the "process" part with tireless, efficient digital collaborators. This elevates the entire business, making it more resilient, more scalable, and less fragile.

It allows human talent to be focused on high-value work like strategy, quality control, and creativity, rather than the manual processing that can sink a budget.

Redeeming the Past

Starting my new business, Acodi, wasn't just a business decision—it was a personal mission born from that painful experience. Through Acodi, I help founders and small business owners implement AI-powered systems that create this kind of digital leverage, building the lean, resilient operations I wish I'd had years ago.

Having to dismantle a team is a crushing weight, and it's part of the reason I'm here now, trying to provide support for people faced with similar challenges. It gave me a deep conviction that there is a better way to build. It drove me to develop the skills to create the kind of efficient, resilient business I wish I could have built all those years ago.

My goal now is to share that capability. I want to help other founders, creatives, and small business owners use these new tools to build their dreams on a foundation of concrete, not sand.

That old failure gave me the perspective and the passion to do what I do today. It's the driving force behind my mission to ensure others can build a future where their success is sustainable, scalable, and secure.

A Cautionary Tale

My story is a cautionary tale for every business owner who has found a measure of success and now feels the pressure to grow.

You're at a crossroads: you have too much work and not enough people to do it. The default decision, the one every business book preaches, is to hire more staff.

My advice is to stop. Take a minute. Before you tie your company's future to the fixed, recurring cost of new salaries, evaluate your options. Modern technology presents a different possibility.

Think about the work that needs doing. Instead of delegating tasks to new employees, imagine overseeing an efficient system that generates value for you. A system where you, the human in the loop, are always on the pulse of the operation, guiding the strategy, not managing payroll.

As a business owner, it's always prudent to evaluate all your options. This one could be the most important you ever consider.

Finding The Right Balance

I'm not suggesting you should never hire—human talent remains irreplaceable for strategy, relationship building, and creative leadership. There are absolutely times when the right hire transforms a business. But I am suggesting that in 2025, "hire first" shouldn't be your default response to capacity constraints.

The question isn't whether to scale, but how to scale smartly. Sometimes that means hiring brilliant people. Sometimes it means building intelligent systems. Often, it means doing both in the right sequence—using AI to handle repeatable processes first, then hiring humans for the uniquely human work that remains.

The key is making this choice deliberately, with full awareness of your options, rather than defaulting to the traditional playbook that sank my business.