The Overpressure Podcast's latest episode features Wes, an Australian digital agency owner of over 20 years who builds websites for trades, construction, and industrial brands out of Melbourne. He spent the business's first decade taking on any client for any reason, learned that trying to please everybody pleases nobody, and eventually niched down after noticing his best clients kept looking strikingly similar to each other.
The best material isn't the niching lesson, though it's a hard-won one. It's the story behind it: a brilliant lead developer growing visibly bored, Wes sensing he might quit, and pitching him a hopelessly complex dating-website project outside the agency's wheelhouse entirely purely to keep the guy engaged. The developer said yes, a proposal went out, the work was won, and a few weeks into a months-long build, he quit anyway, leaving Wes to deliver the whole thing himself through lost sleep and lost money. It's also the admission that followed years later: a controlling streak that kept him from inviting his own team into decisions about how the agency operated, until three or four years ago when he finally opened up the standard operating procedures to their input and watched profitability jump. And it's the mental health thread running underneath the business one a psychologist, breath work, a cold plunge, a sauna blanket, and the blunt observation that he could track the agency's financial improvement against the exact point he started going to the gym.
Quick gems from the episode:
→ Niching down isn't just about focus, it's about speed. Once you've seen the same scenario enough times in one industry, you know how to win fast instead of reinventing the wheel per client.
→ As AI makes the technical build cheap and abundant, the value shifts to guidance: helping someone choose the best option out of too many, rather than just building the thing.
→ The old marketing funnel treated the customer as the endpoint. Wes now treats the customer as the starting point turning finished client projects into case studies, reviews, and story content that only that client could produce, since it can't be replicated by AI-generated filler.
→ Trust has eroded so thoroughly that third-party proof PR, reviews, endorsements is becoming the most valuable marketing asset a business can have, precisely because it's the hardest thing to fake.
→ Large language models are already pulling from media coverage and customer reviews when someone asks about a company. A bad client experience that gets picked up doesn't just cost that client, it can shape what AI tells the next hundred people who ask.
→ One new client turned out to have built commercial landmarks recognizable well beyond their region work they'd been sitting on for years without realizing its marketing value, simply because they were too close to it day to day.
→ Paying for a gym membership he didn't want to waste turned out to be enough of a lever to change his health, which changed his business results proof that the fix doesn't have to be complicated to work.
The thread connecting a costly project taken on for the wrong reasons, a controlling instinct he had to unlearn with his own team, and a business built on human guidance in an increasingly automated industry is what makes this conversation land not as an agency growth story, but as an argument that the hardest lessons in business are usually about people, not technology. Check out the Overpressure Podcast for the full conversation.