FIELD REPORT / BUSINESS PETE DEVKOTA
Business

From $5 an Hour Washing Dishes to Winning Global Clients From Modbury: My Grow In SA Interview

Pete Devkota on Grow In SA: the dishwashing years, winning overseas clients from Modbury, why AI lies, and building a business that runs without you.

Table of contents

I sat down with John Kiama on Grow In SA, a weekly podcast about the people building South Australia’s businesses. John and I have known each other for almost ten years, mostly over the phone, and this was the first time we’d been in the same room.

My Checkout Champ interview was all tactics. This one is the other half: how a kid reselling secondhand books in Nepal ended up washing dishes in Adelaide, and then winning ecommerce contracts against 100-person agencies from Modbury. We also got into how to find clients overseas when you live in the suburbs, why most people get AI wrong, and why you need a business that runs without you.

The full episode is below. If you’d rather read, these are the parts I’d want you to take away.


The jobs that taught me more than the degree

I was born in Nepal and raised between Nepal, Thailand and New Zealand, because both my parents are university professors and their scholarships took me along. I went to China to study aeronautical engineering, and my first real business came straight out of uni: an investor put US$100,000 behind a home solar company in Nepal, at a time when the country was going without power for up to 16 hours a day. I ran it for about six months and it got to US$20K a month. It also showed me what red tape, corruption and a long wait between paying for inventory and collecting the cash can do to a good idea.

I came to Adelaide for a master’s in engineering management, and went from having an office to earning $5 an hour washing dishes. It took hundreds of applications to get even that job. The shifts ran 12 to 14 hours.

Those shifts, and the years I spent in aged care, taught me things an engineering degree didn’t: be tenacious, be humble, stay hungry. I took the aged care work partly to give something back to the people who built the city I’d chosen to live in, and partly to learn from people in their 80s and 90s. It’s why I’ve told my wife I’d love our kids to spend six months in a kitchen or an aged care home.


Sell the outcome. The logo can wait.

Eventually I burnt out and hurt my back, and my wife pointed out that I’d always been good at teaching. So I started a home tutoring agency with no ad budget: a fresh Gumtree ad every day, a free first lesson, a two-hour minimum per session, and free classes for families who referred someone. That grew to $100K a year with no office, just me and a Toyota Corolla covering 200 kilometres a day. I sold English exam packages on the side, and a Chinese builder once paid $3,000 for a daily phone call so he could practise his English. Everything is sales.

Then came the email business. In late 2017, after leading another agency, I quit an $80K job six months before my son was born, with a mortgage, a new car and no income lined up. I had no website, no logo and no office. The business was called Pete Devkota Consulting. What I did have was a clear picture of who I wanted to work with, one specific problem to solve for them, and an offer built so the fee came back at least tenfold in attributable income, with clean data to show it. I collected half a million dollars before I paid a designer in England $500 for a logo.

That order matters. Founders lose months to names, fonts and company structures. Early on, only a few things count: whose problem you’re solving, whether they can pay, and whether you can deliver. Nobody is buying your logo.


How to win overseas clients from the suburbs

About 95% of the clients I work with are outside Australia: the United States, the UK, Dubai, Vietnam. John wanted to know how that happens from Modbury. Two things.

Guanxi. The most valuable thing I brought home from China wasn’t the degree. It was this word, which roughly means who you know. Business comes down to people solving problems for people, so you have to get inside the circles of the people you want to serve, and you get there by giving first.

Distribution before content. Everyone obsesses over what to post and forgets to ask who will see it. My Sniper Method starts the other way around: fill your network with your ideal clients first, then publish. I did it on Facebook, because nobody expects to be marketed to there, and I let its suggested connections do the targeting for me. At my peak I was spending up to 10 hours a day on it: starting conversations, solving problems, and making an introduction when someone else was the better fit.

I think of it as planting a thousand seeds with no expectation attached. People can feel it when you want something from them right now. But keep showing up with value and they’ll eventually click through to your profile, and if that profile shows who you help, with case study after case study, a cold contact has become a warm lead. Time is also the one resource I can’t get more of, so I choose who to invest it in the way I’d pick a horse for the Melbourne Cup. I do the research before I place the bet.

That, plus an obsession with being the best in one vertical, is how I’ve outbid and outperformed 100-person agencies in Boston and New York.


AI will lie to you. Use it anyway.

I paid for ChatGPT the day it launched in Australia, decided it was rubbish, and kept coming back every few months as it improved. I’ve since put more than 10,000 hours into these tools, including ChatGPT, Claude, DeepSeek and Manus, pushing them to find where they break. I never tied myself to one. The tools will keep changing, so my bet was on becoming a better operator.

What those hours taught me: AI isn’t actually intelligent yet, it’s built for the average user, and it will tell you things that aren’t true. If you’re an expert in your field, you’ll find its limits quickly. Put Lewis Hamilton in my car and me in his, and he’d still beat me every single time. Master the craft and the tool amplifies it. We’ve written up the email version of this in our guide to AI email marketing.

Your clients won’t care much either way. A business owner wants someone to take a problem off their plate and own the result.


Know your unit economics

In ecommerce, the cost of acquiring a customer has gone from around $2 to $70 or $80. We feel it too: across AEO, SEO, GEO and Google Ads, we’re seeing about $1,500 to book one qualified call. You can only absorb numbers like that if you understand your unit economics, your customer lifetime value and the market you serve. Everyone has access to the same AI tools now, so judgment and domain expertise are what’s left to sell.


Build a machine that runs without you

John closed by asking what I’d tell someone trying to grow beyond South Australia. Start with whether you actually have a system. A system is a machine that works without you: input goes in, output comes out, and it’s consistent enough to productise. That means SOPs for acquisition, fulfilment and client satisfaction. Without them you’re on the hamster wheel, and it took me almost a decade to really learn that.

Once a system exists you can measure it, and every system has inefficiencies because humans run it. So look for the biggest constraint. If you could move one lever this year to hit your targets, which one is it? It’s the same thinking behind the automated flows and the Profit Loop we build for clients.


Stop tying your self-worth to your net worth

The last point is the one I most want people to hear. Money is a tool and an amplifier, exactly like AI. It isn’t the scorecard. I’m grateful to be alive, and for my family, my friends and my clients, and none of that is tied to my net worth. Attach your self-worth to that number and you start down a long, dark spiral.

Be careful with consumption too. For years I absorbed everything I could read and listen to, until I realised all that absorbing was delaying my success. Now when I hit a problem, I look for the fastest way to solve it and I execute.

My father told me it doesn’t matter what you do, just try to be in the top 1% of it. I still run my decisions through that filter. By my own measure I’ve done maybe 1% of what I want to do.

Thank you to John and the Grow In SA team for the conversation. If you run an ecommerce brand and want to find the constraint in your own email program, talk to us.

Share this post

Add us as a preferred source on Google

Shows our articles higher in your Google results. Takes one click.

WhatsApp