Shivam Agarwal

Quit my A$160k tech job. Now building LicenceMate.

Sharing how to test business ideas, leave a job with less risk & grow a business.

[Shivam Agarwal @shivamA2k]

I had a comfortable A$160k tech job. I didn't even want to quit.

So I tried starting a business without leaving my job first.

That business eventually died. The next one now does A$100k+/month in bookings.

Here's how it happened, what worked, and what I'd do again.

1. I started with A$200 and a pretty basic idea.

I was working as a software engineer, mostly doing backend stuff, and was really interested in crypto at the time.

A friend and I came across an opportunity to build a crypto trading bot.

We built a very basic version and spent around A$100–200 on Telegram ads to see if anyone wanted it.

Got 200–300 signups, some paying customers, and eventually made around A$10k in trading fees.

Not exactly enough to quit a $160k job.

But people were actually paying for something we'd built, and that was interesting.

2. I didn't quit my job straight away.

I actually liked my job. It paid well, wasn't too stressful, and I had no real reason to leave.

So I negotiated a 3-month break instead.

I figured I'd spend those months working on the business full-time and see where it went.

The business grew. Customers were asking for new features, and we could see how building them might make us more money.

We still weren't replacing my salary.

But the opportunity looked big enough that I eventually decided not to go back.

It was a risk, obviously, but at least we had real customers and some revenue before taking it.

3. Then we had to shut the business down.

After around 6–8 months, things went wrong and we had to close it.

Not going into the details here, but it was pretty rough.

By then we were three co-founders, and fortunately we'd made enough money to give ourselves roughly a year of runway.

So we started looking for the next opportunity.

And the next one came from somewhere I really wasn't expecting.

4. My girlfriend's dad was a driving instructor.

He was running a small driving school in South Australia.

My girlfriend and I started helping him with his website and Facebook ads.

Nothing particularly complicated.

But we ended up growing his income by roughly 2–3x.

The interesting part was how little it cost to get a customer compared to how much that customer would spend.

So I brought my other two co-founders in, and we started looking at the industry properly.

How many people learn to drive every year? How many lessons do they take? How much do they spend?

We used public data, competitor prices and the customer acquisition numbers we'd already seen to estimate how much money a business like this could make.

Then worked backwards to figure out how much market share we'd need for it to be worth doing.

It didn't look like some obvious billion-dollar startup.

But the numbers looked pretty good, and we already had some proof that we could get customers.

That was enough to get us interested.

5. We started selling before building the product.

One problem we found was that booking a driving test in South Australia was unnecessarily difficult.

You had to call around different examiners, ask about availability, and hope someone had a slot.

So we started running Facebook lead ads and getting people on the phone.

We already knew a few examiners through my girlfriend's dad.

We'd find a customer, connect them with an examiner, and handle the booking manually.

No fancy platform.

We were making sales before we'd built the actual software.

Eventually we built LicenceMate, where people can search their postcode, see available test slots and book online.

6. Where we are now.

LicenceMate is doing A$100k+ in monthly bookings and makes enough to sustain me in Sydney.

That's booking volume, not profit. There's a pretty big difference.

Our driving test booking business works well, but we're still trying to crack the driving lessons side.

We're experimenting with better offers, progress tracking and guarantees to help students know how much training they actually need.

The goal over the next 12 months is to make the lessons business work at scale and start expanding across Australia.

No idea how smoothly that's going to go, but that's what we're working towards.

Looking back, I think the biggest things I've learned are pretty simple.

You don't have to quit your job to test if a business idea works.

You don't have to build an entire product before getting customers.

And you don't necessarily need some crazy new idea to build a business that makes money.

I'm still learning all of this myself, especially the scaling part.

I'm going to share more about our actual numbers, how we get customers, experiments we're running, mistakes we're making, and what I'm learning about building profitable businesses.

If you're thinking about leaving your job to start a business, or already trying to grow one, feel free to follow along.

Hopefully some of the things I share can save you a bit of time and money.

[Shivam Agarwal @shivamA2k]

@ishubhamsakhuja Poker is good

[Shivam Agarwal @shivamA2k]

@AmanShips Review count is not that high but we have more recent reviews and better review velocity

[Shivam Agarwal @shivamA2k]

Is SEO dead?

Because I'm starting to think 90% of SEO advice on the internet should be.

I've been trying to figure out SEO for my business lately and holy shit, the amount of noise is insane.

We run LicenceMate, a driving education business in Australia.

The problem I'm trying to solve is actually pretty simple.

When someone searches for driving lessons in Adelaide, I want us to show up on Google, Maps, or even ChatGPT.

And I want that person to book with us.

That's literally it.

So I started looking into how to make it happen.

And now I have about a thousand things I apparently NEED to do.

Technical SEO, backlinks, content, local SEO, topical authority, AI search optimisation, programmatic SEO, the list never ends.

Every expert seems to have a different opinion on what matters most.

And the worst part is the FOMO.

You read an article from a month ago and wonder if it's already outdated.

Someone on X posts a new AI SEO workflow yesterday and suddenly you feel like you're missing out by not using it.

Then someone else comes along with an even better one.

At some point I realised I'm spending more time researching the BEST way to do SEO than actually doing anything.

And I think AI is making this problem so much worse.

Because now you can automate practically everything.

But just because you CAN automate 100 things doesn't mean those 100 things need to be done.

I'm starting to think we're getting insanely good at doing things faster without questioning whether those things matter at all.

What if 3 boring changes to our website and Google presence bring us more customers than 50 AI-generated articles?

What if most of the things I'm being told to do have almost no impact on our business?

I don't know the answer yet.

But I think I've been approaching the problem backwards.

Instead of asking how I can do all these things faster, I should be asking which things I can completely eliminate.

And then go really deep on the few things that actually matter.

I'm going to document this experiment as I go.

What advice I ignore. What I choose to implement. How much work it takes. And what actually moves the needle.

Not just rankings or traffic.

Actual bookings.

Because the more I think about it, the more I feel like the biggest advantage in the AI age isn't automation.

It's elimination.

[Shivam Agarwal @shivamA2k]

@AmanShips showing up but not in top 3

[Shivam Agarwal @shivamA2k]

@AlDev0 I am not sure what that means, but will figure out

[Shivam Agarwal @shivamA2k]

@AdrianDittmann 0*100=0

[Shivam Agarwal @shivamA2k]

@DevMarDe https://t.co/GrzcooyQws
https://t.co/65YVVBQi5r

[Shivam Agarwal @shivamA2k]

@imfat I WILL SUCCEED CAUSE I WILL DO WHATEVER IT TAKES!

[Shivam Agarwal @shivamA2k]

@MidcoastWebCo Thanks mate and you’ve identified the major problems. The bigger problem is getting more students in the door

[Shivam Agarwal @shivamA2k]

I had a comfortable A$160k tech job. I didn't even want to quit.

So I tried starting a business without leaving my job first.

That business eventually died. The next one now does A$100k+/month in bookings.

Here's how it happened, what worked, and what I'd do again.

1. I started with A$200 and a pretty basic idea.

I was working as a software engineer, mostly doing backend stuff, and was really interested in crypto at the time.

A friend and I came across an opportunity to build a crypto trading bot.

We built a very basic version and spent around A$100–200 on Telegram ads to see if anyone wanted it.

Got 200–300 signups, some paying customers, and eventually made around A$10k in trading fees.

Not exactly enough to quit a $160k job.

But people were actually paying for something we'd built, and that was interesting.

2. I didn't quit my job straight away.

I actually liked my job. It paid well, wasn't too stressful, and I had no real reason to leave.

So I negotiated a 3-month break instead.

I figured I'd spend those months working on the business full-time and see where it went.

The business grew. Customers were asking for new features, and we could see how building them might make us more money.

We still weren't replacing my salary.

But the opportunity looked big enough that I eventually decided not to go back.

It was a risk, obviously, but at least we had real customers and some revenue before taking it.

3. Then we had to shut the business down.

After around 6–8 months, things went wrong and we had to close it.

Not going into the details here, but it was pretty rough.

By then we were three co-founders, and fortunately we'd made enough money to give ourselves roughly a year of runway.

So we started looking for the next opportunity.

And the next one came from somewhere I really wasn't expecting.

4. My girlfriend's dad was a driving instructor.

He was running a small driving school in South Australia.

My girlfriend and I started helping him with his website and Facebook ads.

Nothing particularly complicated.

But we ended up growing his income by roughly 2–3x.

The interesting part was how little it cost to get a customer compared to how much that customer would spend.

So I brought my other two co-founders in, and we started looking at the industry properly.

How many people learn to drive every year? How many lessons do they take? How much do they spend?

We used public data, competitor prices and the customer acquisition numbers we'd already seen to estimate how much money a business like this could make.

Then worked backwards to figure out how much market share we'd need for it to be worth doing.

It didn't look like some obvious billion-dollar startup.

But the numbers looked pretty good, and we already had some proof that we could get customers.

That was enough to get us interested.

5. We started selling before building the product.

One problem we found was that booking a driving test in South Australia was unnecessarily difficult.

You had to call around different examiners, ask about availability, and hope someone had a slot.

So we started running Facebook lead ads and getting people on the phone.

We already knew a few examiners through my girlfriend's dad.

We'd find a customer, connect them with an examiner, and handle the booking manually.

No fancy platform.

We were making sales before we'd built the actual software.

Eventually we built LicenceMate, where people can search their postcode, see available test slots and book online.

6. Where we are now.

LicenceMate is doing A$100k+ in monthly bookings and makes enough to sustain me in Sydney.

That's booking volume, not profit. There's a pretty big difference.

Our driving test booking business works well, but we're still trying to crack the driving lessons side.

We're experimenting with better offers, progress tracking and guarantees to help students know how much training they actually need.

The goal over the next 12 months is to make the lessons business work at scale and start expanding across Australia.

No idea how smoothly that's going to go, but that's what we're working towards.

Looking back, I think the biggest things I've learned are pretty simple.

You don't have to quit your job to test if a business idea works.

You don't have to build an entire product before getting customers.

And you don't necessarily need some crazy new idea to build a business that makes money.

I'm still learning all of this myself, especially the scaling part.

I'm going to share more about our actual numbers, how we get customers, experiments we're running, mistakes we're making, and what I'm learning about building profitable businesses.

If you're thinking about leaving your job to start a business, or already trying to grow one, feel free to follow along.

Hopefully some of the things I share can save you a bit of time and money.

[Shivam Agarwal @shivamA2k]

@pcshipp Yeah like why? Use Vercel/Railway

[Shivam Agarwal @shivamA2k]

Running an AI model locally doesn't mean you own your intelligence.

I think a LOT of developers are going to spend the next year self-hosting models and still end up locked into someone else's AI stack.

Here's the part I think people are missing.

Imagine you've been running an AI agent for your business for 6 months.

It knows your customers, has access to your tools, remembers previous tasks, knows which actions need approval, and you've spent months fixing all the weird edge cases.

Now imagine you want to switch from Claude to Mistral.

If you lose the agent's memory, execution history, tool permissions, or the tests that tell you whether it still works, what exactly did you own?

The model is only one part of the system.

And ironically, it might be the EASIEST part to replace as open models keep getting better.

I actually think hybrid AI makes the most sense for most businesses.

Run cheap or local models for the easy stuff. Use frontier models when you need them.

But keep the agent's memory, tools, logs, permissions and evaluations somewhere YOU control.

Because the real lock-in might not be which model you're using.

It might be the thousands of hours your business spends teaching an agent how to actually get things done.

And I don't think you want that locked inside someone else's product.

[Guillermo Rauch @rauchg]

Own your intelligence.

Anyone that opposes this has something to sell you (perhaps your own intelligence back to you.)

Local models need to continue proliferating. The “absolute unit” massive models may be out of reach for now locally, so your best option is to, indeed, keep your options open. Own your harness, data, connections, runtime: your agents. Embrace open source.

[Shivam Agarwal @shivamA2k]

@DanielPriestley yeah these people are crazy, no wonder no successful person wants to be in the UK

[Shivam Agarwal @shivamA2k]

@ausstockchick Whats the alternative, whats the dominant strategy?

[Shivam Agarwal @shivamA2k]

@girdley You wanna kill the moving companies haha

[Shivam Agarwal @shivamA2k]

@DanielPriestley How's that unique to UK?

[Shivam Agarwal @shivamA2k]

@almokhtar_al Cybersecurity is the biggest industry alongside marketing

[Shivam Agarwal @shivamA2k]

I wanna meet more people on X who are genuinely trying to make money by building a business.

Not necessarily the next billion dollar startup. Just people trying to build something that actually makes money.

Could be a SaaS, a cleaning company, an agency, a local business, literally anything.

A bit about me:

I quit my A$160k software engineering job to start a business.

My first business was in crypto. Made some money, things went wrong, had to shut it down.

Then started LicenceMate, a driving education business in Australia.

We're now doing A$100k+ a month in bookings, and it makes enough to support me in Sydney.

Still a lot of stuff we haven't figured out.

And honestly, the more I work on this, the more interested I get in how other people actually make money.

How did you find your first customers? How did you know your idea was worth pursuing? How do you decide where to spend money? When do you know it's time to quit your job?

I feel like there's so much to learn from people building completely different kinds of businesses.

I'd especially love to meet people who think about risk, are willing to put in the work, and care about building something profitable rather than just looking successful online.

Doesn't matter if you're making $0 or $1M a month.

If you're building something, or seriously trying to get started, tell me what you're working on.

Would love to follow along and learn from more people here.

Let's connect.

[Shivam Agarwal @shivamA2k]

@MarvellousDev Lets do it!

[Shivam Agarwal @shivamA2k]

@frankezecode_ Done!

[Shivam Agarwal @shivamA2k]

@oAnakin9 Sounds good man, followed you. Is your selling B2B?

[Andrew Gazdecki @agazdecki]

Some founders are going to make millions over the next five years solving problems that everyone else thinks are too boring.