4 Seasons in 1 Day: Why Melbourne Founders Should Stop Finding PMF

If you've lived in Melbourne for more than a week, someone has told you about four seasons in one day. You leave Fitzroy in a T-shirt, cross the Yarra in a downpour, and reach Richmond wishing you'd brought a jacket. Nobody here treats the morning forecast as a promise.

Yet almost every month I meet founders who treat product-market fit as the one forecast that never changes. They found it once - in a pilot, a launch, a strong quarter - and now they're building as if the weather will hold.

It won't.

Product-market fit is not a place you arrive at

Startup folklore frames fit as a destination: search, pivot, find it, scale. In practice, fit is a relationship between what you offer and what someone will exchange for it, and both sides move. In 2026, a competitor can rebuild your feature set faster than you can close your next round.

The founders who get blindsided usually did find fit. They just stopped looking and spent eighteen months scaling a value exchange their market had quietly renegotiated. Signs you've stopped looking:

Your roadmap is mostly what your best customers asked for last year.
Churn gets explained away as "wrong-fit customers".
Nobody can say what customers would pay for if you disappeared tomorrow.

Why this bites harder down here

With a home market of under 30 million, many Australian founders find early fit locally: a few anchor customers, a government contract, a loyal niche. That fit is real, but it's local weather. Push into the US, the UK, APAC or the Gulf and you meet different buyers and a different idea of value. I've watched teams treat expansion as a distribution problem when it was a fit problem wearing a sales costume.

And much of the most interesting Australian work - the energy transition, AI-native services, deep tech from our universities - sits in markets still being invented. I call these liquid markets. They never finish becoming, and they punish anyone who builds as if they have.

The step most founders skip

Across 25 years, one pattern keeps showing up. Founders are good at sensing change and good at moving fast. What they skip is deciding.

They notice a segment cooling or an AI tool eating a workflow they charge for, then jump straight to a new feature, a pivot or a pilot. But sensing plus moving, without deciding, is just reacting. Deciding means answering a harder question: not "how do we add AI to the product?" but "what can our customer now get from us that they couldn't before - and what will they pay for it?"

What it looks like in practice

I learned this firsthand. In 2016 I co-founded a SaaS platform for coworking spaces. We found fit, and by 2019 had to admit it wasn't the business to keep building. We pivoted into PropTech and became HAPLO reOS, where I still serve as fractional CPO. The pivot wasn't the hard part; accepting that fit we'd earned wasn't fit we'd keep was.

A training company that went deep, not wide. During COVID I was fractional CPO for a corporate training company doing around $1 million a year, almost all in person. The instinct was to rush online. Many training businesses did and discovered clients had never paid for content - they paid for an experience that didn't survive the webcam. So we spent a few weeks on strategy before building anything:

Sense - We surveyed clients and participants on what made the programs work: not the syllabus, but the reasons they stayed.

Decide - The founders made one call: online had to carry the same energy as the room. That became the filter for every product decision.

Move - We built a platform that mirrored their in-person delivery, not a generic video portal.

Repeat - Feedback lived inside the learner's workflow, so improvement became part of how the product ran.

By the end of 2025 they were approaching $5 million from online offerings - five times where they started - without going global. The product worked because it matched how people in that country learn and engage, and that doesn't travel by default. Depth became their moat. Going global isn't the automatic next step after fit; sometimes the stronger strategy is knowing your home market so well that no imported product feels as right.

A card scanner that sold the follow-up. Business card scanning is about as crowded as categories get. When I started with this company, their product was a WhatsApp bot conference hosts offered attendees: scan a card, save the contact.

The Liquid Ocean exercise changed their question from "how do we scan better?" to "what is an attendee actually trying to do?" Nobody wants a contact. They want the conversation to continue, yet the card usually ends up in a drawer. Today, one scan saves the contact with company and specialisation, sends an automatic introduction email, and adds the lead to a Google Sheet where you can mark it Hot, Warm or Cold. The engagement starts while the handshake is still warm.

The result: more than 2 million scans in 18 months. The technology barely moved; the value exchange did. They stopped selling a scan and started selling the follow-up.

In all three stories, building was the easy part. The product strategy was the work before it: knowing precisely what customers value and deciding what should change.

A rhythm you can start on Monday

You don't need an offsite. You need a rhythm:

Sense - Monthly, ask five customers what they'd do if you vanished. Track what they pay for, not just what they use.

Decide - Quarterly, finish one sentence: "Our customers pay us for ___, and in twelve months they'll pay us for ___." If both blanks match, look harder.

Move - Place one small, time-boxed bet against the second blank, with a clear signal to kill or scale it.

Repeat - Fit isn't a milestone you report once. It's a reading you take - the way Melburnians glance at the sky before leaving the house.

Carry the umbrella

The founders I admire most don't have the best forecast. They've built a company that can change its mind faster than the market changes its weather.

Every competitive advantage has an expiry date. Your ability to reinvent does not.