Tom built the landing page in a weekend. It wasn’t beautiful – Jas hadn’t joined yet, and Tom’s design instincts ran to “functional.” But it had a clear headline (Fresh local produce, delivered weekly), a description of the two box sizes, a price, and a signup form that collected a name, email, address, and payment details.
The payment integration worked. Tom had used Claude to generate a Stripe setup in an afternoon, and it was solid – one of the few things from those early weeks that didn’t need rebuilding. The confirmation email went out. The landing page loaded fast. The form submitted cleanly.
What the form didn’t collect was a unit number. Or a delivery note. Or any indication of whether the customer lived in a house, a flat, or a unit complex. Tom’s data model had: name, email, street address, suburb, postcode. It seemed like enough at the time.
The flyer
Maya designed the flyer herself. Hand-drawn, because she couldn’t afford a designer and because she wanted it to feel personal. A sketch of a green crate overflowing with vegetables. The Greenbox name in her own handwriting. A QR code that Tom generated, linking to the landing page. And a line at the bottom: Local farms. Weekly boxes. No thinking required.
She printed fifty copies at Officeworks and drove down to the Margaret River farmers’ market on Saturday morning.
The market was where Maya had grown up. Her parents had sold produce from a trestle table at the far end for fifteen years. She knew the rhythms: the early-morning setup, the rush between nine and eleven, the slow afternoon when the stallholders started packing up and the remaining customers got the best deals. She knew the regulars – the retired couples who came every week, the young families with kids running between the stalls, the restaurant owners doing their weekend sourcing.
Maya walked the market with her flyers, talking to everyone who’d listen. Some of them remembered her parents. Most of them liked the idea. A few were sceptical.
“Another subscription thing? I tried one of those meal kit services. Lasted three weeks.”
“This is different. It’s not a meal kit. It’s actual produce from actual farms within fifty k’s.”
“Which farms?”
“Dave Morrison, for one. And Rachel’s place.”
The mention of Dave’s name carried weight at the Margaret River market. People knew Dave. People trusted Dave. If Dave was involved, the vegetables would be good.
By the end of Saturday, twenty-two people had scanned the QR code and signed up. Twenty-two. Maya sat in her car in the market car park and stared at her phone. Twenty-two real people had given her their credit card details and trusted her to send them a box of vegetables.
She called Tom. “Twenty-two.”
“Twenty-two what?”
“Subscribers. We have twenty-two subscribers.”
A pause. Then Tom’s voice, with the particular excitement of a builder who’s just learned that the thing he built has users: “That’s… that’s actual people.”
“Actual people who expect a box of vegetables on Thursday.”
“Right. Thursday. That’s… five days from now.”
“Four, actually.”
Packing day
Wednesday morning. Maya’s kitchen table.
Dave arrived at 5am in his ute, the back loaded with green crates. Tomatoes, zucchini, spinach, carrots, beetroot, a few bunches of herbs. Everything picked the day before. He carried the crates in through the front door, set them on the kitchen floor, and surveyed the operation.
“This is your packing facility?”
“For now.”
“It’s a kitchen table.”
“It’s a large kitchen table.”
Dave shook his head with the expression of a man who had seen many ambitious plans meet their first contact with reality. He left without saying much else, though he paused at the door and said, “The spinach won’t last if you leave it out. Get it in the boxes fast.”
Rachel arrived an hour later in her own ute, with a smaller contribution: bunches of kale, some sweet potatoes, and a crate of capsicums. She helped carry them in and looked at the kitchen table.
“You right?”
“I think so.”
Rachel studied the piles of produce. “You’ll want to pack the heavy stuff at the bottom. Sweet potatoes first, then the root veg, then the leafy stuff on top. If you put the spinach at the bottom it’ll be soup by the time it arrives.”
Maya wrote this down. She hadn’t thought about packing order. The spreadsheet had columns for subscription size, produce allocation, and delivery address. It did not have a column for “which vegetables go on the bottom.”
Sam arrived at seven with boxes. Actual cardboard boxes – she’d sourced them from a packaging company in Welshpool, the cheapest option that was still food-grade. They were plain brown, no branding, because branded boxes cost four times as much and Maya’s budget didn’t stretch that far. Sam had written “GREENBOX” on each one in green marker. It looked homemade, because it was.
They packed twenty-two boxes on the kitchen table. Maya and Sam working side by side, consulting the spreadsheet on Maya’s laptop, weighing produce on a kitchen scale. Tom sat in the living room, monitoring the website and the payment system, feeling useless.
“Can I help pack?” he asked.
“Can you tell the difference between baby spinach and rocket?” Sam replied.
“They’re both green.”
“Stay in the living room.”
The packing took four hours. Maya’s back ached. Sam had produce stains on her shirt. The kitchen looked like a greengrocer had exploded. Nadia, who had taken the day off to help, wrapped the last box in brown paper and taped the address label on with the precision of someone who had decided that if her living room was going to be a warehouse, at least the warehouse would be tidy.
By midday, twenty-two boxes were stacked by the front door. They looked good. They smelled good. Maya took a photo and sent it to Dave. He replied with a single thumbs-up emoji – the most effusive communication she’d ever received from him.
The delivery
Sam had arranged delivery through her mate Callum, who drove a courier van in the southern suburbs. Callum was reliable, Sam said. He’d done deliveries for the trucking company and he knew the Perth metro area. He was also cheap – Sam had negotiated a rate per box that was barely above fuel costs, a favour that Callum would regret by the third week.
The plan was Thursday delivery. Callum would pick up the boxes at midday and deliver them between 2pm and 6pm.
Callum picked up the boxes on Wednesday.
“Thursday,” Sam said, when he turned up a day early. “Thursday delivery. I said Thursday.”
“You said this week. I’ve got a full run tomorrow. Today’s better.”
Sam called Maya. Maya called Callum. Callum was already driving, with twenty-two boxes in the back of his van and the confidence of a man who had been doing deliveries for twelve years and didn’t see the problem.
“Most of these people are at work,” Maya said. “They won’t be home until five or six.”
“I’ll leave them on the doorstep.”
“It’s fresh produce. In a cardboard box. In the sun.”
“I’ll find shade.”
Half the boxes were delivered to empty houses on a Wednesday afternoon. Six were left on doorsteps in full sun. Three were delivered to the wrong addresses because Tom’s address data didn’t include unit numbers – two customers lived in unit complexes, and Callum had left the boxes at the front door of the building, not the individual unit. One box was never found. The spinach, which Dave had warned them about, had wilted in the boxes that sat in the sun for three hours.
Maya’s car
At 4pm on Wednesday, Maya was sitting in her car outside a house in Applecross. She’d driven out to intercept the last few deliveries, hoping to correct the addresses and apologise in person. The house belonged to a subscriber named Mrs Patterson, a woman in her sixties who lived alone on Stirling Highway and had signed up at the market because she liked the idea of someone else choosing her vegetables for the week.
Mrs Patterson wasn’t home. The box was on her doorstep, in the shade at least, but it had been there for two hours. Maya picked it up and opened it. The spinach was limp. The herbs had started to wilt. The tomatoes were fine – tomatoes are forgiving – but the overall impression was not “premium local produce.” It was “vegetables that had been sitting in a box for too long.”
Maya put the box back, sat in her car, and called Mrs Patterson.
“Hello?”
“Mrs Patterson, this is Maya from Greenbox. Your box was delivered today instead of tomorrow, and I’m afraid some of the produce might not be at its best. I’m so sorry. I’m outside your house now and I’d like to –”
“Oh, that’s all right, love. I saw it when I came home for lunch. The tomatoes looked gorgeous. Don’t worry about it.”
Mrs Patterson was kind. She was generous. She told Maya to stop worrying and come back next week with a better box. Maya thanked her, hung up, and sat in her car for five minutes with her hands on the steering wheel, staring at nothing.
She wasn’t crying because Mrs Patterson was angry. She was crying because Mrs Patterson was kind, and Maya felt like she didn’t deserve it. Twenty-two people had trusted her with their dinner, and she’d delivered wilted spinach on the wrong day to the wrong addresses. The spreadsheet, the flyer, the 5am packing session – all of it had produced a result that was, by any honest assessment, a disaster.
She wiped her face, started the car, and drove to the next address.
The recovery
That evening, Maya, Tom, and Sam sat at the kitchen table – the same table they’d packed boxes on that morning – and went through every problem.
Tom opened his laptop and pulled up the customer data. “We need unit numbers. I’ll add a field to the signup form tonight.” He paused. “I should have thought of that.”
“We all should have,” Maya said.
Sam had a list on her phone. “The delivery window is non-negotiable. Thursday between 3pm and 7pm. Not Wednesday. Not whenever Callum feels like it. I’ll find a different courier if I have to.”
“Can we afford a different courier?”
“Can we afford to lose customers?”
Maya conceded the point.
They went through every failure. The spinach problem was timing – they’d packed too early. If they packed on Thursday morning and delivered Thursday afternoon, the produce would be hours old instead of a day old. Dave had told them this. They hadn’t listened, or rather, they’d listened and then let the logistics override what they’d heard.
The address problem was data. Tom fixed the form that night, adding fields for unit number and delivery instructions. He also added a confirmation step that showed the customer their full address before they submitted, so they could catch errors.
The delivery timing was Sam’s domain. She called three courier companies on Thursday morning and found one – a woman named Jen who ran a small delivery business in Fremantle – who could guarantee a Thursday afternoon window. Jen was more expensive than Callum, but she answered her phone, confirmed delivery times, and understood that fresh produce and hot doorsteps were a bad combination.
By the following Thursday – week two – the process worked. Pack on Thursday morning. Deliver Thursday afternoon. Address data includes unit numbers. Jen delivers within the confirmed window. No boxes in the sun. No wrong-day deliveries. No wilted spinach.
It wasn’t smooth. Sam spent Thursday afternoon texting Jen for updates. Maya called three customers to confirm their boxes had arrived. Tom refreshed the delivery tracker – a shared Google Sheet that was the entire “operations platform” – every fifteen minutes. But the boxes arrived. The produce was fresh. Nobody called to complain.
Mrs Patterson emailed on Friday morning: “Much better this week! The carrots were beautiful.”
Week three
By week three, the process was routine. Pack at 6am Thursday. Jen picks up at 10am. Deliveries between 2pm and 5pm. Sam confirms each delivery by text. Tom monitors the payments. Maya handles the farm coordination – checking in with Dave and Rachel on Monday about what they’d have available, confirming quantities on Wednesday, adjusting the packing list if something fell short.
The rhythm emerged not from a plan but from the accumulated learning of things that went wrong. Every mistake in week one became a process in week three. Unit numbers on the form. Packing order: heavy at the bottom, leafy on top. Delivery window confirmed 24 hours in advance. A shared spreadsheet tracking every box from packing to delivery.
Sam started a simple feedback system – an email sent to every subscriber on Friday asking how their box was. Most people didn’t reply. The ones who did were either very happy or very specific about what they didn’t like. One subscriber requested no coriander. Another asked if they could get extra tomatoes. A third wanted to know which farm her carrots came from.
Maya answered every email personally. She learned the subscribers’ names, their preferences, their quirks. Mrs Patterson didn’t like beetroot. A young couple in Northbridge were vegetarian and wanted more variety in leafy greens. A family in Cottesloe had three kids and needed quantity over variety. A retired teacher in Mosman Park wanted whatever Dave recommended, because she’d been buying from Dave at the market for years and trusted his judgement.
These weren’t user personas on a whiteboard. They were real people with real kitchens and real opinions about coriander.
The email
On a Friday afternoon in the third week, an email arrived from a subscriber named Claire. It was three sentences long.
Hi Maya, just wanted to say thanks. I haven’t thought about what’s for dinner since I started getting the box. That’s worth more than the vegetables.
Maya read it three times. She read it standing at the kitchen counter while Nadia made tea. She read it again before bed. She didn’t have the language for what Claire was describing – not yet. The phrase “job to be done” was months away. But she felt the shape of it. The box wasn’t about vegetables. It was about something else, something larger and harder to name, and Claire had just told her what it was.
The box was about one fewer decision in a day full of decisions. Open the door, pick up the box, cook what’s inside. No planning, no shopping, no standing in a supermarket aisle at 6pm wondering what to have for dinner. Trust the box. Trust the farm. Trust Maya.
She saved the email in a folder she named “Why We Do This.” It was the only email in the folder. Over the next year, it would have company.
Growth
Twenty-two subscribers became twenty-six in week two. Word of mouth – the people who got the box told the people who didn’t. By week four, thirty-one. By week six, thirty-eight.
Maya hadn’t run a single ad. The growth was entirely organic: market flyers, word of mouth, and a short piece in the local Fremantle newspaper that Sam had arranged by calling the editor and saying, “We’re three people packing vegetables on a kitchen table and delivering them to your neighbours. Want to write about it?”
The editor did want to write about it. The article ran on a Wednesday and produced nine signups by Friday. Small numbers, but each one was a person who’d read about Greenbox and decided to trust a stranger with their weekly dinner.
Thirty-eight subscribers was encouraging. It was also nowhere near enough. And the manual operation – Maya, Sam, and a kitchen table – couldn’t scale. Every Thursday was a full day of packing and coordinating. Maya was spending Monday on farm calls, Tuesday on the packing list, Wednesday on logistics, Thursday on packing and delivery, and Friday on customer emails. That left no time for anything else. Tom was building the platform as fast as he could, but the platform was for 200 subscribers and they needed to stop packing by hand long before then.
The seed round would change that. Maya had been talking to investors since before the first box shipped. The terms were clear: reach 200 active subscribers within three months of funding, and the next round follows. Miss the target, and Greenbox is done.
Two hundred. From thirty-eight. In twelve weeks.
Once the money came in, they could hire a proper team, move out of the living room, and build the systems to replace the kitchen-table operation. The pilot subscribers – the thirty-eight people who’d trusted Maya with their Thursday dinners – would keep getting boxes through the manual process for now. But the platform Tom was building had to be ready before the numbers got any higher. You can hand-pack thirty-eight boxes on a kitchen table. You cannot hand-pack two hundred.
Maya looked at the subscriber graph – a line on a spreadsheet that she checked every morning at 5am, before her run, before coffee, before anything else. The line was going up. Slowly. Steadily. But 200 was a long way from 38, and the gap between them was filled with packing days and delivery runs and emails about coriander and a team of three people who were already working as hard as they could.
She needed more people. She needed an office – Nadia’s patience with the living room situation was genuine but not infinite. She needed a developer who wasn’t Tom, because Tom was one person and the codebase was growing faster than one person could manage. She needed money.
The seed round had to close. And once it did, the clock started. Three months. Two hundred subscribers. Build the platform, grow the customer base, prove the model. Tom was already building as fast as he could, using LLMs to generate code at a pace that felt miraculous. They’d shipped a subscription system, a landing page, a basic farm portal, all in weeks.
The question Maya couldn’t answer – the question that would define the next three months – was whether all that speed was pointed in the right direction.