In the beginning

I've been meaning to write this down for a while, and a first newsletter seems as good a place as any. Not the product news, that can wait, but the why. Why someone who has spent his life in software ended up, on nights and weekends, building a way for neighbours to buy vegetables.

Sustainable living

My wife and I have cared about sustainability for as long as we've been together. One of the load-bearing beliefs underneath that, if I'm honest, is an uncomfortable one: the food system is broken. Fixing it means moving away from monoculture and big food, and back towards smaller producers: organic, permaculture, the sort of farming that works with the land rather than against it.

Before any of this I worked in the big city, in London, in the years before covid. When we'd had enough of it, we moved to rural Canada to look for a patch of land where we could actually practise what we kept preaching. For a while I carried on working remotely, often from our car (the back seats make a surprisingly serviceable office), while my wife WOOF'ed, working on organic farms in exchange for food, board and an education. We learned, first hand, the joys and the trials of being a small organic vegetable farmer. Long days bent over a bed of greens. Sleeping in a tent. And produce that, somehow, made all of it worth it. The farm where she spent that time, Catena Farm, is still going strong.

Nothing local

After the organic farm we moved to a small town, we still lived amongst farms, but most of them were what's politely called "conventional", the petrol and chemical model that, for all its air of permanence, has only really existed since the 1970s. The organic produce we actually wanted had to be shipped in.

The way people got around that was something the locals called a buyers club. A handful of neighbours, or one organised soul thinking ahead, would club together on a bulk order. You'd hit the supplier's minimum, cut out a middleman, and everyone came away with cheaper, better food. Out in the country it wasn't only about the money. Often it was simply the only way to get the things we were after: the organic grains, the rest of it.

Near the end of that chapter, while we were still hunting for land of our own, covid arrived and rearranged our plans for us. Rural house prices shot up as people fled the cities, and we were priced out. We ended up in a small city instead, had two kids, and eventually admitted the thing we'd been avoiding: the reasons we'd come to rural Canada no longer held. So we packed up once more and moved to my wife's country of origin, South Africa.

South Africa

We both still cared, deeply, about the food system. Through a neighbour my wife heard about a website for running a buyers club, though here the same idea had been rebranded as a food club, and had been quietly ticking along for years. Someone, a while back, had built a site to make organising one a little less painful.

She ran a food club on that platform for nearly two years. I'll be honest: the software was a struggle. Hard to use, and wearing the look and feel of a much older internet. In the two years she used it, it didn't really change.

A system that got in the way

Almost everyone we spoke to said the same thing: the tools put them off. That matters more than it sounds. If people are going to change their habits and step away from the big chains, the alternative can't be a chore. It has to be a pleasure.

Then there was the money. For most of that first year, load shedding ran four to eight hours a day, which meant keeping several fridges and freezers alive, appliances we'd bought out of our own pocket (a chest freezer doesn't care that the grid is down). When you did the sums, the modest markup my wife put on top of wholesale was almost entirely eaten up, and a third of it went to the platform. We were about breaking even. Subsidising the thing, really, and she was fraying her nerves in the process, because none of it was easy.

What "better" would look like

I was busy at work, but for the better part of a year I kept turning it over. Instead of four tools, one. People expect to pay online. Stock should be the software's job, not something you hold together with scraps of paper, a torrent of messages, and a quiet, constant hum of stress.

Most of all, two things had to be true at the same time. The platform had to be as good as the supermarkets for the shopper. Better, ideally. And the person running the market had to be paid fairly for real work. Otherwise it would only ever be possible for people like my wife: lucky enough not to need the income, doing it purely because they cared. For this to go mainstream it had to be self-sustaining, exactly like the farming it set out to support.

Around then the tech market was contracting. The company I'd been with for eight years was shrinking its software teams, and offered me a way out, along with a little runway to go and build something of my own.

A way forward

With time on my hands, I went to the people behind the existing system and offered my time and what I know: twenty-odd years in tech, most of it across the full stack. They were clearly passionate; software just wasn't their world, which explained a great deal about the product. It was a creaking thing, built on an ageing enterprise framework and handed from one developer to the next until no one was left who really understood it.

I suggested, as gently as I could manage, that the honest move was to start over. Understanding and rebuilding the old code would likely cost more than a clean slate, and a modern foundation would let a small team move fast and ship something that actually worked. The offer was declined. I never learned exactly why. Sunk cost, I suspect. Years and a lot of money had gone in, and that's a hard thing to walk away from.

Then ordinary life took the wheel for a few months. Christmas, school holidays, small children, and my free time evaporated. In the new year I came back with a different idea: instead of telling, show. I'd build a working proof of concept, proof that a fresh platform was not only possible but the obvious move. I was happy to do it entirely on spec, if it nudged the food system in the right direction.

A few months of nine-to-five and it came together, built on nothing but public information. On top of the basic shop (products on shelves, a checkout), I added collection slots, online payments, and pricing that varied by storage type: dry goods, fridge, freezer. For that last part there was no reliable public data, so I went to the suppliers directly. They're public companies, after all, on the lookout for new business, and most have a contact address of one sort or another.

I sent off a batch of emails. This is where things got strange.

When plans go awry

What I'd hoped might become a collaboration turned into something more adversarial. There was a misunderstanding about who I was and what I was up to, and the door I'd hoped to walk through quietly closed. I won't relitigate it here. It's enough to say the partnership I'd imagined wasn't to be, and for a while it left me where I'd started: alone, with a platform maybe ninety per cent finished and a community not quite sure what to make of me.

It was, I'll admit, disheartening.

Going ahead alone

But I still cared, and I still wanted to do something about it. So I pushed on with helyi by myself. I worked through around fifty suppliers' product sheets by hand to build the catalogue, arranging it the way I'd been picturing for months: like a proper shop, products on shelves, shelves in aisles, the whole thing sorted by how it needs to be stored. Plenty of the data was already a little stale (most of the lists were months old by then), but close enough to start.

The first market

I built like a man possessed. Sixty-hour weeks, because a date had been set: the first Local would open, with collection two weeks later. It worked. Some labels were off, because by the end I was writing scripts to parse the bewildering zoo of formats supplier lists arrive in: pdf, Excel, csv, Word, take your pick. But everything I did, I knew, would be easier next time. The data only improves, and every new Local stands on the grunt work of the first.

I sat at the back of that first market and kept coding while it ran, watching how people actually used the thing. That's where features like the shared packing and collection list came from. Organiser and member both see it, each ticking items off, watching live what the other has, with no room left for the usual confusion about what did or didn't make it into the bag. There were bugs, of course. Software always has bugs. The difference this time was that I'd built it and understood every line, so most were gone within minutes of being found. Everyone got what they ordered. Every invoice went out automatically, and correctly. Every payment was tracked.

The second market

Yesterday we ran the second market, and it was a breeze. Local stock management meant we ordered only what was actually needed. People arrived on time, helped along by a reminder an hour before their slot with a link to their basket. The first event was about as stressful as one on the old platform: better, but new. This one was simply easier, and you could feel the difference.

What's next

Now that the early kinks are out, the data's in good shape and the platform's sound, I want to widen the reach of helyi, and of localmarkets in general. This doesn't have to be something only the privileged can pull off. Anyone should be able to start one and make it work, with every tool they need in one place.

That's the short version of how we got here. More soon on where we're going. If you'd like to follow along, you're already in the right place.

If you live in South Africa, and it sounds like something you'd be interested in, check out helyi