I run a software agency in Malta. I've been writing code for over ten years. And this year I started a courier company.
When I tell people this, most of them just look confused. You're a software engineer. Why would you choose to become a courier?
I think the question assumes a courier company is a van and somebody to drive it. Most of it isn't. The parcel getting from A to B is the easy part. The harder parts are working out which parcels should go on the same run, knowing whether you can actually deliver tomorrow before you tell a merchant that you can, giving a driver something useful to do when nobody answers the door, and being able to tell a merchant where their money is on a Tuesday afternoon. That's all software, and software is what I've been doing for ten years.
A lot of it is also work that used to need a person doing it. Most courier companies have a dispatcher, someone who knows the island and knows the drivers and spends the first part of the morning working out who takes what. I'm trying to build this one so that there isn't a human dispatcher, and the system does that job instead.
Building software stopped being the hard part
For most of my career, the moat in software was the building.
If you could actually ship a working product, you were already ahead of most people who had the same idea. The code was the barrier. That's what you got paid for, and that's what protected you once you were out there.
That barrier is gone.
I'm not being dramatic about AI here. I use Claude Code every day at Solveita and it has genuinely made us faster. But I've watched what that speed does to the maths. Something that used to take me a week takes a day now. Which is great, until you notice it takes everyone else a day too.
If you're building a SaaS product today, whatever you spent six months on can be rebuilt by a competent person in a weekend. Not a rough version of it. A real one. The feature gap that used to buy you a year of breathing room now buys you a fortnight.
Which moves the whole problem somewhere else.
If the building isn't the scarce part any more, then the scarce part is getting anyone to care that you built it. That was always the harder half of the business. The difference is that it used to be one half of two, and now it's most of what's left.
And it's the half AI is no help with at all. It'll write your onboarding flow in an afternoon. It won't get the thing in front of a single person. It won't make a shop owner pick up the phone to you, or talk a customer who's been let down before into trying somebody new. It can't compress the year it takes to build a name in a market this small, because that year isn't work. It's just time passing while you keep turning up and not messing it up.
So you get two people with more or less the same product, and the one who does better usually isn't the one who built it better. It's the one who already had somewhere to put it.
Where I think the moat went
I think it went physical.
Start with the van. Not because a van is hard to come by, but because buying one takes money and a decision you can't cheaply undo, and that on its own thins the field. In software, somebody can look at what you've built on a Friday and have their own version running by Monday, and it costs them a weekend. Nobody does that with a courier company. Before anyone can come at you here, they have to put real money down, and then keep paying for it every month whether the work shows up or not.
And the van is the easy part. It's everything around it that takes time.
The trust, mainly.
That's the thing about software. You build it, and then it works without you. It's serving people at three in the morning while you're asleep. That's the whole appeal, and I've spent ten years enjoying it.
This doesn't work like that. Somebody has to be in the van. Not a process, not a job that runs on a schedule, a person, actually out there in the rain. Somebody has to knock on a door that doesn't open, drive back that evening, and knock again. Every parcel is a small promise that a person has to physically keep, and there's no deploying your way out of it.
That is what a shop owner is really weighing up when they decide whether to hand you their parcels. You're not just carrying a box. You're carrying their customer's last impression of them, and on cash on delivery you're carrying their money as well. Nobody hands that over because your pricing page is good. They hand it over after you've turned up, week after week, in whatever weather, and not messed it up once. There's no quick version of that, and nothing I can build that skips it.
And the businesses I know best in Malta are run on WhatsApp, a notebook and somebody's memory. That's not me having a go at anyone. It's just what happens when the people who understand the operation have never had someone build software for the way they actually work. That's the part I thought I could help with.
So for me it was never software instead of physical. It was a physical business with real software underneath it. Own the slow, expensive, hard-to-copy part, and build your own tools to run it.
Which is a bit contradictory, and I'm aware of it. The same thing that took away my advantage in software is what makes this business work. AI removed the edge I used to get from being able to build something quickly. It gives me a different one here, because an operation like this is full of small decisions that used to need a person making them, and a lot of them don't any more.
It started with my wife's shop
My wife runs Orden Beauty, a salon in Swatar, Birkirkara. Alongside the salon she sells products, and for a long time the couriers were making her life hell.
Lost parcels. Three to five days to cross an island you can drive end to end in an hour. Printing labels at midnight. No pickup unless she had three orders ready and waiting. Nothing moving on a Saturday. And drivers marking parcels as "missed" while her customer sat at home all afternoon waiting for a doorbell that never rang.
Sales are exciting. Deliveries are stressful. If you've ever run an online shop, you know exactly what I mean. She'd spend hours building a perfect experience for her clients, and then one bad delivery would undo all of it. Because when a delivery goes wrong, the customer doesn't blame the courier. They blame the shop. They leave the review on her page, not the courier's. And they don't come back.
So I took her deliveries off her hands. That was supposed to be a stopgap. What actually happened is that I found I liked the problem, and that I could see the gaps from the inside in a way I never could have from a spreadsheet.
It wasn't just her
Before I committed money to this, I wanted to know whether my wife's experience was bad luck or the actual shape of the market. So I spent months talking to small Maltese sellers. Not surveys. Actual conversations, with people running a shop out of a garage, a spare room, or one small unit in an industrial estate.
It was the shape of the market.
Courier services here are built for volume. That's a completely rational way to run a logistics company and I don't blame anyone for it. But it means the systems are shaped around merchants shipping tens of parcels a day:
- Upload a manifest, when what you actually have is five parcels on a table and a laptop you only open at night.
- Hit a morning cutoff, so an order that comes in at midday sits in the corner of your living room until tomorrow, while the customer who paid for it starts wondering.
- Commit to a monthly minimum, and pay for parcels you never sent in a month that was quieter than you hoped.
- Wait thirty days for your cash-on-delivery money, while the stock you need to buy today is sitting in somebody else's account.
If you're doing fifteen orders a week, every single one of those is a problem.
And it's not just admin. It's the message at half nine on a Sunday asking where the parcel is, when you have no more idea than the customer does. It's refunding an order you already paid to pack and ship. It's apologising for a mistake that wasn't yours.
So the small ones end up doing one of three things. They deliver themselves after closing, which means their evening is gone. They lose the order. Or they decide selling online isn't worth the aggravation and stop.
It's not really about price. Nobody is getting rich off the difference. It's that the smallest merchants in the market are the worst served, and they're the ones with the most to gain from someone taking them seriously.
Why Malta is a strange and good place for this
Two numbers from Eurostat made the decision for me.
The first one usually surprises people. Malta is not a fast-growing ecommerce market. Across the EU, the share of internet users buying online rose from 62% to 78% between 2015 and 2025, a jump of sixteen percentage points. Malta rose eight. We're among the slowest movers in Europe on that measure.
But that's the wrong number to stare at, because Malta was already near the top. What matters is how often people buy, and there Malta is genuinely first in the EU. In 2025, Malta was the only country in the union where the single most common online shopping frequency was more than ten purchases in three months. A third of Maltese online shoppers, 33%, sit in that bracket. Nobody else in Europe looks like that.
Think about what that actually means in physical terms. It means parcels. A small island, short distances, and a population that orders constantly.
Now the other side. Across the EU in 2024, only 23.6% of enterprises made any online sales at all, and the gap runs by size: 21.4% of small enterprises against 31.7% of medium ones. The smaller the business, the less likely it is to be selling online. That pattern is exactly what I kept walking into in Malta.
Put the two together and you get the line I keep coming back to:
Malta's customers are ahead of Malta's shops.
The demand is already here. It's the businesses that haven't caught up. And when you ask small merchants why, it's almost never the website. It's everything that happens after the order.
What I'm building
Kwip is a courier service for Malta built for that merchant specifically.
The courier service itself is deliberately plain. No minimum volume. No contract. You book a pickup when your order is actually packed and ready, not before some cutoff you had to plan your whole morning around. Next day from pickup at the latest, and often sooner. We run Malta Monday to Saturday, and Gozo on Saturdays.
None of that is clever. Every one of those is a direct answer to something a merchant said to me, or something I watched go wrong in my wife's shop. That's just what comes out when you design for the small shop first, instead of taking a system built for someone much bigger and trying to squeeze them into it.
The software is the reason it works, and there is far more of it than people expect. Dispatch and routing, so parcels heading the same way end up on the same run instead of crossing the island twice. That's the piece I care most about, because in a normal courier company it's a person, and a person is exactly what keeps a small operation from growing. Orders arrive all day in no particular order. Somebody has to keep deciding what fits on which run, what can still make tomorrow, which pickup is worth the detour and which one isn't. I'm building that decision into the system instead, so the runs assemble themselves as the orders land and a driver opens their phone to a route that already makes sense. Capacity, so we know what we can promise before we promise it, worked out from what's actually on the road rather than from somebody's feel for it. Pickup availability and delivery windows. What the driver sees on their phone when they're at the door, and what happens when nobody answers it. Proof of delivery. Cash on delivery reconciled back to the right merchant without anyone doing sums at the end of the week. A merchant who can see where every one of their parcels is without phoning anybody to ask.
None of that is visible from the street, which is probably why a courier company looks like just a van from the outside. We're building all of it ourselves. It's the part I actually know how to do.
The person at the door
Everything above is about the merchant, because the merchant is the one who pays. But the delivery doesn't happen to them. It happens to their customer, and that's whose afternoon gets taken.
Think about what a delivery normally tells you. Nothing, or a message first thing saying it's coming today, somewhere between now and the end of the day. So you stay in. You don't go out for lunch. And if you do slip out for twenty minutes, that's the twenty minutes the van comes, and now you're on the phone to a depot about a card that was never actually left.
That was the exact thing that used to happen to my wife's customers, and she was the one who got blamed for it.
So the person receiving the parcel gets told when it's actually coming, with an estimate that means something rather than a working day. They can follow it while it's moving instead of refreshing a status that changes once. And if tomorrow afternoon has stopped working, or they'd rather it went to their mother's, they say so and it moves.
That last one sounds small and it's the hardest thing here. Any courier will reschedule if you phone up and somebody re-plans the morning by hand. Doing it on demand, on the day, while the runs are already out, means the plan has to be able to rebuild itself every time anything changes. Somebody re-planning it by hand can't keep doing that all day without the whole thing coming apart. Software can, and that's the only reason we can offer it as a normal thing instead of as a favour.
It's really the same problem on both sides. The merchant wants to stop apologising for deliveries, the customer wants to stop waiting around for one, and both come down to the system making decent decisions all day without someone having to be asked.
Where this goes
Courier is the way in, not the business.
I've spent years building ecommerce systems for other people's businesses. Checkouts, stock, order flows, and the joins between a shop and everything sitting behind it. Every one of those projects showed me where a small retailer actually loses time and money, and it is almost never the part they expect. For the smallest sellers here, delivery is the most expensive thing nobody has sorted out, so that's where I started.
It's also a business designed to grow without me in the middle of it. I've stayed close to the operation from the start, because that's how you find out where it actually breaks and how you earn a first customer, not because that's the plan. Vans and drivers are things you add. The system that decides what goes where, and keeps everyone told, gets built once and then carries whatever volume you put through it. This is also why the dispatcher matters so much. In a normal courier company, more vans eventually means more people coordinating them, and the coordination is what gets expensive and error-prone long before the driving does. If the deciding is done by software, a third van mostly just costs you a third van. That's the part I think actually scales.
What I'm actually interested in is the layer underneath small Maltese commerce. A shop that can take an order over the phone and have it at the customer's door the next morning is running an ecommerce business, website or no website. A lot of shops here are one decent system away from that, and as far as I can tell nobody has bothered to build it for them.
So the plan is to start with the boring, physical, hard-to-copy bit, get it right, and build outward from there. Not a platform hunting for merchants. Merchants first, then the platform they actually asked for.
I still run the agency
Solveita is running as it always has. I'm still building software for clients and I still enjoy it.
If anything, Kwip has made me better at the agency work, because I'm on the other side of it now. I'm the business owner with an operation to run, staring at my own system at seven in the morning with a van half loaded, wondering why something that made perfect sense on a whiteboard is so annoying in real life. You don't get that perspective from the outside.
The thesis isn't "software is dead". It's that software on its own, as the entire product, with nothing physical underneath it, is a much worse bet than it was five years ago. Software attached to something real is a much better one.
Ask me in two years
I might be wrong about all of this.
Maybe SaaS is fine and I've just talked myself into something because I got restless. Maybe logistics is hard in ways I haven't hit yet, and there's a good reason the people already doing it run their businesses exactly the way they do. That's normally how it goes when someone from outside an industry decides they've spotted something obvious.
What I do know is I'd rather be wrong about something I can see and touch. If this doesn't work, it won't be because somebody in another country shipped a better version of it while I was asleep. It'll be because I got the operations wrong, and that's a problem I can actually fix.
So that's the bet. A van, a system I built myself, and a few hundred small Maltese shops who deserve a better option than the one they've got.
Questions I get asked about this
What is Kwip? Kwip is a courier service in Malta, founded in 2026 by Duane Chetcuti, built specifically for small and independent merchants. It offers next-day delivery from pickup at the latest, with no minimum volume and no contract.
Why would a software developer start a courier company? Two reasons. The personal one is that my wife runs an online shop and the existing couriers were costing her customers. The strategic one is that AI has made software cheap to build and distribution expensive to win. A physical operation, meaning vans, routes and relationships with shop owners, takes months or years to replicate, which makes it a far better moat in 2026 than a feature set. The software is still the advantage; it's just no longer the product on its own.
Who is Kwip for? Small Maltese merchants shipping anywhere from a handful to a few hundred parcels a month: home businesses, single-unit retailers, Instagram and Facebook sellers, and shops that take orders over the phone or on WhatsApp without ever calling themselves an ecommerce business.
Where does Kwip deliver, and how fast? Malta, Monday to Saturday, with next-day delivery from pickup at the latest and often sooner. Gozo is served on Saturdays.
Does Kwip use AI? Yes, in its operations. Kwip is built so that dispatch, the job of deciding which parcels go on which run and what can still be delivered tomorrow, is handled by the system rather than by a human dispatcher. In most courier companies that is a person's full-time job, and it's the point at which a small operation stops being able to grow.
Can customers track a Kwip delivery? Yes. The person receiving the parcel is notified with an estimated arrival time and can follow the delivery in real time while it's on the road, rather than being given a status that updates once a day.
Can a customer change the delivery time, date or address? Yes, on demand and on the day. Because dispatch is handled by software rather than by a human dispatcher, the plan can rebuild itself when something changes, so rescheduling or redirecting a parcel doesn't depend on someone re-planning routes by hand.
Does Kwip handle returns? Yes. Returns and collections run through Kwip Send, which is an on-demand courier service rather than the next-day network.
How is it different from the couriers already operating in Malta? Existing couriers are optimised for high-volume shippers: manifests, morning cutoffs and monthly minimums. Kwip drops all three. You book a pickup when the parcel is packed, with no volume commitment.
Do you still run Solveita? Yes. Solveita, the Malta software agency I founded, is operating as normal and continues to take on client work. Kwip is a separate company with its own operation.
If you're one of those merchants and any of this sounded like your week, that's exactly who Kwip is for.