There’s a woman in my neighbourhood who hasn’t left her house in three weeks. Not because she’s sick. Not because she doesn’t want to. She just doesn’t have a way to get to the café down the road, and nobody’s expecting her there anyway.
She’s not unusual. There are people like her all over the country. Elderly people. People with disabilities. People who moved cities and never quite landed. People whose partner died and whose world shrank to the distance between the couch and the kettle.
We call all of this loneliness. And the way we talk about it (the campaigns, the mental health funding, the “reach out to someone today” posters) frames it as an emotional problem. Something internal. Something to feel your way out of.
I think that’s mostly wrong. I think loneliness, for a lot of people, is a logistics problem.
The chain is broken at every link
What stops an isolated person from going to lunch is not a shortage of wanting to go. It’s that the chain between “wanting to go” and “being there” is broken, and it’s broken at every link.
There’s no reason to go. Nobody invited them. There’s no event on. They don’t know what’s happening in their area today, and the things that do exist (community centres, council programmes) feel institutional. Nobody wakes up excited about attending a “social connectedness workshop”.
There’s no way to get there. They don’t drive. Public transport is inaccessible or confusing or doesn’t go where they need. They’d need someone to pick them up, and there’s nobody to ask. A taxi is about $30 each way, and they’re on a pension.
There’s nobody expecting them. Even if they could get there, they’d arrive alone. No familiar face. No table saved. No one who’d notice if they didn’t show up.
That’s three logistical failures, not one emotional failure. And nearly the whole public response is aimed at the emotional layer. We fund therapy. We run awareness campaigns. We tell people to check on their neighbours. All fine. All necessary. And none of it puts a person in a car at 11:30 on a Tuesday.
Three things
The fix has the same shape as the failure. A reason to go. A way to get there. And people who notice when you’re missing.
A reason to go means a real calendar of real things. Not a programme. Not a service. A long lunch at a local café. A morning walk with a group. A cooking class. A pub quiz. Things that feel like life, not like being managed.
A way to get there means door-to-door logistics. A ride. A companion if they need one. Accessibility built in from the start, not as an afterthought. Someone who shows up at 11:15, rings the doorbell, and says “ready to go?”
People who notice when you’re missing means human nodes, not an app sending push notifications: a café owner who knows the regulars, a volunteer who checks in, a network that picks up the phone when someone doesn’t show.
Together they do something no amount of awareness-raising can do: they physically move a person from isolation into the world.
One text message
I keep coming back to one interaction, the simplest possible version of this.
You get a text: “Want to go to lunch tomorrow? Café on Ponsonby Road. We’ll pick you up at 11:30.”
That’s it. One message. Behind it, everything is orchestrated: the event is booked, the ride is arranged, the companion is confirmed, the café knows to expect you, and someone will notice if you don’t come. From the user’s side it’s just an invitation. A normal, human invitation to go have lunch.
That compression, from a sprawling coordination problem to a single friendly text, is what software makes possible now. Matching outings to a person’s interests and mobility, orchestrating rides and companions and venues: the coordination cost that used to make this impossible for anything larger than a village has collapsed.
The important thing is what’s not technology. The lunch is real. The café is real. The companion is a real person, not a chatbot. The other people at the table are really there. The software handles the logistics so that the human part (the showing up, the sitting together, the being expected) can happen at all.
What the numbers say
The honest picture of loneliness in New Zealand is less dramatic than the posters suggest, and more specific. Older New Zealanders are not, on average, the loneliest group: in the official social survey it is the young who report the most loneliness, and the 65–74 band the least. (Ministry of Social Development 2016) Nor is the country an outlier. Across the OECD, about 6% of people report feeling lonely most or all of the time, and New Zealand sits near that average. (OECD 2024)
What the averages hide is the concentration. For the people who are genuinely isolated (living alone, recently bereaved, no longer able to get out unaided), the condition turns chronic, and chronic isolation is dangerous in a measurable way: across 148 studies covering more than 300,000 people, those with stronger social relationships had 50% higher odds of survival over the study periods. (Julianne Holt-Lunstad 2010) The unusual thing about this risk factor is that the treatment is not a drug. It is other people.
The economics point the same way. A bed in residential aged care costs roughly $60,000–$90,000 a year (MoneyBalance 2026), some of it paid by residents themselves and the shortfall topped up by the government. (Work and Income 2026) Keeping someone connected (getting them out of the house three times a week) might cost $5,000–$10,000 a year in rides, coordination, and venues. That second figure is my estimate, not a study’s, so hold it loosely. But the gap between the two numbers is wide enough that the conclusion survives any reasonable error: prevention is cheaper than crisis, and nearly all the money sits on the crisis side. Disability support budgets, aged care budgets, ACC, PHO funding: the money exists. It is allocated to reacting.
Delivery and participation
The system we do have treats isolation as a health problem to be managed. Fund a service. Write a care plan. Assign a support worker. The person gets a visit once a week from someone wearing a lanyard, and we call that “community participation”.
That’s not participation. That’s service delivery. Service delivery brings the world to you: a support worker visits, a meal is delivered, a phone call is made. It keeps you alive and it reinforces the walls, because you’re still inside, still the recipient. The help comes to your door and your door stays closed. Participation gets you out. You go somewhere. You sit with people who are there because they want to be, not because they’re paid to be. The door opens outward.
And participation can’t be delivered by relabelling the programmes. The moment something feels like a programme for lonely people, lonely people don’t want to go. Nobody wants to self-identify as the kind of person who needs a loneliness intervention. But everyone wants to go to a good lunch. The café doesn’t have a ramp because it’s compliant; it has one because that’s how you build a café.
I’m not an expert on loneliness. I’m not a social worker or a policy researcher. I’m a founder in Auckland who makes drinks and builds software. But I notice things, and what I notice is that the people in my life who are lonely aren’t missing feelings. They’re missing rides. Get someone to the table and the feelings follow the logistics.
That’s the argument, anyway. I believed all of it when I first wrote it down, and I still believe most of it. Here’s the part I had wrong.
The word doing too much work
Look again at the text message. “We’ll pick you up at 11:30” is logistics, and logistics can deliver it. The table being booked is logistics. But “the café knows to expect you” is a different kind of clause. To be expected is not a logistical fact. It’s a relational one. It means there’s a person at the other end for whom your absence would register, who knows your name and your usual and whether you take milk. Logistics can put you in the chair. It can’t put that person across the table. A driver can guarantee you a seat. Only time and repetition can make the seat yours, so that the group notices the week you’re not in it. And that second guarantee is the one people mean when they say they’re lonely.
I let solving the logistics stand in for solving the loneliness, because the logistics is the part you can solve, and there’s a strong pull toward calling the solvable part the whole problem. It isn’t. Loneliness has two layers: a logistical layer, which is about getting there, and a relational layer, which is about being known once you’ve arrived. The first is an engineering problem. The second is not, and no amount of excellent dispatch reaches it.
You can watch the gap open up at a well-run event. The van arrives on time, the room is warm, the lunch is good, and a person is delivered, on schedule, into a crowd of strangers, where they eat quietly and go home having been out without having been with anyone. Presence was manufactured. Belonging was not. If the studies of befriending schemes and organised outings show what I suspect they show (attendance up, measured loneliness barely moved), this is the mechanism: the thing that was missing was never the attendance.
What can honestly be promised
So the promise has to be smaller than the one I started with, and I’ve come to think the smaller promise is the better product. Belonging can’t be booked. It grows out of repetition: the same faces at the same table, often enough that recognition turns into being known. What logistics can do is create the conditions (the same group, the same driver, a standing Tuesday) and then get out of the way while time does the part it can’t.
That’s the bet behind Homerun, the service I’m building. Treat the reason, the ride, and the return as one problem. Handle them well enough to be boring. Then hold the same people together, at the same tables, long enough for the harder thing to have a chance.
The most a service like this can honestly promise is not that it will make you less lonely. It’s that it will get you reliably into the room, often enough and with enough of the same people, that connection has a fair chance to happen on its own. The ride is the easy part. It’s worth building precisely because it clears the way for the part nobody can do for you, which was always the point.