Try this. Open the phone you carry. Look at the apps you keep on the first screen — the few that have earned a place there because you actually use them. Now think back ten years. How many of the apps on that first screen, ten years ago, are still on it now?
Maybe one. Maybe none.
Some of them are gone by your own choice. You replaced them with something better. But many of them are gone for a different reason. The company sold to someone who didn't want it. The team moved on to the next thing. The pricing changed. The product went away. The data came out, if you were quick about it, in a format nothing else could read.
You probably did not notice each one at the time. They went away in twos and threes, on weekday afternoons, with an email and a thirty-day window. But the cumulative loss is a particular kind of damage, and the household has felt it harder than anywhere else.
Sunrise, the calendar everyone in the house loved, sold to Microsoft. Wunderlist, where the shopping list lived, sold to Microsoft. Both shut down a few years later. Mailbox, sold to Dropbox, then closed. Acompli, the mail app you tried and loved, sold to Microsoft. Reeder you still have. RSS you still have. But Sunrise is gone. Wunderlist is gone. The calendar invites you accepted in 2014 are buried somewhere in Outlook now, if they are anywhere at all.
This piece is about that.
Specifically, it is about why the household has been worse-served by software disappearance than any other domain in the personal-software field, why that has happened, and what household software might be if it were built by people who understood that the household was meant to outlast it.
The Pattern
There is a recurring shape to this. Once you see it, you cannot un-see it.
A small useful app is built by a small team. The team is talented, the design is careful, the price is reasonable. The app fills a real gap — the shopping list, the family calendar, the recipe collection, the notes, the password store, the meal plan, the kids' school forms. People love it. People recommend it. People build a life around it.
The team is offered an acquisition. The price is meaningful for the team. They take it.
The acquirer is large. The acquirer's interest is in something — a feature, a patent, the team itself — that does not need the standalone product to keep existing. The product enters a period the press will later call "winding down". The product's existing users receive an email explaining that, in the spirit of integration with the acquirer's broader offering, the product will close on a future date, and an export tool will be provided.
You export. You receive a zip file of data in a shape no other product can read directly. You spend a Sunday afternoon trying to find a replacement. You find one. You re-enter the data, or some of it. The new product is not quite as good as the old one. You stay anyway, because the cost of moving again is the cost of moving again, and you are tired.
Eighteen months later, the new product is acquired.
Across the household-software field of the last fifteen years, this pattern has run dozens of times. Pick a domain. Calendars: Sunrise. Tempo. Cue. Mynd. Tasks: Wunderlist. Astrid. Catch Notes. Springpad. Notes: Vesper. Catch. Lists of beautiful little notes apps with thoughtful, careful design, each of them quietly vanishing on a Wednesday afternoon with thirty days of warning. Photo libraries that were not made by Apple or Google have had a difficult decade.
The Australian household has seen its own version, and the Australian household has felt it slightly harder, because the local market is smaller and the loss is sharper when there are not three alternatives waiting. The local budgeting app two adults loved for tracking joint money without forcing a single account, acquired by a larger fintech a few years back; the standalone product slowly wound down inside the acquirer's stable. The local meal-plan service that was your sister's favourite, gone — closed on a Tuesday in March, with thirty days' notice and an export tool that produced a CSV nobody else's app would read. The mortgage tracker your accountant recommended, sold to a bank, withdrawn from public sale within eighteen months — the data became part of the bank's offering instead, which was not the same product and required a bank login the household member who had been using the tracker did not have. The recipe site that ran for fifteen years and then closed on a quiet announcement in March, with the export tool producing a zip nobody could read in any other app, and the founder writing a heartfelt blog post that ended with thanks-for-the-memories and a forwarding link to a recommended replacement which itself closed eight months later.
If you have been paying attention to Australian household-tech for fifteen years, you have lived through enough of these closures to recognise the email shape before you open it. The subject lines are similar. Important news about your account. Changes to [product name]. A message from our founder. You open it and you already know, before reading the first paragraph, that you have a window — usually thirty days, sometimes sixty if the founder is being generous — to extract what you can. You set a reminder. You forget to act on the reminder. The window closes. The data is gone.
When the pattern is named at all, the framing is acquirer-centric. The story is told as "Microsoft acquires Wunderlist." The story is rarely told as "the shopping list that twenty million households used will close in eighteen months." The press release names the new home. The press release does not, as a rule, name what is being left behind.
What is being left behind, when this pattern runs, is one of two things. Sometimes it is just data — a list of recipes, a list of dates, a folder of notes. Data can be recovered, with effort, into a new home. The cost is the Sunday afternoon and the partial loss and the years of accumulated context that does not survive the export.
Sometimes, though, what is being left behind is more than data. It is a practice. The way you and your partner had divided the week. The way the older child had learned to check the shared list before adding anything to it. The way Sunday morning had organised itself around the shape of the app. When the app goes, the practice goes too, because the practice was held in the app's particular shape. You cannot port a Sunday-morning rhythm into a different shape.
That is the harder loss, and it is the one the press releases do not mention.
The Domestic Lens
To understand why this hurts the household more than other domains, you have to look at what makes the household different from work.
A workplace tool is replaceable. A workplace tool has an owner — the company — and a refresh cycle. A workplace tool's lifespan is measured in fiscal years. When a workplace tool gets replaced, somebody decides, somebody migrates the team, somebody runs the change. The transition is somebody's job. It might be a bad transition. It might be a long transition. But it is a project, and the project has an owner.
A household tool is different. A household tool has no owner separate from the household itself. There is no IT department. There is no procurement cycle. There is no refresh cadence. When a household tool gets replaced, the work of replacing it falls on whoever in the house has the energy to do it, on a Sunday afternoon, between other things. The transition is not a project. It is friction. It is friction the household pays for, in evenings and weekends and the kind of small, accumulating tiredness that erodes how you feel about the things you keep.
There is more. A household has time horizons workplaces do not have. A workplace's planning horizon is the quarter, maybe the year. A household's planning horizon is the marriage, the children, the long arc of ageing in place. The recipe collection your mother started thirty-five years ago is still being added to. The address book has names in it going back twenty years. The school records folder has been growing since the first child started kindy, and will be needed until the youngest finishes school. The lease folder is the lease folder. The first-aid box is the first-aid box.
The household keeps things that workplaces do not keep, and it keeps them for longer than workplaces keep anything. The wooden spoon in the kitchen drawer has been there for twenty years. The bread tin is older than the youngest child. The salad servers are the wedding present from somebody's aunt. None of it is replaced because none of it has stopped working. The household has its own quiet rule: if a thing still does its job, the household keeps it.
This is the value system the household applies to its software. It applies the same rule. It does not occur to most households to replace a piece of software that is still doing its job, any more than it would occur to most households to replace the bread tin. The bread tin lasts as long as the household lasts. So should the calendar.
But software does not work this way. Software is sold on a different value system — one that measures success in growth, in funding rounds, in acquirer interest. The household's measurement is duration. The software industry's measurement is velocity. The two systems disagree about everything. The disagreement is not visible at any single moment. It only becomes visible in cumulative form, fifteen years on, when you look at the apps you used to depend on and find that almost none of them are still there.
The kitchen drawer keeps the same scissors for forty years. The household's software has not lasted four.
Imagine a Sunday morning in a household you know — yours, perhaps, or one you have stayed in. The two adults are in the kitchen. One is making coffee. The other is leaning on the bench, looking at a recipe taped to the inside of a cupboard door. The recipe is handwritten on lined paper, the ink faded slightly along the bottom edge where the cupboard door has rubbed against it for years. It is from one of their mothers. Neither of them remembers when it was put there. The cupboard has been opened twenty thousand times. The recipe is still there.
The kettle clicks off. The wooden spoon is in the drawer where it always is. The bread tin, on the shelf, has been on that shelf for longer than the youngest child has been alive. The lease folder, in the third drawer down of the desk in the hallway, has been added to once a year for a decade, and every page is still there.
One of them picks up their phone, to check what is on this week. The calendar app shows the week. The calendar app does not show the week from 2017, because the calendar app from 2017 is not the calendar app being used now. The calendar app from 2017 was Sunrise. Sunrise had a feature, much-loved at the time, where it pulled in the family birthdays and the school holidays and the public-holiday adjustments without being asked. The household has not had that feature in any of the calendar apps it has used since. The household has also not had the events it accepted between 2014 and 2016, because the migration from Sunrise was imperfect, and the destination's later promises did not include the original events.
The kettle is the same kettle. The kitchen drawer has the same scissors. The recipe is the same recipe. The calendar is a different calendar.
This is the moment the household first notices — if it notices at all — that the things it owns are not all equally permanent. The physical objects in the house behave according to a rule the household has known since it became a household: a thing kept well, kept. The digital objects in the house behave according to a different rule, one the household did not write and was not told about: a thing kept well, gone in five years.
Most households do not articulate this. They feel it. They feel it as a slight tax, a slight friction, a slight wrong-feeling when the calendar app changes again, when the shopping list moves to a different home, when the photo library asks them to sign in to a service they did not know they had been opted into. The feeling does not have a name in the productivity vocabulary, because the productivity vocabulary is not addressing it. The feeling does have a name in the domestic vocabulary, where the relevant word is kept. Things that are kept are kept. Things that are not kept have not been chosen.
A household-software product designed for kept is a different product from a household-software product designed for more efficient. The two products look superficially alike. They have lists, calendars, reminders, search. The difference is that one of them was built to still be there in twenty years, and the other was built to be acquired in three. The household, given the choice, picks the first one. The household has, mostly, not been given the choice.
There is a domestic register for things that stay. Australian houses still have it. You can see it in the un-thrown-away serving dishes, in the wedding-present linens, in the recipe cards in your grandmother's handwriting that are still legible because they were laminated by an aunt in 1987 who happened to have a laminator. The register is permanence as warmth. Things that have been kept are warmer than things that are new, because they have been chosen by the household, by accumulation, over time.
Software that disappears every two years cannot enter this register. It is the wrong material. Software that lasts can. But to last, it has to be designed for lasting, which is something almost no piece of household software has been.
Why It Keeps Happening
The pattern is not bad luck. It is structural.
There are four forces. They are well-known. They are rarely named together as the explanation for the household's particular suffering.
The first force is acquisition as the dominant exit. Almost every venture-funded software company is built with the assumption that it will, at some future date, be acquired. Acquisition is the standard ending. The team's compensation, the investor's return, the founder's freedom — all of these are anchored to the acquisition outcome. If the acquisition does not happen, the company is, by the industry's own measure, a disappointment, even if it is doing what it set out to do. The standalone, slowly-growing, never-acquired household-app company is rare enough to be remarked on. Most acquirers do not particularly want the product they have acquired. They want the team, the patents, the user base, the integration story. The product is incidental. After the acquisition closes, the product enters a quiet hospice.
The second force is subscription churn as the operating model. Modern household software runs on subscriptions, because that is what investors have learned to value. A subscription business needs constant new growth to offset constant churn. Constant new growth means constant marketing, constant feature releases, constant pricing experiments, constant attempts to find adjacent markets. None of these things are bad in themselves. But they pull the product away from the household's pace. The household does not want a new feature every six weeks. The household wants the feature it has to keep working, for years, without becoming worse. The growth requirement and the household's value system disagree at the molecular level. The product, over time, drifts toward the growth requirement, because that is what the company is being measured on. The household watches its tool become a different tool. Eventually the household stops using it.
The third force is the talent market. The people who build household software are, on average, in their late twenties or early thirties when they ship their best work. They are talented; they care; they make beautiful things. But they have not run a household for forty years. They have not held onto a kitchen drawer of inherited objects for two decades. They do not, yet, feel in their own bodies what it is to want a tool to be the same tool in twenty years' time. They will feel it, in another twenty years, when their own households have accumulated. By then they will be working on something else. The next generation will be in their late twenties or early thirties, making beautiful things, not yet having felt it.
The fourth force is less visible than the others, and it is the one that closes the loop. When an acquirer buys a household-software company, the deal model rarely includes the existing user base as a primary asset. Acquirers buy for one of three reasons, in roughly this order: the team, the technology, the market segment. The product's existing users are mentioned in the press release as a number, often as evidence that the team can build things people want, but the number is a symbol of capability, not the thing being purchased. The acquirer does not, in most cases, intend to keep the standalone product running for those users. The acquirer intends to absorb the team into a larger effort, port the relevant technology into a different surface, or use the market segment as a beachhead for an adjacent product the acquirer was going to build anyway.
The existing users, in this model, are a quantity that will decay. The decay is built into the deal terms. The decay is not, by itself, regrettable to the acquirer, because the deal was not premised on the users staying. The deal was premised on the team, the technology, or the market segment — all three of which can be extracted without the standalone product surviving. When the product is, after a year or two, wound down, the press release frames it as a strategic decision. The strategic decision was made on the day the deal closed; the wind-down is the deferred execution.
The first three forces explain why companies sell. The fourth force explains why the products always die after they sell. The acquirer's deal model does not have a budget for keeping a quietly-running side product alive for the original users, because the original users were never the thing being bought. The product enters the acquirer's portfolio with no internal champion whose career depends on the product staying alive. Within eighteen months, almost without exception, there is no one inside the acquirer's company whose job is to keep the household using the product. Within thirty-six months, the product is gone.
These four forces are not anybody's fault. They are not a conspiracy. They are not a moral failing on the part of any individual founder or acquirer or team. They are simply the operating physics of how household software has been built. The acquisition is the standard exit; the subscription is the standard model; the talent is, on average, young; the acquirer's deal model treats the existing users as decay. Put them together and you get the pattern.
The pattern does not run because anyone wants it to. The pattern runs because none of the people involved have an incentive strong enough to interrupt it. The founders are doing what their training and their funding ask them to do. The acquirers are doing what their corporate development teams ask them to do. The users are doing what users do, which is to find the best available tool and hope for the best.
To interrupt the pattern, you need a different operating system for the company — one whose measurement is duration rather than velocity, whose exit is no exit, whose model does not require constant growth, whose people have themselves lived inside a household long enough to feel what staying means.
That is a small list of companies. It has not historically included most household-software companies. But it is not an impossible list. The household has a small number of brands that have served this register well — the kitchen-equipment companies that have made the same bread tin for sixty years, the linen-makers whose sheets are still the sheets, the publishing houses whose recipe books are still in print fifty years after first publication. Software has not yet produced its equivalents at scale. It could.
What Stays Looks Like
Set aside the cynicism for a moment. The pattern is the pattern, but it is not a law of physics. It is a description of how things have tended to go. The interesting question is what household software would look like if it were built differently — if it were built by people whose explicit aim was that it should still be there, doing the same job, in twenty years.
This is not a thought experiment. It is the shape of a particular bet some of us are making. Here is what the bet looks like, point by point.
It is privately held.
The single most important structural feature is ownership. A privately-held company can choose its own pace. It can decline acquisition offers. It does not have an investor base whose returns depend on a future liquidity event. It can grow slowly, deliberately, in the direction the household needs, rather than in the direction the funding round needs. This is not a small thing. It is the foundation everything else sits on. A household-software company with venture-capital ownership is, structurally, on a path that ends in acquisition or closure. A privately-held one is on a path that ends only when the people running it choose to end it. The household has a right to know which kind it is using, before it makes the tool central to its week.
It has no acquirer.
This is the second piece, and it is harder than the first. Privately-held does not, by itself, mean unsellable. Many privately-held companies are sold the moment a number large enough is offered. To be the kind of software that stays, a company has to make a more particular commitment: that no number will be enough. This sounds like an unusual claim to make publicly, and it is. But it is a claim the household has a right to hear. The household is being asked to make a long-term commitment to the tool. The tool has to be willing to make a long-term commitment back. "We are not for sale" is the form that commitment takes. Without it, the household is making a commitment that the tool may, at any future date, refuse to keep.
It is in a jurisdiction the household can find.
There is a quieter piece. Households increasingly use software whose operating entity is in a jurisdiction they cannot easily reach. When something goes wrong — when a price changes, when an account is closed, when data is lost — the household has limited recourse. The household has a right to know who is responsible, in what country, under what consumer-protection law. Household software that stays is software whose company can be named, whose office can be visited, whose jurisdiction is the household's own. For an Australian household, that means Australian. The household will not, in practice, ever visit the office. But the option of visiting matters. The option is the structural protection.
It exports cleanly.
Even if all of the above holds, the household still has a right to its own data. Export must be a first-class operation, not a buried compliance feature. Export must produce data in a format that other tools can read. Export must include not only the rows but the context — the dates, the people, the structure, the relationships between things — that make the data useful in another tool. Export must be testable; the household must be able to perform a complete export and inspect the result, before it ever has reason to need one. A household-software company that has not made this work, in the open, before it has any users, has not yet finished building the product. Export is part of the product. It is what makes the household's commitment to the tool a bounded commitment rather than an open-ended one. The bounded commitment is the only commitment a careful household will give.
It improves slowly.
This is the piece the industry will find hardest, and it is the piece the household will recognise immediately. Software that stays does not get rewritten every two years. It does not change its colour scheme on a whim. It does not move the button you have used twenty thousand times. It does not bury a feature in a settings page accessible only through a new gesture. It improves, but it improves at the pace at which households improve, which is to say slowly, in the direction of being more like itself rather than more like something else. Australian houses do this well. The good Australian house is not renovated every five years; it is tended, looked after, gradually made more itself over decades. Household software at its best does the same thing. The tools that have stayed in your house for fifteen years are tools that have been tended, not optimised.
It is honest about what is not there.
There is one more piece. A household-software company that wants to stay has to be honest about its limits. It cannot promise everything. It cannot pretend to be the household's only tool. It must say, plainly, what it does and what it does not, and it must be willing to say that there are other tools, and other companies, that the household may want for other purposes. The honesty is the trust-building. A company that overpromises is a company that is, by its own measurement, going to disappoint. A company that scopes itself tightly and keeps that scope for a long time is a company that can be relied on. The household has a finely-tuned sense of overclaim. It does not forgive a company that has overclaimed.
And, underneath all of these, it is willing to be small.
This is the foundational feature. It is the one the other six rest on. It is also the one the industry will find hardest to write into a press release.
A household-software company that wants to stay has to be willing to be small. Not as a target, not as a marketing position, but as a structural commitment about what success looks like. The household-grade company does not need to be the largest household-software company in the world; it does not need to dominate its category; it does not need to reach every household. It needs to reach the households it can serve well, charge them a price they can afford, run on a budget that allows the team to keep working without external funding rounds, and stay that shape for years.
A household-software company that is unwilling to be small ends up, eventually, doing something that breaks the other six features. It accepts an investment round, which puts a clock on the privately-held commitment. It enters a market segment beyond its competence, which breaks the honest-about-limits feature. It rewrites the interface to chase a new user demographic, which breaks the improves-slowly feature. The other six features are the visible commitments. The willingness to be small is the invisible commitment that makes them durable.
This is the hardest commitment to verify from the outside. A company can claim it without meaning it. A company can mean it for a few years and then quietly stop meaning it when the funding environment changes. The household has to read the company over time, watch what it does, and decide whether the willingness is structural or rhetorical. The signals are slow ones: did the company take outside investment when offered; did the company expand its scope when expansion was available; did the company keep the price stable when prices in the category rose. Each of these is one data point. The household needs several years of data points before it can know.
Software that stays starts with the willingness to be small, and earns trust by demonstrating it, slowly, over years.
If you put these seven features together — privately held, no acquirer, in a jurisdiction the household can find, exporting cleanly, improving slowly, honest about its limits, and willing to be small — you do not get a fast-growing software company. You get something else. You get a household-grade company. You get the software equivalent of the bread tin.
This is what is meant, in some quieter corners of the field, by marathon business. The phrase comes out of the European Mittelstand tradition — the family-held, decades-old manufacturing companies that have furnished the workshops and the kitchens of half a continent for generations, without ever being household names themselves. The Mittelstand companies do not appear in the technology press. They appear in the kitchen drawer, in the linen cupboard, in the tool shed. They appear, in fact, in exactly the places where household software ought to appear and has not yet learned to.
A household-software company that takes the marathon shape is doing something the industry has not, historically, rewarded. It is, in particular, doing something the industry's measurement systems have trouble seeing. Its growth curve is unremarkable. Its acquirer-interest is, by design, zero. Its press coverage is, on average, sparse. It is not the kind of company that an investor will tell you about at a dinner party.
But it is the kind of company that, twenty years from now, will still have your calendar.
Closing
Go back to the phone in your hand. Imagine, for a moment, the apps that would be on the first screen ten years from now, if the pattern were interrupted.
Some of the apps would be different. You would have replaced some of them, on purpose. That is healthy. The household chooses what it keeps; the household is allowed to change its mind.
But many of them would be the same. The calendar would be the same. The list would be the same. The papers would be in the same place. The kids' school records would be in the same place. The recipe collection would be the same recipe collection, with another ten years of additions, still readable, still exportable, still in the same place.
It would feel different to have a phone like this. It would feel like the kitchen drawer. The wooden spoon would be in there. The scissors. The corkscrew you've had since before the children. Each thing would carry the weight of having been chosen, and kept, and tended, for as long as the household has been a household.
There is no urgency about a household-software industry that produces this kind of phone. The industry will not produce it because households have asked for it. It will produce it, if at all, because some small number of household-software companies decide to be the kind of company that produces it — privately held, no acquirer, in jurisdiction, exporting cleanly, improving slowly, honest about limits.
The bet is that this is possible, on a long enough horizon. The bet is also that some households are looking for exactly this, and have been looking for it for a long time, and will recognise it when it appears.
The household has always known what it wants from the things it keeps. The household wants them to stay.
That is the whole essay. That is the whole bet.
The household that runs is the household whose tools stay.