It Works for People Who Hate Technology

Software is designed, almost without exception, by people who like software. This is not a scandal and it is not avoidable, since building something requires caring about it enough to spend years on the details. But it produces a consistent blind spot, and the blind spot is visible in nearly every product that has ever asked a new user to complete a tour. The assumptions run all the way down: that discovering a feature is enjoyable, that a settings panel is an invitation, that a term like workspace or sync conflict or nested tag carries obvious meaning, that a person who has not used the product for a month will happily re-learn where things moved. Each assumption is reasonable about someone who finds technology interesting. Each is false about the majority of people who actually need the tool.

The majority is large and it is not composed of the incompetent. It contains bakers, ship officers, nurses, electricians, farmers, pastors, hairdressers, mechanics, and the enormous population of professionals whose expertise is deep, hard-won, and entirely unrelated to computing. These people are not intimidated by complexity; they handle more of it in an ordinary Tuesday than most software imposes in a year. What they lack is patience for complexity that does not belong to their work. A baker who can hold four proofing schedules, six supplier relationships, and the behaviour of dough at three humidities in their head is not confused by a note-taking application. They have simply decided, correctly, that their attention belongs somewhere else.

What that population needs is frequently described, wrongly, as a simpler tool, and the industry response to that description is usually to remove capability. This is the wrong operation. A tool with fewer features is not easier to use; it is merely less useful, and it fails these users for a second time when their needs grow past it. The correct goal is considerably harder and much rarer: capability that imposes no cost when it is not being used. A feature you have never opened should be invisible rather than intimidating, and the tool should not require its user to have an opinion about it.

There is a second requirement, less discussed and arguably more important, which concerns explicability. People whose work carries real consequence develop a strong instinct about tools they cannot explain. A ship officer will not rely on equipment whose behaviour they could not describe to a colleague at three in the morning. A tradesperson will not trust a component they cannot reason about. This instinct is routinely misread as technophobia, and it is nothing of the sort. It is the same professional judgment that makes them good at their actual job, applied to a new object, and it produces a demand that most software cannot satisfy: tell me where my work is, in terms I can verify by looking.

Most software fails that demand not through malice but through architecture. The honest answer to where does my work live is frequently a paragraph involving accounts, regions, synchronization states, and caches, and no part of that paragraph can be confirmed by the user opening anything and seeing it. The result is a relationship built on assurance rather than observation, which sophisticated users tolerate because they understand the tradeoff and which non-technical users experience as a low-grade unease that never quite resolves. They keep a paper copy. They print things. They are told this is irrational, and it is not: they are compensating for a system they cannot inspect.

The third requirement is that the tool must survive absence. Professionals who do not enjoy technology do not use their tools continuously and lovingly. They use them in bursts, abandon them for a fortnight during a busy period, and return expecting to find things as they left them. Any product that punishes a gap, by having moved its interface, changed its terminology, or introduced a modal explaining what is new, has just spent that user’s remaining goodwill. The tools that survive in this population are the ones that are the same on the ninetieth day as on the first.

Put together, these requirements describe a design goal that is considerably harder than either simplicity or power taken alone: full capability that costs nothing when unused, an architecture a non-specialist can verify by opening a folder, and an interface stable enough to be forgotten and returned to. When that combination is achieved, something distinctive happens. People who describe themselves, cheerfully and repeatedly, as not technology people end up operating systems of real sophistication, without tutorials, without support, and without ever having thought of themselves as having learned anything.

Two of them, six thousand miles apart and in professions with nothing whatsoever in common, agreed to compare notes.

Two people who both said it in the first ninety seconds

Yasmin set up the call, because Yasmin set up everything, and because she had decided some weeks earlier that her father would enjoy this more than he would admit in advance.

‘I do not need a translator,’ Amir Hassan said, as she angled the laptop on the steel worktable at the back of the bakery.

‘You need somebody to press the green button.’

‘I can press a green button.’

‘You can press a green button,’ Yasmin agreed. ‘You have pressed several. Some of them were the right one.’

It was the middle of the afternoon in Toronto, which was the dead hour, after the morning trade and before anyone came in wanting something for dinner. The ovens were off. The room smelled the way it always did at that hour, of cooling bread and the particular sweetness that gets into the brick after enough years. Amir had wiped down the table twice before sitting at it, which Yasmin had noticed and not mentioned.

The call connected. On the other side was a small office with painted breeze block walls, a window with a view of container cranes, and a woman in uniform trousers and a plain shirt sitting very upright in a plastic chair.

‘Captain Nkechi Obi,’ she said.

‘Amir Hassan. I have a bakery. This is my daughter Yasmin, who arranged this and who will now go away.’

‘I will go away in a minute,’ said Yasmin, off-camera.

Nkechi smiled. ‘You are in a kitchen.’

‘I am in the back of a bakery,’ Amir said. ‘I would apologise for the background but it is the cleanest room I own. Where are you?’

‘Durban. Port office. We are alongside for four days, which means I have a chair, a door that closes, and a connection that works most of the time, and I intend to use all three before we sail.’ She glanced at something off-screen. ‘I should say at the start that I am not really a technology person, so if this call was arranged expecting expertise, somebody has been misled.’

Amir sat back and laughed, genuinely, for the first time.

‘I said exactly that to Yasmin this morning,’ he said. ‘Word for word. I told her, do not put me on a call with a computer person, I will have nothing to say and I will be embarrassed.’

‘Then we are two people who have been tricked in the same direction.’

From somewhere behind the camera Yasmin said, ‘Nobody has been tricked, you are both extremely capable and this is why I set this up,’ and then, audibly, left.

There was a short pause of the kind that happens when two strangers have to decide who goes first.

‘May I ask something rude,’ Nkechi said.

‘Please.’

‘How long did it take before you stopped expecting it to go wrong?’

Amir considered the question seriously, which she appreciated.

‘About five weeks,’ he said. ‘And I can tell you the exact moment. I had not opened it for nine days, because we had a wedding order and a fridge failure in the same week and I did not open anything except the fridge. And on the tenth day I opened the laptop expecting to have to remember how it worked, and it was exactly the same, and everything was there, and nothing had asked me to update anything or read anything.’

‘That is precisely my answer,’ Nkechi said, ‘except mine was six weeks and it was a crossing rather than a wedding.’

‘You were away longer.’

‘Twenty-two days with no connection at all,’ she said. ‘Which for most software is a death sentence. I came back into range and nothing had happened. Nothing was waiting for me. Nothing had queued up, nothing had failed to sync, nothing wanted my attention.’ She paused. ‘I do not think I can explain to somebody who works ashore how unusual that is.’

‘I have never been to sea,’ Amir said, ‘and I understand you completely.’

What a bakery actually needs to remember

‘Show me,’ Nkechi said. ‘I want to see what a bakery keeps.’

Amir turned the laptop, decided that was not going to work, and shared his screen instead, which took two attempts and produced a brief commentary from off-camera that he ignored with dignity.

‘Four pages at the top,’ he said. ‘Recipes. Suppliers. Daily. Business.’

‘That is fewer than I expected.’

‘It was more, at the start, because Yasmin made a beautiful structure with many branches and I did not use it.’ He said this without any embarrassment. ‘I told her it looked like somebody else’s filing cabinet. So we cut it back to what I actually reach for, and then it grew again, but underneath, where it belongs.’

He opened Recipes, and beneath it were sub-pages: Breads, Pastry, Cakes, Seasonal, and one called Failures which made Nkechi lean in.

‘Explain the last one.’

‘Everything that did not work and why.’ Amir opened it. ‘Nine entries. The sourdough that would not rise in the winter of the bad radiator. The croissant batch I ruined by trusting a thermometer that lied. A cake somebody ordered that I should not have agreed to make.’ He shrugged. ‘I have been baking for twenty-two years and I have made every mistake at least twice, and I only stopped making them a third time when I started writing them down.’

He opened a recipe entry and the body filled with a formatted document: a heading, a list of ingredients with quantities, numbered steps in bold where the timing mattered, and a table at the bottom.

‘This looks like a recipe book,’ Nkechi said.

‘It looks like a recipe book because I made it look like a recipe book, which took no learning at all.’ He pointed at the toolbar. ‘Bold, lists, headings, a table. I use maybe six of those buttons. The rest are there and I have never touched them and they do not bother me.’

‘What is the table for?’

‘Batch sizes.’ He scrolled to it. ‘One column for a single batch, one for a double, one for the big Saturday quantity. Because the arithmetic is where mistakes happen. Anybody can multiply by two. Nobody multiplies eleven ingredients by two point five at five in the morning without getting one of them wrong.’

Beneath the body were collapsed section bars, which Nkechi asked about immediately because she used the same feature differently.

‘Three sections on every recipe,’ Amir said. ‘Notes on the dough, which is feel, temperature, what it should look like. Notes on the oven, because my two ovens are not the same and one of them lies by fifteen degrees. And Customers, which is who orders this, whether anyone has an allergy, and that the family on Beaumont wants it darker than anybody else in the city wants anything.’

‘That last one is not a recipe.’

‘That last one is the business,’ Amir said. ‘The recipe is the easy part. The recipe has been the same for twenty years. What changes is who wants it and how, and if I lose that I am a bakery that makes correct bread for the wrong people.’

He moved to Suppliers, where each entry was a company and the attachments were invoices.

‘Every invoice, photographed with the phone and attached to the supplier.’ He opened one and a scanned invoice sat inside. ‘Which was Yasmin’s idea and which I resisted for two weeks because it seemed like extra work, and then the flour price went up and I could see, in about four seconds, that it had gone up three times in fourteen months and that nobody had mentioned it twice.’

‘Did you do anything about it?’

‘I had a conversation,’ Amir said, ‘with numbers in it, and it went differently to the conversation I would have had with a feeling in it.’

Nkechi laughed properly at that.

‘The daily notes?’

‘Short.’ He opened Daily, which had entries by date. ‘Four lines most days. What sold out, what did not, the weather, anything a customer said that stuck. That is it. And it looks like nothing, and after two years it is the most valuable page I have, because now I know that the second Saturday in October is always bigger than I expect and I stopped being surprised by it.’

‘You are describing a log.’

‘I am describing what my father kept in a notebook by the till,’ Amir said. ‘He had thirty years of them in a cupboard. I could not read most of them, because his handwriting was a private language, and I could not search them at all, because they were paper. That is the only difference. The practice is his.’

He showed her labels last, of which he had five: order, seasonal, price, allergy, and one simply called check.

‘Allergy is the important one,’ he said, and his tone changed slightly. ‘Anything involving a customer with a serious allergy carries it. When somebody new works the counter, that filter is the first thing I show them. Nine entries. It takes two minutes and it is the most important two minutes of their training.’

‘And check?’

‘Things I am not sure about and have not resolved.’ He grinned. ‘Yasmin says this is bad practice and that I should resolve them.’

‘Yasmin is right and it does not matter,’ Nkechi said. ‘I have the same label under a different name.’

What a ship keeps, and why it cannot be lost

Nkechi shared next, and her tree was as spare as his.

‘Three pages,’ she said. ‘Ship Operations. Certification Study. Personal. That is the whole top level and it has not changed in a year and a half.’

‘Three,’ Amir repeated. ‘I feel better about four.’

‘The depth is underneath.’ She opened Ship Operations, and beneath it sat pages for inspections, drills, cargo operations, maintenance, and handovers. ‘This is the formal spine of my work and it is arranged the way the work is arranged, which is not a design decision, it is just what the job looks like.’

She opened an inspection entry and it had six section bars.

‘Sections by area of the vessel,’ she said. ‘Forward, engine spaces, accommodation, deck, cargo holds, safety equipment. Every inspection entry, the same six, always in that order.’

‘Why always the same?’

‘Because I am reading them at four in the morning, tired, looking for one thing.’ She said this as a matter of fact rather than a complaint. ‘If the fourth bar is always deck, my hand goes there before my eyes have finished reading. And because if a section is empty, that means I did not inspect it, and that is something I need to know rather than something I want hidden.’

Amir made a small noise.

‘I have that,’ he said. ‘Not for safety. For the oven notes. If the oven section is empty on a recipe, it means I have never written down what that oven does to that bread, and that is a gap.’

‘It is the same instinct,’ Nkechi said. ‘An empty box asks a question. A blank page does not.’

Drills were next, with attachments: signed muster sheets, photographs of equipment, a scanned checklist.

‘These are records,’ she said. ‘They have to exist, they have to be findable, and they have to be defensible if somebody official asks. And the reason they are here rather than in a binder is that a binder cannot be searched and a binder stays on the ship when I do not.’

Amir asked the question he had been holding since the beginning.

‘What happens if the computer breaks?’

Nkechi looked genuinely pleased to be asked.

‘That is the correct question and almost nobody asks it,’ she said. ‘My material is files in a folder. Not inside the program. In a folder, on the machine, which I chose when I set it up. So every Sunday I copy that folder to a memory stick, and there is a second one in the ship’s safe, and if this laptop went into the sea tomorrow I would lose a laptop.’

‘That is it? Copy a folder?’

‘That is it. Copy a folder.’ She spread her hands. ‘I have been on ships where a system failed and the answer involved contacting somebody ashore, and there is nobody ashore at three degrees south. Whatever I depend on has to be something I can restore myself with what I have on board.’

She showed him Certification Study, which had entries with due dates, and then the timetable, which he had not seen before.

‘My renewals,’ she said. ‘Certificates expire. That is not an inconvenience, that is my licence to work. So each one is an entry with a due date, and the calendar shows me the year, and the sidebar tells me what is coming.’

‘And it just knows?’

‘It knows because I told it once,’ Nkechi said. ‘Which is the only kind of knowing I trust.’

Then she showed him Threads, and this was the part Amir did not immediately understand.

‘During a watch I cannot sit and write properly,’ she said. ‘I am standing, I am moving, something has happened and I have perhaps eleven seconds before the next thing. So I open this and I type one line into it, like a message.’

Amir watched her do it. It looked, to him, like texting.

‘It looks like texting.’

‘It is exactly like texting and that is why it works,’ she said. ‘Nobody has to learn to send a message. And the point is not that the line is good. The point is that the line exists, with a time on it, so that at the end of the watch I can turn six lines into one proper entry while I still remember what they meant.’

‘And if you do not?’

‘Then I have six lines with times on them, which is enormously better than nothing, which is what I had for the first four years of my career.’ She shrugged. ‘Half of what I capture never becomes anything. That is fine. The half that does is the half I would otherwise have lost.’

Amir was quiet for a second.

‘I need that,’ he said. ‘At five in the morning my hands are in dough and something occurs to me and by seven it is gone. Every day. For twenty-two years.’

‘Then that is the thing you should take from this call.’

Asking it a question in ordinary words

‘May I test something,’ Nkechi said. ‘I want to see whether the thing I use works the same way for you, because I suspect we use it differently and I would like to know.’

‘Test away.’

‘When you want something and you do not know where it is, what do you do?’

‘I ask it,’ Amir said, as though this were obvious, and then paused. ‘Is that not what everyone does?’

‘Show me what asking looks like for you.’

He opened the panel at the side and typed, slowly, with two fingers, the sentence: what did I do about the flour that was too wet.

Nkechi watched the results appear.

‘You typed a question,’ she said. ‘An actual question, in a sentence.’

‘I did. Should I not have?’

‘You should absolutely have, I am simply noting that you did it without anybody telling you to, which is not how most people approach a search box.’ She was leaning toward her camera. ‘What came back?’

‘Three things.’ Amir read them out. ‘A recipe entry with a note in the dough section about a wet delivery in the spring. A supplier entry, because I complained to them and wrote down what they said. And one of the failures, which is the batch I threw away before I understood the problem.’

‘And the order?’

‘The recipe first. The supplier second. The failure last.’ He shrugged. ‘I do not know why that is the order and it is the order I would have chosen.’

‘It is the order because of where the words are,’ Nkechi said. ‘That was explained to me once and it stuck. A match in the title counts for the most, then the labels, then any text pulled out of pictures inside the entry, then the body, then the sections, then the files attached to the entry, then the files attached to a section.’

Amir absorbed that.

‘So the photographs of my invoices.’

‘Are readable, and count, and sit at the bottom of the ordering, which is correct, because an invoice with a word in it is less about that word than a note titled with it.’

‘I did not know the pictures were readable,’ Amir said slowly. ‘I have been searching for supplier names and finding things and I assumed I had written the names somewhere.’

‘Try one you never typed.’

He thought, and typed a company name he was fairly certain appeared only on paperwork. Two results returned, both invoices attached to supplier entries.

‘Well,’ he said, after a moment. ‘That is a thing I have owned for a year without knowing.’

‘This is the pattern of this entire call,’ Nkechi said, amused. ‘Mine is less impressive and more urgent. Show me yours, I said, and here is mine: I ask it things like what was the finding on the forward hold in the last inspection. Full sentence. Because at the moment I need that I am not in a state to construct a clever query, I am in a state to ask a colleague a question, and this is the only piece of software I have used where the two are the same thing.’

‘And it answers.’

‘It answers, and when it does not answer well the reason is always my titles.’ She was matter of fact about it. ‘When I title an entry properly, describing what it actually is, it comes back first every time. When I title it something lazy, it hides in the body and I have to work. Which taught me to title things carefully, and nobody had to tell me to.’

‘The program taught you.’

‘The program did not teach me anything,’ Nkechi said. ‘It rewarded me and I noticed. That is different and I think it is better.’

Everything they learned by pressing something wrong

It was Nkechi who started it, and she started it because she was still slightly embarrassed about her own example.

‘I want to ask what you found by accident,’ she said. ‘Because I have a list, and I would like to not be the only one.’

‘I will go first and you will feel much better,’ Amir said. ‘The round button.’

‘The one in the corner.’

‘The one in the corner.’ He was already laughing. ‘For about three weeks I kept touching it by mistake. My hands are large and the laptop is on a steel table and I lean. And every time I touched it something opened, and I closed it, and this happened perhaps twenty times.’

‘Did you ask Yasmin?’

‘I did not ask Yasmin, because I did not want to be a man who asks his daughter about a button.’ He said this with enormous dignity. ‘So instead I created eleven empty notes over three weeks and blamed the computer.’

Nkechi put a hand over her face.

‘And then?’

‘And then I touched it on purpose, once, to see what would happen if I did not close it.’ He opened it now to demonstrate: a new entry, immediately, from anywhere. ‘It is for making a note quickly. That is all it is. And it is now the thing I use more than anything else in the whole program, because that is exactly what I need at five in the morning. Not a menu. Not a page. One button, write the thing, close it.’

‘You spent three weeks fighting the feature you needed most.’

‘I spent three weeks fighting it, and then I used it twice on the first day and I have used it every day since.’ Amir shook his head. ‘And Yasmin says the lesson is that I should have asked her. I say the lesson is that the button was in the right place and I was wrong.’

‘I think you are both correct and it is more interesting than either of you is saying,’ Nkechi said. ‘It found you. You did not read about it. You did not go looking. It was in the way, in the good sense, until you understood it.’

‘Now yours.’

‘The tabs in the sidebar.’ She showed them: Recent, Due, Expiring, Tools. ‘I had used this thing for four months without touching any of them. I assumed they were a menu of things I did not need, because that is what tabs usually are.’

‘What happened?’

‘I clicked Due while trying to click something else, on a bad night in the Gulf of Aden with a headache.’ She said it flatly and Amir laughed. ‘And it showed me everything with a date coming up, all in one list. Every certificate renewal. Every follow-up I had promised somebody. Every inspection that was falling due.’

‘You already knew all of that.’

‘I knew all of it and I was carrying all of it,’ Nkechi said, and the distinction was clearly one she had thought about. ‘There is a difference between knowing a thing and having to hold it. I had eleven dates in my head at all times, and I checked them mentally about four times a day, and I did not know I was doing it until the list existed and I stopped.’

‘That is not a small thing.’

‘It is not a small thing at all, and I found it by misclicking.’ She spread her hands. ‘Recent I use constantly now, because most of what I want is something I touched yesterday. Expiring I use once a month with some anxiety.’

‘My second one is the typing,’ Amir said.

‘The search?’

‘I was typing a supplier’s name in the box at the top, expecting to press something afterward, and before I had finished the name a small list appeared underneath with the thing I wanted in it.’ He mimed the moment. ‘And I said out loud, to an empty bakery, that the computer was reading my mind.’

‘It was not reading your mind.’

‘It was not reading my mind, it was looking at what I had typed so far, and I understand that now, and at the time I genuinely put my hands up.’ He grinned. ‘I use it constantly. I almost never do a proper search. I type four letters and the thing is there and I click it. My daughter says I use the fastest path in the program by accident.’

‘Mine is the least dramatic and the one that changed my behaviour most,’ Nkechi said.

‘Go.’

‘I used to press save.’ She said it as a confession. ‘Constantly. After every sentence, sometimes after every line, because everything I have ever used has punished me for not saving and I have lost work at sea in ways that mattered.’

‘And it saves itself.’

‘It saves itself, and there is a small indicator that says so, and I did not notice the indicator for something like two months because I was too busy pressing.’ She laughed at herself. ‘One evening I noticed it change while my hands were nowhere near the keyboard. And I sat there and watched it for a while, and then I tested it, deliberately, by writing something and closing the machine without saving and opening it again.’

‘Did it survive?’

‘All of it.’ She nodded slowly. ‘And then I did it again a week later, because trusting something once is not trusting it. And then I stopped pressing.’

‘Two months of pressing a button that was not needed.’

‘Two months, and I would do it again,’ Nkechi said. ‘I am not going to apologise for testing something before relying on it. But I will say that the indicator being there is what let me stop, because the alternative was hoping, and I do not hope about records.’

Amir was nodding.

‘My last one is smaller,’ he said. ‘The star.’

‘Favorites.’

‘I thought it was decoration. A little star. What is that for.’ He clicked one. ‘And then Yasmin used it on my four base doughs, and now there is a place in the side where those four things live, and they are the four things I open more than everything else combined.’

‘So the thing you thought was decoration is your front page.’

‘The thing I thought was decoration is my front page,’ Amir agreed. ‘I would like it recorded that I have never once been right about what a feature is for before using it.’

The long list of things neither of them touches

‘Can I say something that may be impolite,’ Amir said.

‘You have been polite for forty minutes and I have been waiting.’

‘There is a great deal in here that I do not use.’ He gestured at the sidebar. ‘A great deal. And I want to say clearly that this is not a complaint, because for about a month I thought it was one.’

‘Say the list.’

‘Encryption. I do not use it. I am a bakery. If somebody steals my sourdough method they will still have to get up at four in the morning for twenty years, and honestly they are welcome to it.’ He counted on his fingers. ‘Version history, I have never opened. The board with the cards, which Yasmin showed me and which is clearly excellent and which I closed and have not returned to. The filters with the dates and the file types, which I understand and have never needed because I have four pages and I can find things.’

‘That is a substantial list.’

‘And here is the thing I want to say about it,’ Amir said, and he leaned toward the camera. ‘None of it is in my way. Not one of them. They are not on my screen. They do not ask me anything. Nothing has ever suggested I try them. There is no red dot telling me I have not discovered something.’

Nkechi was nodding before he finished.

‘That is the observation I would put at the centre of this conversation,’ she said. ‘Because my list is different from yours and the experience is identical.’

‘Say yours.’

‘The feed reader. I have never once opened it and I know exactly why, which is that at sea there is nothing to read from and ashore I have four days and other priorities.’ She counted her own. ‘The suggestions that show related entries, which I have looked at twice and which are clearly for somebody with a larger library than mine. The charts. There are charts. I saw them once, they were handsome, and I have never had a question that a chart would answer.’

‘Do you feel you are using it wrong?’

‘I felt that for about a fortnight,’ Nkechi said. ‘There is a habit people have with a new tool, which is to assume that value equals coverage, and that if you are using four things out of forty you are somehow getting a tenth of it.’

‘That is exactly what I thought.’

‘It is nonsense, and I worked out why while doing something else.’ She shifted in the plastic chair. ‘I use perhaps six things from that program. Those six things are worth what I paid several times over. The other thirty-four are worth exactly nothing to me and they cost me exactly nothing, which means the arithmetic is fine. What would not be fine is if the thirty-four were in my way.’

‘The board with the cards would be in my way if it were on the front,’ Amir said.

‘The charts would be in my way if something congratulated me for looking at them.’ She said this with some feeling. ‘That is the thing that drives people like us away, and it is never the complexity itself. It is the software being disappointed in you. The prompt asking whether you have tried something. The tour that will not close. The notification about a feature you did not ask for.’

‘I have left two things because of that,’ Amir said. ‘Not because they were hard. Because they kept talking to me.’

‘And this one does not talk to you.’

‘This one has never once talked to me,’ he said, ‘and I have used it every working day for over a year, and I could not tell you the name of a single thing I have not used, because I have never been introduced to any of them.’

Nkechi laughed.

‘That is the highest praise I have heard anyone give software.’

‘It is the only praise I know how to give,’ Amir said. ‘It stayed out of my way while I worked.’

Trust, explained by people who cannot explain encryption

The conversation turned, without either of them steering it, toward the thing underneath.

‘I want to tell you why I actually trust it,’ Nkechi said, ‘and I want to warn you that my reason is not technical, because I am not technical.’

‘Mine is not either. Go.’

‘I can explain it.’ She said it simply. ‘That is the whole reason. On a ship there is a rule that everybody knows and almost nobody states, which is that you do not rely on something in an emergency if you cannot explain how it works. Not because explaining is useful during the emergency. Because if you cannot explain it, you do not actually know its limits, and a thing whose limits you do not know will surprise you at the worst possible moment.’

‘And you can explain this.’

‘In three sentences, to anybody, including a cadet on their first voyage.’ She held up fingers. ‘It is one file. I open it and it uses a folder that I chose. Everything I write is in that folder as files.’ She dropped her hand. ‘That is it. That is the whole explanation and there is nothing behind it that I have left out because it was too complicated.’

Amir was quiet for a moment.

‘I have a version of that,’ he said, ‘and it is less dignified.’

‘Say it anyway.’

‘I opened the folder.’ He was slightly sheepish. ‘Not for any good reason. I did not trust it, in the first week, and I could not have told you what I was afraid of. So I opened the folder on the laptop to see whether anything was actually there.’

‘And?’

‘And there was a file, and there were my notes, and there was a folder with my invoice photographs in it, with the pictures just sitting there as pictures.’ He shrugged. ‘I opened one of my notes in the thing that opens text, the plain one, and my writing was in it. Not in a code. My writing.’

‘That is exactly the test I ran,’ Nkechi said, ‘and I ran it for exactly the same unnamed reason.’

‘Yasmin thought it was very funny.’

‘Yasmin is young,’ Nkechi said, not unkindly. ‘People who grew up with everything in the cloud do not have the instinct, because they never had the earlier arrangement to compare it against. My mother kept records in ledgers. My grandmother kept them in her head and a tin. The idea that a record exists somewhere I cannot go and look at it is genuinely strange to me, and I do not think that instinct is old-fashioned. I think it is correct and everybody stopped honouring it.’

‘I have paper in a drawer,’ Amir admitted. ‘Still. From before.’

‘Do you add to it?’

‘I stopped about eight months ago and I did not notice I had stopped until my wife asked why the drawer was not full.’ He looked slightly surprised at himself. ‘Which I suppose is the answer to the question of whether I trust it.’

Nkechi nodded.

‘That is the only measure that means anything,’ she said. ‘Not whether you say you trust something. Whether you stopped keeping a backup you made out of fear.’

They compared what they had each been afraid of, which turned out to be the same fear in two costumes.

‘Mine was that it would stop,’ Amir said. ‘Not break. Stop. That one day it would want something from me. Money, a password, an update, an account, and I would be standing in a bakery at five in the morning unable to see my own recipes because a company somewhere had decided something.’

‘Mine was distance,’ said Nkechi. ‘That the thing I needed would be somewhere else. Which is not paranoia at sea, it is the ordinary condition. Everything is somewhere else. The company is somewhere else. The office is somewhere else. If my records are also somewhere else then in the moment I need them I have nothing.’

‘And the answer to both is the same answer.’

‘The answer to both is that the folder is here,’ Nkechi said, ‘and nothing has to be asked, and nothing can be withdrawn. I do not need to understand encryption to understand that. I need to understand that a folder is a folder.’

Amir laughed.

‘I am going to say something and you will think I am being sentimental.’

‘I will allow it. We are nearly finished.’

‘My father kept everything in notebooks,’ he said. ‘Thirty years of them, by the till, in his own handwriting. And when he died I had the notebooks and I could not use them. They were his and they stayed his.’ He turned a hand over. ‘What I have now is the same practice, and I can read it, and Yasmin can read it, and if something happens to me tomorrow it is a folder on a laptop and somebody opens it and everything is in ordinary words.’

Nkechi did not say anything for a second.

‘That is not sentimental,’ she said. ‘That is succession planning, and it is more than most companies I have worked for manage.’

The invitation, and what they each took away

The port office had begun to fill with the noise of somebody else’s shift change, and Nkechi glanced at the door twice before Amir noticed.

‘You have to go.’

‘I have about four minutes and then I have to be somewhere with a hard hat on.’

‘Then take this,’ he said. ‘If you ever dock in Toronto, you come to the bakery. Not to buy anything. You come in the back, you sit at this table, and I feed you until you object.’

‘I will hold you to that and I do not object easily,’ Nkechi said. ‘And I will bring you spices from here, because you will have opinions about them and I would like to hear the opinions.’

‘I will have opinions.’

‘I am counting on it.’ She was gathering her things. ‘I am taking the invoice habit, by the way. Photograph and attach, at the supplier. We have stores paperwork that goes into a folder nobody opens, and it should be attached to the thing it concerns.’

‘And I am taking the messages,’ Amir said. ‘The one line during the watch. Because five in the morning with my hands in dough is my watch, and I have been losing thoughts every single day for twenty-two years and treating that as normal.’

‘It is normal. It is also fixable.’

‘Yasmin is going to be unbearable,’ he said. ‘She has been telling me to write things down mid-task for a year.’

From somewhere off camera, immediately, came the words, ‘I have been in the room the entire time.’

Both of them laughed, and Nkechi stood, and the call ended the way calls between people who have enjoyed each other end, with slightly too much waving.

The person who set it up, briefly on camera

Yasmin came back into frame after it ended, took the laptop, and closed three things her father had left open.

‘That went well.’

‘She is coming to the bakery,’ Amir said. ‘Eventually. When a ship allows it.’

‘I heard. I heard all of it.’ She sat on the edge of the table. ‘Can I say the thing I have been not saying for a year?’

‘I cannot stop you and I never could.’

‘When I set this up for you, I was ready for it to fail,’ Yasmin said. ‘Not because of you. Because I had already tried twice with two other things, and both times you used them for about nine days and then stopped, and both times you were polite about it and I could tell you were relieved.’

Amir considered this.

‘That is true and I did not think you had noticed.’

‘I noticed.’ She shrugged. ‘And the reason I think this one held is that I did not sell it to you. I did not say it would change the business. I said one sentence, which is that it was one file, and then I did the setup and got out of the way.’

‘You made a beautiful structure that I dismantled.’

‘I made a beautiful structure that was wrong, because I do not run a bakery,’ Yasmin said. ‘That was my actual mistake and I have thought about it a lot. I built the shape I would have wanted, with branches for everything, and you looked at it and said it was somebody else’s cupboard, and you were right. What worked was four pages that matched what you already thought about.’

‘So the lesson is that you should have asked me.’

‘The lesson is that whoever sets this up for somebody should build almost nothing and let them grow it.’ She pulled up the sidebar. ‘Look at this now. There are sub-pages under Recipes that I did not make. There is a page called Failures that I would never have thought of and which is the best idea in the whole vault. You built all of that without me and you did not notice you were doing it.’

‘Because I was not filing,’ Amir said. ‘I was writing down the bread.’

‘That is exactly it.’ Yasmin was grinning. ‘And I will tell you the other thing I did that mattered, which is that I never once explained a feature to you. I set it up, I showed you how to make a note, and then I refused to answer questions for the first month unless you were actually stuck.’

‘That was deliberate?’

‘That was completely deliberate and slightly cruel and it worked,’ she said. ‘Every time I explain something to you, you learn a fact and forget it. Every time you find something yourself, including by pressing it eleven times by accident, you own it forever.’

Amir looked at the screen for a moment.

‘The round button.’

‘I watched you fight the round button for three weeks,’ Yasmin said, ‘and it was the hardest thing I have ever done, and I would do it again.’

The hardest design goal there is

The temptation, reading a conversation like that, is to conclude that the tool succeeded by being simple. That conclusion is wrong in an instructive way, because neither of those systems is simple. A bakery running four page branches, structured recipe entries with three standard sections each, supplier records with years of attached invoices, a two-year daily log, and a five-label vocabulary including one reserved for allergies is operating a genuine knowledge management system. A ship officer running inspection entries with six fixed sections, drill records with evidentiary attachments, certification tracking with due dates and a calendar, quick capture during watch, and a weekly backup discipline is operating something a great many organizations would struggle to maintain. Neither person would describe what they have built as sophisticated, and both would be wrong.

What the tool achieved was not simplicity but the far harder property of capability that costs nothing when unused. Both of them listed, without prompting and without complaint, substantial features they have never touched: encryption, version history, boards, advanced filters, feed reading, related-entry suggestions, analytics. What matters is the second half of what they said. None of it is on the screen. Nothing prompts them about it. No indicator suggests they have failed to discover something. The features exist entirely for other people, and the cost of their existence to these two users is zero. That is the design achievement, and removing those features to make the product simpler would have made it worse for everybody without making it better for them.

The second observation is about discovery, and it undercuts most of what the industry believes about onboarding. Between them, these two users found the quick capture button, the sidebar time tabs, typeahead search, autosave, and favorites, and every one of those discoveries happened by accident, several of them after weeks of misunderstanding. Nobody read documentation. Nobody completed a tour. What made the discoveries stick was that each feature turned out to be exactly what the person already needed, sitting where they were already looking. A quick capture button that a baker keeps brushing at five in the morning has been placed correctly, even during the three weeks he was fighting it. A due list found by misclicking on a bad night removed a mental burden its user had not known she was carrying. This is discovery through use rather than through instruction, and it is slower, funnier, and considerably more durable.

The third and most important observation concerns trust, and it arrives here from a completely different direction than it usually does. Neither of these people can evaluate a cipher. Both of them tested the system in precisely the same way, independently, in their first weeks and for reasons neither could articulate at the time: they opened the folder and looked. There were files. There were photographs sitting there as photographs. A note opened in a plain text editor and contained ordinary words. That test is available to anyone who can open a folder, and it converts an abstract promise into something confirmed by sight. It is the same argument a security consultant makes about minimizing dependencies, expressed as a ship officer’s rule that you do not rely on what you cannot explain, and as a baker’s quiet decision to stop adding to the paper in the drawer.

That last measure is the honest one. Not what somebody says about a tool, but whether they stopped maintaining the fearful backup they had been keeping alongside it. A drawer that stopped filling. A system explicable in three sentences to a cadet on their first voyage. A folder that can be copied to a memory stick on a Sunday, from which everything can be restored by a person alone in the middle of an ocean with no support line to call.

Two people, six thousand miles apart, in professions with nothing in common, each built a working system without a tutorial, without a support ticket, and without a single moment of frustration directed at the tool. Both describe themselves, cheerfully and often, as not technology people. Neither of them learned anything they would call technical. That outcome is not luck and it is not a low bar cleared by a simple product. It is what it looks like when software is designed for people whose attention belongs somewhere else.

FAQs

Why do most tools fail people who do not enjoy technology?

Because they are designed by people who do enjoy it, and the assumptions run all the way through the product. Feature discovery is treated as a pleasure, settings panels as an invitation, terminology as self-evident, and returning after a gap as an opportunity to introduce what has changed. Every one of those assumptions is reasonable about someone who finds technology interesting and false about someone who finds it tolerable. The users in question are rarely intimidated by complexity, since most handle far more of it in their actual work. They have simply decided their attention belongs elsewhere, which is a judgment rather than a limitation.

Is a simpler tool the answer for non-technical users?

Usually not, because reducing capability fails those users twice: once when the tool cannot do something they need, and again when their requirements grow and they have to migrate. The property that actually helps is capability that imposes no cost when unused. A feature nobody has opened should be invisible rather than intimidating, and the product should never require a user to have an opinion about it. That combination is considerably harder to build than either a minimal tool or a powerful one, which is why it is rare, and why users who find it tend to stay for years. The test to apply is not how many features a product has but whether any of them can be ignored without consequence, which is a very different question.

What makes a tool explicable enough to trust?

The ability to state how it works in a few plain sentences that leave nothing important out. Professionals whose work carries consequence develop a strong instinct against relying on anything they cannot explain, not because the explanation helps in the moment but because an unexplained system has limits they cannot predict. A description like one file, a folder chosen by the user, and every note stored inside that folder as a file satisfies this instinct completely, because each clause can be verified by looking rather than accepted on assurance. Most software cannot offer a comparable sentence. The instinct behind this is routinely misread as technophobia when it is in fact the same professional judgment these people apply everywhere else in their work.

What is the folder test and why do non-technical users perform it?

It is the simple act of opening the storage folder and looking at what is actually there, and users frequently do it in their first weeks without being able to explain why. What they find is a file for the application, a repository file holding entries and structure, markdown sidecar files containing note bodies, and a folder of attachments in their original formats, all readable with ordinary tools. This converts a promise into an observation. It requires no technical skill whatsoever, only the ability to open a folder, and it is the single most reassuring thing many users report doing. Doing it once tends to settle the question permanently, which is why it is worth suggesting to anyone who seems uneasy but cannot say why.

How do you know when someone actually trusts a tool?

Not from what they say, but from whether they stopped maintaining the parallel backup they were keeping out of fear. Non-technical users very commonly run a shadow system for the first months: a paper notebook, a drawer of printouts, a duplicate spreadsheet. That redundancy is rational rather than superstitious, since it compensates for a system they cannot inspect. The moment worth noticing is when the shadow system quietly stops growing, often without the person registering it until somebody else points out that the drawer has not filled. That moment is usually noticed by somebody else first, which is a good reason to ask a new user whether their paper backup is still growing rather than asking whether they like the software. The answer to the first question is considerably more informative.

What should a small business actually keep in a system like this?

The operational knowledge that currently lives in one person’s head. For a bakery that means recipes with formatted steps and batch conversion tables, sections for dough behaviour and oven quirks and customer preferences, supplier records with photographed invoices attached, a short daily log, and a handful of labels covering the categories that matter. The most valuable page is frequently the one recording failures, since mistakes repeated a third time are the expensive ones, and a written record is what breaks that cycle. None of this requires technical knowledge, only the decision to write things down where they can be found. Starting with whatever currently lives only in one person’s head is the fastest route to something useful, since that material is both the most valuable and the most at risk.

Why attach invoices and paperwork to the supplier rather than filing them separately?

Because the useful question is almost never about a document in isolation. It is about a relationship over time: how often a price has changed, whether a delivery problem is a pattern, what was agreed and when. Photographing an invoice and attaching it to the supplier entry takes seconds and makes that history retrievable in one place. Users regularly report that this single habit produced a conversation with numbers in it rather than a conversation based on an impression, and impressions about costs tend to be wrong in the direction that favours the other party. It also means paperwork stops needing good filenames, since the document is identified by what it is attached to rather than by what somebody typed in a hurry.

How does a short daily log become valuable?

Through accumulation rather than through any individual entry. Four lines a day recording what sold, what did not, the weather, and anything a customer said looks like nothing for the first several months. After a year or two it becomes the only record of the seasonal shape of a business, and it turns recurring surprises into expectations. The practice is old and predates computers entirely; the difference is that a written log can be searched, read by somebody other than its author, and survive the person who kept it, which handwritten notebooks in a cupboard generally cannot. Keeping it deliberately short is what makes it survive, since a log requiring ten minutes will be skipped on precisely the busy days that are most worth recording.

What does a shipboard or field system need that an office system does not?

Independence from everything outside it. At sea there is no support line, no colleague ashore who can fix something, and frequently no connection at all for weeks. Anything relied upon must be restorable by one person with what is physically present, which in practice means data stored as files in a known folder and a routine that copies that folder to removable media. It also means the tool must survive long absences without demanding attention on return, since a system that queues up failures during a crossing has effectively failed during the crossing. The requirement extends to the tool itself: anything demanding a connection, an account, or a periodic check simply cannot be relied upon in those conditions.

How should backups work for someone with no technical support available?

As simply as possible, since a procedure nobody can perform alone is not a backup. Copying the vault folder to a memory stick on a fixed schedule, keeping a second copy somewhere physically separate, and confirming occasionally that the copy opens correctly covers the realistic failure modes. Because the folder contains ordinary files rather than an application database, the copy is immediately usable rather than requiring an import step. A hardware failure then costs a piece of hardware rather than a working archive, which is the whole objective. Testing the copy occasionally by opening it matters as much as making it, since an untested backup is a hope rather than a plan. A fixed day of the week works better than a resolution to do it regularly.

What is quick capture during a shift or a watch used for?

Recording something in seconds while standing, moving, or holding something else, when writing a proper entry is impossible. A chat-style capture surface works because nobody has to learn how to send a message, and each line carries a timestamp. The purpose is not to produce good notes; it is to ensure the thought exists at all so it can be turned into a real entry later while the context is still fresh. Practitioners generally accept that a substantial share of captured lines never become anything, since the ones that do would otherwise have been lost entirely. Because each line carries a timestamp, even captures that never become entries retain enough context to be interpreted later, which is the difference between a rough note and a useless one.

Why does the quick capture button work so well for people who dislike software?

Because it removes every decision except the writing. There is no menu to navigate, no page to select, and no filing to do at the moment of capture, which matters when a person has a few seconds and both hands occupied. Placement is what makes it discoverable: users frequently find it by pressing it accidentally, sometimes repeatedly over weeks, before realising what it does. A control that keeps getting touched by someone leaning over a work surface has been positioned correctly, even during the period when they are still confused by it. It also lowers the threshold for writing anything at all, which matters because the material most often lost is the thought that seemed too small to justify opening a program for.

What are the sidebar time tabs and why do they matter?

They provide views organised by time rather than by structure: recently modified entries, items with upcoming due dates, and things approaching expiry. Their value is that they remove a mental burden the user may not know they are carrying. Somebody tracking eleven renewal dates in their head, and checking them mentally several times a day, is doing continuous unpaid work that a list eliminates entirely. The recent view tends to become the most used of the three, since a large share of everyday retrieval is for something touched within the last day or two. They also require no configuration, which is part of why people find them by accident and keep them: nothing has to be set up before the view becomes useful.

Why does typeahead search suit non-technical users so well?

Because it shortens the path from intention to result and requires no concept of searching at all. Typing a few characters into the main box produces suggestions drawn from titles, bodies, labels, attachment names, and attachment content, and the desired item is frequently visible before the word is finished. Users regularly describe the experience as uncanny the first time it happens. In practice it handles most everyday retrieval, meaning the full search surface is reserved for genuinely difficult questions rather than routine lookups. It also teaches the shape of the library incidentally, since seeing what appears alongside a search reminds people of material they had forgotten owning. Nothing about it has to be learned, which is why it tends to become the most-used path within days.

Why does an autosave indicator matter more than autosave itself?

Because trust requires evidence rather than assurance. Many users arrive with a habit of pressing save compulsively, formed by earlier tools that punished them for not doing so, and no amount of documentation breaks that habit. A visible status showing that work has been saved lets them observe the behaviour, test it deliberately by closing and reopening, and then stop. That progression, observe, test, then rely, is exactly how careful professionals adopt anything, and a silent autosave gives them nothing to observe. The habit it replaces is deeply ingrained and often born of real losses, so it deserves to be dismantled by evidence rather than dismissed as unnecessary caution. Users who test it deliberately and then stop pressing describe the change as a genuine reduction in background anxiety.

What are favorites best used for?

The small number of items opened far more often than everything else. For most people that is between three and six things: base recipes, a recurring checklist, a reference document consulted weekly. Placing them in a dedicated panel turns the sidebar into a front page for the actual working set. Users commonly dismiss the star as decorative before someone applies it to their most-used entries, at which point it becomes the fastest route into the system and one of the features they would miss most if removed. Reviewing the starred set occasionally is worthwhile, since the working set changes over time and a favourites panel full of items nobody opens has stopped being a front page. Keeping it to a handful is what preserves its usefulness.

Is it a problem to use only a small fraction of a tool’s features?

No, and the arithmetic explains why. If six features are worth more than the price several times over, and the remaining thirty-four are worth nothing to that user and cost nothing, the trade is straightforwardly good. The assumption that value equals coverage is common among new users and produces unnecessary guilt. What would genuinely be a problem is if the unused features imposed a cost by occupying screen space, generating prompts, or implying failure, which is a different matter entirely from their mere existence. The guilt is common among new users and worth naming, because it occasionally drives people to adopt features they do not need and then abandon the system as too complicated. Coverage is not the measure; the value of what is used is.

What actually drives non-technical users away from software?

Rarely complexity, and very frequently the software expressing an opinion about them. Prompts asking whether they have tried something, tours that resist being closed, notifications about features they did not request, indicators implying incomplete discovery, and congratulation for ordinary actions all produce the same reaction. Users describe this as the product talking to them, and it exhausts goodwill faster than any difficulty in the interface. The tools that survive in this population are the ones that stay quiet while their user works and are identical on return after a long absence. The pattern is consistent enough to be a useful evaluation criterion: a product that has opinions about how much of itself you have explored will eventually exhaust the patience of anyone whose attention belongs elsewhere.

How do standardized sections help someone who is not systematic by temperament?

By making the structure carry the discipline instead of the person. When every entry of a type has the same sections in the same order, the writer knows what to fill in and the reader knows where to look, and this remains true when tired or interrupted. The most valuable property is that an empty section is visible: a blank space in prose says nothing, whereas a section heading with nothing beneath it plainly indicates something was not done. Users in very different fields independently describe this as the feature that changed their practice. Because the template supplies the questions, consistency stops depending on discipline and becomes a property of the form itself, which is what makes it survive tired days.

Can a system like this be handed on to somebody else?

More easily than most alternatives, because the material is ordinary files in a folder written in plain words rather than records inside a service. There is no account to transfer, no subscription to maintain, and no company to contact. A successor receives the folder and opens it. This matters particularly in family businesses and in roles where records must outlive the individual holding them, and it addresses a problem many people have experienced personally: inheriting notebooks in someone else’s handwriting that remain, in every practical sense, that person’s property forever. Leaving a short note explaining the structure alongside the material makes the handover better still, and it takes a few minutes to write.

Does using a tool like this require learning anything technical?

Almost nothing, judging by how people describe the process. The setup consists of opening one file and choosing a folder. Formatting is done with a toolbar whose buttons behave as expected, and most users report relying on a handful of them. The features that end up mattering most are typically discovered by accident during ordinary use rather than learned deliberately. What does require effort is the human part: deciding what is worth writing down and developing the habit of doing it, which is a working practice rather than a technical skill. The people who succeed with it are generally those who decide what is worth recording rather than those with any particular aptitude for software.

How long does it take before someone stops expecting the tool to fail?

Commonly around five or six weeks, and users can usually identify the exact moment. It tends to be the first return after a genuine gap: a busy fortnight, a long crossing, a period when nothing was opened at all. Finding everything unchanged, with nothing queued, nothing failed, and nothing demanding attention, does more for confidence than any amount of correct behaviour during daily use. Until that point people generally keep a parallel record, and the disappearance of that habit is the real marker of adoption. Knowing that the moment is coming makes the first weeks easier, since the appropriate response to early scepticism is to keep the parallel record and let it become unnecessary on its own.

What should someone write down first if they have never kept a system?

Whatever they currently keep in their head and worry about forgetting. For most people that is a short list: the details customers or colleagues have mentioned, the recurring problems that keep recurring, deadlines with consequences, and the reasons behind decisions that will be questioned later. Starting with the material that already causes anxiety produces an immediate return, which is what makes the habit survive its first fortnight. Structure can be added afterward, and adding it later is considerably easier than designing it in advance. Formatting, structure, and labels can all wait; the only thing that matters in the first fortnight is that the writing happens at all and can be found again afterward.

Can questions be asked in plain sentences rather than keywords?

Yes, and non-technical users frequently do it without being told they can, which is a reasonable indication that it matches how people naturally look for things. A full question such as what did I do about the flour that was too wet is treated as a set of terms weighted across the whole vault, and the results come back ordered by where the matches occurred. This matters most under pressure, when constructing a clever query is not realistic and the user is in the frame of mind to ask a colleague a question. The plain sentence and the effective query being the same thing removes a skill barrier entirely.

How does search ordering reward careful titling?

Matches are weighted by how specific they are likely to be: titles first, then labels, then text extracted from images inside an entry, then the body, then sections, then attachments and their names, then attachments belonging to a section. The practical consequence is that an entry titled to describe what it actually is will surface ahead of one titled lazily, whose content then has to be found through its prose. Users notice this pattern on their own and begin titling more carefully, which is a habit acquired through reward rather than instruction and tends to stick for that reason. Improving titles on existing entries is therefore one of the cheapest improvements available, since it lifts every future search at once without anything else changing.

Are photographed documents searchable by their contents?

Yes. Text is recovered from images and from scanned pages, so a photographed invoice, a signed checklist, or a handwritten note contributes its words to the same search as everything else. Users are commonly surprised by this, having assumed for months that they were finding documents through something they had typed themselves. It matters disproportionately in trades and small businesses where a large share of incoming paperwork arrives as a photograph taken on a phone, and it means the effort of good filenames becomes largely unnecessary. Accuracy depends on legibility and is rarely perfect, which does not matter much, because the objective is making the document findable rather than transcribing it. A single recognised company name or reference number is usually enough.

What is the best way to set this up for a parent, colleague, or employee?

Build almost nothing and let them grow it. The common mistake is constructing an elaborate structure that reflects how the person setting it up thinks, which the actual user recognises immediately as somebody else’s arrangement and quietly abandons. A small number of top-level pages matching categories the person already uses in conversation is enough. Structure added later by the user themselves tends to be better, because it reflects the real shape of their work, and it also produces ownership, which is what determines whether the system survives its first busy fortnight. Checking back after a month to see what they have added is more useful than any amount of instruction, since what they built themselves shows what they actually needed.

Should features be explained to a new non-technical user?

Sparingly, and there is a real argument for restraint. People who are shown a feature tend to learn a fact and forget it, while people who discover one themselves, sometimes after weeks of confusion, retain it permanently and use it confidently. The practical approach is to demonstrate how to create a note, ensure the setup is correct, and then answer questions only when the person is genuinely stuck rather than pre-empting them. Watching somebody misunderstand a control for several weeks is uncomfortable and generally produces a better long-term outcome than explaining it on day one. The exception is anything with a real cost to getting wrong, such as where the data folder lives and how to back it up, which should be shown clearly and confirmed.

What happens when someone returns after weeks away?

Nothing, which is the point and which users describe as the moment they stopped expecting failure. There is no synchronisation to reconcile, no queue of pending changes, no notice about what has changed in the interface, and no prompt requiring dismissal before work can resume. Everything is where it was left. For people who use their tools in bursts around genuinely busy periods, this property matters more than almost any feature, because a product that punishes a gap has spent its remaining goodwill at exactly the moment its user is least able to absorb the interruption. This is worth testing deliberately during evaluation, since a product’s behaviour after a long gap reveals more about whether it will survive real use than its behaviour on day one.

Why does an unchanging interface matter so much?

Because people who do not enjoy technology do not re-learn willingly, and every change spends credit they extend reluctantly. A control that has moved is not a small inconvenience to somebody who located it by habit rather than by understanding; it is evidence that the tool is not stable enough to stop thinking about. Stability also compounds: a person who has found the same button in the same place a hundred times stops registering the software as software at all, which is when it finally gets out of the way of the work. It also makes informal support possible, since a colleague or family member can help without first working out what has moved since they last looked.

How do labels help in a small business?

By marking the categories that cut across whatever page structure exists, and by staying few enough to be remembered. A short set covering order status, seasonal items, price changes, and anything requiring particular care is usually sufficient. The highest-value single label is often one marking safety or allergy information, since filtering by it produces a complete list in seconds and doubles as training material for anybody new to the counter. A vocabulary that grows beyond roughly eight tends to be applied inconsistently, which quietly removes its usefulness. Auditing the list occasionally and merging anything that has become a near duplicate keeps the set honest, and takes only a few minutes once or twice a year.

Is it worth recording failures and mistakes deliberately?

It is frequently the highest-return page in a small business vault. Mistakes repeated a third time are the expensive ones, and the pattern of repetition is invisible without a record, because each instance feels like bad luck in the moment. A page holding what went wrong and why converts scattered frustrations into a reference that can be consulted before a similar decision. Practitioners in very different fields report the same experience: the failures page looks morbid to visitors and is the one they would least willingly lose. It also makes training somebody else far easier, since the reasoning behind a rule is more persuasive than the rule itself when the person has to apply it under pressure.

Why keep customer or client detail alongside operational records?

Because the operational part is usually the stable one and the human part is where the variation lives. A recipe, a procedure, or a specification may go unchanged for years while the requirements of individual customers change constantly, and losing that second layer means doing correct work for the wrong people. Keeping it as a section within the same entry means the preference is encountered at the moment the work is being done rather than remembered separately, and it survives staff turnover, which is where informal knowledge normally disappears. Keeping it as a distinct section rather than mixed into the main text also means it can be read on its own by somebody covering a shift.

Does the same approach work for regulated record keeping?

It works well where the requirement is that records exist, are findable, and can be produced when asked. Drill records, inspection logs, and compliance checklists benefit from consistent sections and from attaching the signed or photographed evidence to the section it belongs to. Searchability is the addition a paper binder cannot offer, and the fact that records travel with the individual rather than remaining in a particular physical location is frequently valuable. Any specific regulatory obligations about retention and custody remain the responsibility of the organisation involved. Consistent sections help here as much as anywhere, since an inspection with an empty area plainly indicates that the area was not covered rather than leaving the omission invisible.

What does a good handover between shifts look like?

Consistency is what makes it work: the same sections in the same order on every entry, so the person taking over knows where to look before they have read anything. Anything unresolved should be stated as unresolved rather than omitted, since an incoming colleague cannot distinguish between nothing happened and nobody wrote it down. Quick capture during the shift feeds this directly, since a handful of timestamped lines can be assembled into a proper entry at the end while the details are still accurate rather than smoothed by memory. Writing it in the same shape every time is what allows the incoming person to read it quickly, which is the actual measure of whether a handover works.

How much does maintaining a system like this actually cost per day?

Most users describe it in minutes rather than in a block of time set aside. A few seconds of quick capture during the working period, and five to ten minutes at the end turning captures into entries or writing a short log. The work only survives when it is attached to the activity it records rather than treated as separate administration, which is why reflections written while tidying up and notes written while something is still running are the versions that actually get done. Set aside as a separate task, it is almost always the thing that gets dropped. The cost also falls over time as the habit forms, and most people report that the daily portion becomes automatic within about a month of consistent use.

What is the single most persuasive argument for this kind of tool?

That it can be explained completely in three plain sentences and every clause can be verified by looking. One file runs the application, a folder chosen by the user holds everything, and each note exists inside that folder as a file. Nothing important has been left out of that description, which is unusual, and it means a person with no technical background can confirm the arrangement themselves rather than accepting it on assurance. For anyone whose instinct is to distrust what they cannot explain, that property does more than any list of features. Everything else, including the capabilities that never get used, sits behind that sentence without complicating it, which is the design achievement worth noticing.

Does working offline actually matter for people who are always connected?

More than they expect, because the benefit is not only about connectivity. A tool that operates entirely from the local machine has no synchronisation state to reconcile, no server that can be slow, no account that can lock, and no maintenance window that arrives at an inconvenient moment. Users in well-connected cities describe the same relief as those at sea: nothing is ever waiting for them, and nothing can be unavailable at the moment they need it. The absence of an entire category of interruption is easy to overlook precisely because nothing happens, which is the point. It also removes any dependence on a provider’s availability, which is the kind of thing nobody thinks about until the morning it becomes the only thing that matters.

What is the value of everything living in one file plus one folder?

It collapses the number of things a person has to understand into something they can hold. There is no installer, no background process, no separate database, and no hidden application directory whose contents mean nothing to the person who owns them. The result is that the answer to where is my work is short enough to state and simple enough to verify. Users repeatedly identify this as the property that persuaded them to try it at all, since the alternative descriptions they had encountered required trusting a paragraph they could not check. It also makes moving to a new machine trivial, since copying the folder and the application file reproduces the entire working setup with nothing left behind.

How do people who dislike software usually discover features?

By accident, during ordinary use, and often after a period of misunderstanding. Almost nobody in this group reads documentation or completes a tour, and the features that end up mattering most are the ones encountered by pressing something unintentionally. What makes the discovery stick is that the feature turns out to be exactly what the person already needed, positioned where they were already reaching. This is slower than instruction and considerably more durable, because a capability someone worked out for themselves is one they use with confidence rather than one they half remember. The implication for anyone introducing the tool to somebody else is that placement matters far more than explanation, and patience matters more than either.

Is it a bad sign if a new user creates empty or messy entries at first?

Not at all, and it is frequently how the most valuable habits begin. Empty entries created by accident are a sign that a control is where the person’s hands naturally go, which is exactly where it should be. Messy early entries reflect somebody writing before they have decided what the system is for, which is the correct order. Cleaning up is trivial later, whereas a person who waits until they know the right way to use something usually never starts. Volume of imperfect material beats a small amount of careful material in the first months. Deleting the accidental ones takes seconds, and the habit of writing quickly is worth considerably more than a tidy first month.

What should be done with a system when the person who built it is unavailable?

Very little, if it was built sensibly, because the material is ordinary files in a known folder and requires nothing to open beyond the application file sitting alongside it. The practical preparation is to make sure at least one other person knows where the folder is, has a copy of any passwords used for encrypted entries, and has been shown how to open it once. That takes minutes and removes the single most common failure in small businesses and small teams, which is a working system that only one person can reach. Writing that information down somewhere outside the vault, such as with other business records, closes the last gap in the arrangement.

Does this approach suit people who work primarily with their hands?

It suits them particularly well, since the requirements are exactly the ones such work imposes: capture must take seconds, retrieval must survive interruption, and the system must not demand attention at moments when both hands are occupied. Quick capture, favourites for the handful of frequently used items, and plain-sentence search cover most of it. The material recorded tends to be practical rather than analytical, covering what worked, what did not, who needs what, and what to check next time, which is the kind of knowledge most likely to be lost otherwise. Voice notes and photographs attached to entries extend this further, since both can be captured in seconds and remain findable by their content afterward.

How does this compare to keeping a paper notebook?

The practice is identical and the difference is retrieval, legibility, and succession. A notebook by the till has served small businesses for generations and has three well-known failure modes: it cannot be searched, it can only be read by whoever wrote it, and it stops being usable when that person is no longer there. A digital equivalent preserves the habit while removing those limits, which is why people who inherited unreadable notebooks from a parent or predecessor tend to be the most enthusiastic adopters rather than the most resistant. Many people keep both for a period, which is sensible, and the paper version usually stops growing on its own once the digital one has proved itself.

What is the strongest sign that a tool is well designed for this audience?

That its users cannot name the features they do not use. Somebody who has worked with a product daily for a year and could not list what they have ignored has never been prompted, nudged, or made to feel incomplete by it. That silence is difficult to achieve, since almost every commercial incentive pushes in the opposite direction, and it is the property most often cited by people who describe themselves as disliking technology. The tool stayed out of the way while they worked, which is the whole of the praise they have to offer and the highest available. A useful question to ask any long-term user is what the product has ever asked them to do; a short answer is a very good sign.