Organize documents without folders: the starting point for people with no time

The daycare wants the form. Not just any form, the one from two years ago, with the signature you scribbled in the cloakroom between rain pants and a lunchbox. You know it exists. You just don't know whether it's in the kitchen cupboard, in an email attachment, or in the folder your wife set up, sorted by a logic only she understands.

I know that moment well. A job, a side business, two kids, an apartment we rent out. Four areas of life, and each one produces documents that somebody will want to see at some point. The tax advisor asks for a receipt. The electrician sent his invoice, at some point, by email, with a subject line that didn't contain the word invoice. For years I searched for things that were somewhere, just never where I happened to be sitting.

This page is for people wearing too many hats. For self-employed people without an accounting department who still have to survive an audit. And for couples where one person knows the system and the other stands helpless in front of the shelf when it matters.

The promise is modest: by the time you reach the bottom, you'll know what to keep, how to find it again, and how to get started in five minutes.

For context: I build Paperarchive, a tool for exactly this problem. This page stays tool-neutral until the last section anyway.

What document management actually solves

Storing is not the problem. Everyone can store. The problem is finding things again.

The cost of searching isn't the twenty minutes. The cost is the moment when someone is waiting. The landlord who has a question and signs with someone else if you don't answer by Thursday. The tax advisor whose follow-up sits for three weeks because you can't find the receipt and keep putting off the search. The daycare manager who smiles politely and quietly files you under "sends forms late".

There's a difference between "I have it somewhere" and "I have it right now". The first sentence is a promise to yourself. The second is an answer to the person asking. Document management, whichever kind, has exactly one job: turning the first sentence into the second.

With several hats, this explodes. The job has one filing logic. The side business has another. The family has none, it has a kitchen cupboard. The rented apartment has a folder that hasn't been opened since the purchase. Four systems, one head, and at nine in the evening that head is no longer in any shape to remember four systems.

Why folders fail

Folders demand something impossible: you have to know the structure before the document exists. Where does the invoice from the electrician who worked in the rented apartment go? Apartment? Tradespeople? Taxes, because it's deductible? All three are correct. The folder forces you to pick one answer, and in two years you have to remember which.

A colleague of mine has the problem in its purest form. His wife filed everything, neatly and completely, in folders. The system works. It only works for her. When she's not there, he finds nothing, and he has stopped trying. That's not a criticism of her. Every folder structure is the image of one head, and heads are hard to share.

My brother tried rules. A scanner with programmed menus: Inbox, Travel expenses, Receipts. Press a button, the document lands in the right folder. Then came the first document that belonged in two menus. Then the second. At some point you stop maintaining the rules, and from that day on everything lands in Inbox, which is the expensive name for a shoebox.

File names don't help either. scan_0047.pdf says nothing. Invoice_final_final2.pdf mostly says something about your afternoon.

My opinion, and you can see it differently: folders are not a character flaw. They're the wrong tool for things you only need every few months. For the car insurance policy you look for once a year, a folder is a memory test with poor odds.

Search instead of structure, the principle

The alternative is simple: file without deciding, find by content. When you need something, you type what you know about it, and the tool finds it because it has read the content.

For that, the content has to be readable. A scan without text recognition, called OCR, is a photo, not a document. Only OCR turns the photo into text you can search. That's the first ground rule for any tool: without text recognition, you have a digital shoebox.

The second level is semantic search, meaning search by meaning rather than exact words. My favorite example: you type "car insurance" and get the motor insurance policy even though the word "car" appears nowhere in the document. It says "motor vehicle", "liability", a license plate. A good search knows that's the same thing.

A little structure remains, just in the right place. Categories are fixed and the same for everyone: invoice, contract, insurance, official letter. Tags, on the other hand, are free and cut across everything. An invoice, a contract and a letter from a lawyer can all carry the tag "move", even though they sit in three different categories.

And instead of rules up front, there are corrections afterwards. If a document gets sorted wrong, you change the category, and the system learns to handle the next similar document correctly. That's the opposite of my brother's scanner menus.

For couples, this means something concrete: the other person doesn't have to know the system. They only have to know what they're looking for. "rental contract apartment" is enough.

What belongs in, and what doesn't

Roughly by area of life, these are the document types that belong in an archive:

• Invoices and receipts: tradespeople, online orders, anything deductible
• Contracts: rent, phone, electricity, gym, employment
• Insurance: policies, amendments, claims
• Doctors and health: findings, vaccination records, invoices if you're privately insured
• Authorities and taxes: assessments, returns, correspondence
• Daycare and school: contracts, forms, certificates for your tax return
• Vehicle: purchase contract, registration, workshop invoices, expert reports
• Apartment and house: purchase or rental contract, utility statements, protocols
• Work and pension: payslips, references, pension statements

By the way, the main source isn't the mailbox on the street. Most documents arrive as email attachments today, paper is the minority. So above all, the archive has to be able to accept email.

Paper still arrives, and for paper the rule is: scan once, then decide whether the original stays. For most receipts it can go. Certificates, notarized contracts and a few official documents stay on paper, and if you're self-employed in Germany, the rules for throwing things away are in the article on replacement scanning.

What doesn't need to go in: newsletters, advertising, shipping confirmations once the return window has passed. And here's an opinion, with a caveat: private bank statements don't need to go into the archive. The bank keeps them for ten years anyway. The caveat is twofold. If you switch banks, the access is gone, so an export before switching is worth it. And for the self-employed, bank statements are accounting records, they belong in the archive, no discussion.

How long you have to keep what

For private individuals in Germany the situation is more relaxed than most people think. There is almost no legal obligation. The one exception is invoices from tradespeople for work on your house or apartment, which you have to keep for two years. Everything else is a recommendation, and it goes roughly like this:

• Tradespeople invoices for building work: two years by law, closer to five for larger jobs because of warranty claims
• Tax documents and receipts: five years as a practical value
• Contracts and insurance: term plus two to three years
• Warranty receipts: two years, five for expensive purchases
• Employment contracts, references, pension documents, property documents: permanently

The details with the legal references and a table are in the article on retention periods for private individuals.

For the self-employed a different logic applies, and it's strict. Accounting records, meaning incoming and outgoing invoices, cash receipts and bank statements: eight years. Books, inventories and annual accounts: ten years. Business letters and emails with business content: six years. Periods change occasionally, the eight years for records only apply since 2025, before that it was ten. The overview for both worlds is in the article on retention periods, private and business. This isn't legal advice, when in doubt ask your tax advisor.

My practical rule runs across every retention table: when in doubt, keep it digitally. Storage costs nothing, searching costs time, and the document you deleted after the period expired is always exactly the one somebody asks about the following year. A vehicle damage assessor I spoke to lives off this: expert reports, photos, correspondence with insurers and lawyers, and the follow-up questions come years later. Deleting isn't an option for him, and I don't think it should be for private individuals either.

GoBD and e-invoices in five sentences (for the self-employed in Germany)

The GoBD are the German tax authorities' principles for how tax-relevant documents have to be kept digitally, and at their core they demand three things: the document must not be alterable without a trace, it must be traceable who did what with it and when, and the original must be preserved in the format it arrived in.

Since January 1, 2025, every business in Germany has to be able to receive e-invoices, small businesses included, and an e-invoice isn't a PDF with a logo but a structured data record: ZUGFeRD is a PDF with an XML file inside it, XRechnung is pure XML with no pretty view.

The mistake you can't see: if you print a ZUGFeRD invoice or re-save it via "Print to PDF", the XML is gone, and the file still looks the same.

I tested this with a real invoice from SevDesk, forwarded it to my archive address, downloaded it again and compared both files with SHA-256, a checksum that changes completely if a single bit is different: both were identical, the embedded XML was untouched, and that's what "preserve the original" means in practice.

This gives you a hard criterion for choosing a tool, namely whether it leaves files untouched and whether your tax advisor can access them directly instead of receiving ZIP files by email; the instructions for the hash test are in the article on archiving e-invoices.

The question everyone asks: who can see my documents?

The honest answer for every hosted service is: technically, there is always an admin access. At Google, Dropbox, iCloud and every small provider. Anyone who tells you otherwise either has end-to-end encryption, meaning encryption where only you hold the key and the provider can't read anything, and then search inside the content doesn't work. Or they're talking nonsense.

So the real question isn't "can he?" but "is it his business model?". And a business model isn't made of access, it's made of analysis. Advertising profiles. Usage analytics. Training AI models on your content. Sharing with partners. With free or ad-funded storage you pay with exactly what you file, and bank statements, medical letters and employment contracts are a pretty good currency.

What to look for, in five questions: How does the provider make money? Where are the servers? Is "no training on your data" in the contract or only in the blog? Does deleting really mean deleting, or does a copy remain? And is there even a function that lets staff read other people's documents, or was it never built?

My opinion, which you can see differently: documents with health data, contracts and bank statements don't belong in a tool whose operator earns money from data analysis. Not because anyone there is evil. But because the incentive points in the wrong direction, and in the long run incentives beat every good intention.

Why I know this so precisely: I didn't start Paperarchive as a product, but as a tool for my own problem. My family and business documents still live there in exactly the same way. That's why there is no analysis and no view of other people's documents, I would never have built something like that for myself. It only became a product when people around me kept approaching me with the same issue. Solo founder means: the subscription is the business model, not the data, and a single incident would end the product. The question "can you see my documents?" is still the one I hear most often. No, there is no function for it. Server locations, contracts and details are on the security page.

Briefly on chat tools: they're convenient for questions about a single document. They're not an archive, because they neither accept email nor keep things nor share them with your partner. And the question about training and storage location applies there too, if anything more sharply.

Working as a pair: partner, tax advisor, colleagues

The most common pattern is the duo. One does the bookkeeping or the family mail, the other still needs access. Not daily, but exactly when the first one is on vacation, in hospital or simply unreachable.

The solution is shared spaces instead of shared passwords. At home, my wife and I each have our own space for the things that only concern ourselves. Family is a shared space where we both see everything: daycare, insurance, the apartment, the pediatrician. Nobody has to know the other's password.

The same principle applies to the tax advisor, just with a different role. The tax advisor should be a user, not a recipient of ZIP files. One space for everything tax-related with read access, and the annual exercise of "collect all receipts and email them" disappears.

How to start, in five minutes

The most important advice first: don't start with the backlog. The backlog is the reason most people never start. Start with your inbox, that's where this week's documents are.

Step one: pick out three or four emails with attachments, the electricity bill, the insurance policy, the letter from school. Forward them to your tool's archive address. Don't rename, don't pre-sort. Forwarding is enough.

Step two: wait a moment and search for something that isn't in the file name. The amount. The name of the case worker. "car insurance". If that hits, you've understood what this is about. If it doesn't, the tool is wrong, not you.

Step three: forward everything new for a week, sort nothing. At the end of the week you have a small archive you've used, instead of a large structure you've built.

Only then comes the question of whether the backlog moves. A customer from Florida did this exemplarily. Years of nested folders, then three test documents first, then everything. When he moved apartments, the landlord had a question about his paperwork, he found it immediately, and the lease was done the same day instead of a week later.

And the scanner? Set up scan-to-email to the archive address, done. No programmed menus, no Inbox folder. The scanner is then just another sender.

Typical mistakes at the start

Wanting to build the perfect structure first. If you spend a week thinking about categories before the first document is in, you've missed the point.

Migrating everything at once, then never searching. Two weekends of scanning, 800 documents uploaded, and then the archive sits there like the folder before it. Just bigger.

Only filing, never searching. The aha moment is the search, not the filing. If you don't search, you never notice that it works.

Keeping the paper "just in case" anyway. Then you search twice, first digitally, then in the folder, because you don't trust the scan. Check once properly that the scan is complete, then throw it away.

"Organizing" documents with a chat tool and wondering why nothing comes back. A chat forgets, that's its nature.

There's a moment when it clicks. Not when uploading, not when setting up. It's the first search that hits. You type two words, the document is there, and you notice you didn't just think about where it might be. From then on you don't want to go back.

Everything up to here works with any tool that accepts email, recognizes text, leaves files unchanged, can share spaces and comes from a provider whose business model you understand. Paperarchive is my attempt to build exactly that. 30 days free, no card. Throw three documents in and search for them.

For comparison: the comparison page shows where Dropbox, Google Drive, Evernote, Notion and Paperless-ngx stand in relation to it, including the places where they're better.

Frequently asked questions

Can you, as the operator, see my documents?

No. There is no function that lets me read other people's documents, and I deliberately never built one. My own documents live in the same system. Technically every hosted service has admin access, what matters is that the business model is the subscription and not your data.

Can I file e-invoices without them losing their status?

Yes, if the tool stores the file unchanged. I tested this with a ZUGFeRD invoice: forwarded, downloaded, SHA-256 identical, the XML was untouched. You can repeat the test with any invoice of your own.

What happens to my documents if I cancel?

You export everything, documents and metadata, at any time with one click. After account deletion, all data including the files is permanently deleted within 30 days.

Can my partner or my tax advisor work in it too?

Yes, via shared spaces with their own roles. Your partner sees the family space, the tax advisor sees the tax space, each only what concerns them. Nobody needs your password.

Do I have to migrate my entire old backlog?

No. Start with your inbox and forward everything new for a week. Whether the backlog follows is your decision afterwards, and if so, start with three test documents.