Own Your Data Online: A Practical Guide

To own your data online means holding your own copy of it, in a file format you can open without the app that made it, on storage you control. That is the whole definition. A login is not ownership. A privacy setting is not ownership. If the only copy of your photos, notes and messages lives inside somebody else's account, you are renting your own memory, and the lease can end without notice.

This guide is the practical version, aimed at a normal person with a phone, a laptop and no interest in becoming a system administrator. It covers what to export, which formats survive a decade, how to keep a backup you can actually restore, and how to tell whether a service will let you leave. It also covers what data ownership does not fix, because that part gets left out of most advice.

What owning your data actually means

Ownership has three parts and most people only have one. You need a copy you hold, a format you can open without the original app, and somewhere to put it that is not another account you could lose. Miss any one and the other two stop mattering.

Laws help less than people expect. The GDPR in Europe and the CCPA in California give you the right to request a copy of your data and, in some cases, to have it moved. That is a right to ask. It is not a habit of keeping. If you have never used it, you do not have your data. You have permission to request it later from a company that still has to exist.

Here is the honest test: if the service vanished tonight with no warning and no email, what would you still have tomorrow morning? Whatever survives that question is the part you own. Everything else is borrowed. For the broader idea behind this, see what data sovereignty means.

Start with an inventory, not a migration

People fail at this by trying to leave everything at once. Do not. Spend one evening writing down where your things actually live and how each one gets out. The list is usually shorter than you fear.

Then rank each row by two questions: how much would it hurt to lose, and how hard is it to get out? Start where the answers are a lot and not very. That is almost always photos, contacts and documents, and getting those three home is most of the benefit for a fraction of the effort.

Run the exports before you need them

Every large service has an export tool, buried but real. Google Takeout, Apple's Data and Privacy page, Meta's Download Your Information, Mastodon's account archive. Each produces a zip you download once. None of them run themselves.

Two things surprise people. First, exports are slow. A large account can take hours or days to build, and the download link expires, so a request you ignore for a week is a request you have to make again. Second, archives arrive split into chunks, often 2GB or 50GB files, and if one chunk fails to download the whole set will not extract cleanly.

So the rule is simple: open the archive and look inside before you trust it. Find five photos you remember. Open the MBOX file in Thunderbird. Confirm your notes came out as readable text and not an opaque blob. Then set a calendar reminder to do it again in six months. Google Takeout can schedule repeat exports every two months for a year, which is the closest thing to automatic that these tools offer.

Formats that outlive the app that made them

The quiet form of lock-in is not the account. It is the file format. A folder full of files that only one discontinued program can open is not a personal archive. It is a hostage situation with extra steps.

Formats that have aged well, because many unrelated programs can read them:

• Text and notes: plain text and Markdown (.txt, .md)
• Records and tables: CSV, or SQLite for anything structured
• Documents: PDF for finished things, ODT or DOCX for editable ones
• Photos and video: JPEG, PNG, and MP4 using H.264
• Mail, contacts, calendar: MBOX or EML, vCard, iCalendar

Convert away from anything with DRM, proprietary notebook formats, and closed backup blobs only their own app can restore. Two honest exceptions: Apple's HEIC photos and camera RAW files are not open in the friendly sense, but both are widely supported and worth keeping for quality. The working rule is that if two programs from two unrelated developers can open a file, it will probably still open in ten years.

A backup you can actually restore

Exports plus one hard drive is not a backup plan. It is a single point of failure wearing a costume. The old rule still holds: three copies of anything you care about, on two kinds of storage, with one copy somewhere else.

Software that does this well, all free: restic and borg for encrypted, deduplicated backups, rsync for straight copies, Time Machine if you are on a Mac, Duplicati if you want buttons instead of a terminal. Syncthing keeps folders matched across your own machines, which is useful, but sync is not backup. Delete a file by mistake and sync helpfully deletes it everywhere.

Encrypt before anything leaves the house, so the offsite copy can sit on cheap rented storage without you having to trust the company holding it. Then do the part everyone skips: restore something. Pull one folder back out and open it. A backup you have never restored is a hypothesis. Testing it once a year turns it into a fact.

Where self-hosting fits, and where it does not

Running the software yourself is the strongest form of data ownership, because the files sit on a disk in your house and nobody can revoke your access to them. Nextcloud handles files, contacts and calendars. Immich is a genuinely good replacement for Google Photos. Vaultwarden holds passwords, Jellyfin holds media, Home Assistant keeps your home automation local instead of routed through a vendor's cloud.

The tradeoff is real and worth saying plainly: self-hosting makes you the administrator. Updates, failing disks, expired certificates and the 2am outage are now yours. Email is the hardest case and most people should not self-host it. A paid provider on your own domain name gets you nearly all of the portability with none of the deliverability pain.

Data can also stay local at the network level, not just on the disk. Community networks keep traffic between neighbours on hardware the neighbours own, which is the same idea one layer down. See how mesh networks work, and if you want a gentler on-ramp than a home server, de-Googling step by step is the practical order to do things in.

Portability: test the exit before you need it

The best moment to check whether you can leave a service is the day you sign up, while you have nothing invested and no urgency. Ask for the export on day one. If it is missing, partial, or produces a format nothing else can read, you have learned something important for free.

Signs a service will let you go: a documented export in an open format, support for standard protocols such as IMAP, CalDAV, CardDAV, RSS or ActivityPub, and data stored as ordinary files in ordinary folders rather than a proprietary database. A self-hostable version of the same software is the strongest signal of all.

The highest-leverage single move is owning your domain name. An address at a domain you control can follow you between providers forever. An address at somebody else's domain cannot move at all. Be honest about the limits, though: moving a Mastodon account brings your followers but not your old posts. Some things do not survive the trip, and knowing which ones in advance beats finding out mid-migration.

What owning your data does not fix

This is the part most guides leave out. Holding your own copy is worth doing, and it does not do everything.

• It does not delete their copy. Exporting is copying, not clawing back. Deletion is a separate request, and backups or legal holds can keep data alive after it.
• It does not cover data you never held. Advertising profiles, device telemetry, credit records and the messages your friends kept were never in your account to export.
• It moves risk onto you. You can lose the drive, leak the folder, or forget the passphrase. Encryption without a key you can find again is a slow way to delete things.
• It is not anonymity. A server in your house still has an address, and traffic to it is still visible.

There is a time cost too, and it is ongoing rather than one-off. For plenty of people the right answer is a reputable paid provider plus a disciplined export habit. That is a legitimate way to own your data online, and it beats an ambitious home server that stopped getting security updates eight months ago.

A realistic first month

Small and repeated beats one heroic weekend. Four short sessions get most people to a genuinely defensible position.

1. Week one: write the inventory, then request every export. Requesting is quick and waiting is the slow part, so start it early.
2. Week two: the archives arrive. Open each one and check it. Buy one external drive, set up restic or Time Machine, and copy everything onto it.
3. Week three: convert the fragile files to open formats. Register a domain name and move your email address onto it, even if you stay with your current provider.
4. Week four: add the offsite copy, encrypted. Then restore one folder from it to prove the whole chain works end to end.

After that it is one recurring reminder every six months: re-export, verify, restore one thing at random. Twenty minutes, twice a year. That is the entire maintenance burden, and it is the difference between people who own their data and people who intend to. More guides are in the learn index.

Frequently asked questions

Do I legally own my data online?

Not the way you own a chair. Laws like the GDPR and CCPA give you rights to access, copy, port and delete personal data, but the service keeps its own copy and often keeps derived data too. Treat the law as a tool for getting exports out, not as a substitute for keeping your own copies.

What file formats are safest for a long-term personal archive?

Formats many independent programs can open. Plain text and Markdown for notes, CSV or SQLite for records, PDF for finished documents, JPEG and PNG for images, MP4 with H.264 for video, MBOX for mail, vCard for contacts, iCalendar for calendars. If only one company's app can open a file, convert it now rather than later.

Is cloud storage the same as owning my data?

No. Cloud storage keeps your files for you inside an account that can be suspended, repriced or shut down. It becomes part of owning your data only when you also hold a local copy you could restore from tomorrow without that provider. Cloud as a second copy is sensible. Cloud as the only copy is not.

How often should I export my data from Google or Apple?

Every six months covers most people, and Google Takeout can schedule repeat exports every two months if you would rather not remember. Export sooner when something changes: before cancelling a subscription, before switching phones, or after a service announces a shutdown. Always open the archive and check it before deleting anything from the original account.

Do I have to self-host to own my data online?

No. Self-hosting with tools like Nextcloud or Immich is the strongest version, but it makes you responsible for updates, disks and outages. A reputable paid provider on your own domain name, combined with regular exports and a tested backup, gets most of the benefit for far less work. Pick the version you will actually maintain.

What is the difference between file sync and a backup?

Sync keeps copies identical across devices, so a mistake spreads instantly. Delete a folder, or let ransomware encrypt it, and sync repeats that everywhere within seconds. A backup keeps older versions you can go back to. Restic, borg and Time Machine keep that history. Syncthing, Dropbox and iCloud Drive mostly do not.