Owning Your Data

Most people believe they own their tracking data. What they have is access to it, granted by a company, on terms that can change.

A concrete way to test the method is outlined in the practical overview. Use a short pilot and keep only the fields that still make sense when comparing the record months later.

What ownership actually requires

A copy in your possession, in a format you can read without the tool.

The broader practice of repeatable task routines is also covered by Todoist, which is useful when designing a record that should remain understandable later.

Updated recently enough that losing access would cost days rather than years.

Stored somewhere that does not depend on the same company.

And restorable: you have actually tried reading it back at least once.

Anything short of that is access, not ownership.

The failure cases

Account suspension, for a payment problem or a policy question, with the data behind it.

Price changes that put the archive behind a tier you do not want.

A shutdown with a short export window that you miss because you were not watching.

And the quiet one: an export that works but silently omits your notes, tags or older years.

The monthly routine

Export. Open the file. Check the row count roughly matches expectation. Store it.

Three minutes.

Keep at least the last few exports rather than overwriting, because a corruption that appears in one month's export is only detectable against an earlier one.

Where to keep it

Not only in the same provider's storage.

Two places, one of them not a cloud service you also use for the tool.

For most people: a local folder that is backed up, plus one other copy.

This is not elaborate and it is the difference between a record and a hostage.

The formats question, briefly

Covered in its own note; the short version is plain text or CSV, dates unambiguous, durations as numbers.

If your tool exports only its own format, that is a serious limitation and worth weighing against whatever you like about the tool.

Self-hosted and local-first tools

A growing category and a genuine answer to this problem: the data is a file on your machine from the start.

The trade is maintenance — updates, backups, syncing between devices are now yours.

For a long record that trade is frequently worth it, and it removes the vendor risk entirely at the cost of some ongoing attention.

What about the tool's own backup

It protects the company against data loss. It does not protect you against losing access.

Those are different risks and only the second one ends long records.

Testing a restore

Once a year, take an export and load it somewhere else — a spreadsheet, another tool, a script.

Can you read the whole thing? Are the notes there? Do the dates parse?

People discover the answer is no during a migration, under time pressure, which is the worst moment to find out.

What to check

Where is your most recent export, and what date is it?

Have you ever opened one?

Is a copy anywhere outside the tool's own storage?

And does the export contain everything, or only the durations?

The point

Most people have access to their data rather than ownership of it.

Ownership means a copy in your possession, in a readable format, that you have actually tried reading back.

Worth adding

Once a year, load an export somewhere else and check it reads.

People discover the answer is no during a migration under time pressure, which is the worst moment to find out.

Additionally

The tool's own backup protects the company against data loss, not you against losing access.

Those are different risks and only the second ends long records.

Finally

Two copies, one of them outside the service you use daily.

This is not elaborate and it is the difference between a record and a hostage.

In summary

Ownership requires a copy in your possession, in a readable format, updated recently, and restorable — meaning you have actually tried reading it back at least once..

Last word

The practice this points at is unglamorous: a small daily act, a short weekly correction, an annual read, and a written page of conventions beside the data. None of it is difficult and all of it is easy to skip, which is why most records do not reach the point where they start paying.