Own Your Records: The Case Against Renting Your Financial History
There is a pattern worth noticing in personal finance tooling. Nearly every popular expense app asks for the same thing up front: connect your bank, grant read access, let us ingest the feed.
It works. It also means your complete transaction history lives with a third party, under their terms, for as long as they choose to operate.
That is a trade some people make happily. It is worth knowing it is a trade.
What you actually give up
Three things, and none of them are hypothetical.
Custody. The canonical version of your records sits on infrastructure you do not control. If the service is acquired, changes pricing, or shuts down, your export options are whatever they decide to offer.
Format. Proprietary storage means migration is a project. A vendor that exports to a locked format has effectively locked your history.
Privacy. A bank connection is not a snapshot. It is an ongoing feed of everything you buy, which is considerably more than what an expense record requires.
What the requirement actually is
Strip out the product marketing and the obligation is narrow. You need documents showing merchant, date, itemized description, amount, and payment method, kept legible for as long as the retention window runs.
The IRS puts that window at three years in the standard case, six years where more than 25% of gross income went unreported, and four years for employment tax records.
Nothing in that requires an OAuth grant against your bank account. It requires legible files that still exist in year three.
Why the paper original is not the answer either
Self-custody does not mean keeping the slips. That fails for a different reason.
Store receipts print on thermal paper, which contains no ink. A colorless leuco dye and an acidic developer react under heat from the print head, and the reaction is reversible. Per the thermal paper reference, the dyes return to their original colorless crystalline form when stored in hot or humid conditions, with UV light, oils and friction accelerating it.
So the physical archive degrades on its own schedule regardless of how carefully you file it. Paper is not custody. It is a slowly expiring copy.
A self-custodied workflow
Four steps, no accounts, no integrations:
- Capture at purchase. A photo at the register, while the print is dark.
- Rebuild anything above a threshold into a proper itemized PDF. Set the number to match your situation.
- Store in a plain directory tree. Year, then month. Filenames date-first so they sort and search.
- Back it up wherever you already back things up. The records are just files, which is the entire advantage.
For step two, a browser-based free receipt generator covers it without an install or an account. Enter the details you already hold, download a PDF, file it. Nothing is connected to your bank, nothing syncs, and nothing needs cancelling later.
Why plain files keep winning
Because they outlive the tools. A directory of dated PDFs opens on any operating system, in any decade, without the software that created it still existing.
That is the same reasoning behind every argument for portable, self-held data. The value of a format is not what it does today. It is whether it still works when the vendor is gone.
What you actually gain
The practical difference is worth naming precisely: one fewer party holding a continuous record of what you buy, for a job that never required one.
Your bank and the card networks still hold the transaction history. What changes is that an expense tool does not join that list, and your archive stops depending on a company's roadmap. Receipt Mango does the formatting step and holds nothing, which is roughly the right amount of involvement for a tool whose job takes sixty seconds.
The takeaway
Convenience defaults are worth examining, not rejecting. Sometimes the connected option genuinely is better.
For expense records specifically, it is not. The requirement is legible files kept for a few years, and that is a problem the filesystem solved a long time ago.