splitpocket

What your expense app knows about you

Spending data is some of the most revealing data there is. What to look for in a policy, and what a private expense app looks like.

4 min read

People guard their chat history and hand over their spending without a second thought. It's the wrong way round. A month of expenses is one of the most revealing datasets a person produces: more precise than a diary, and unlike a diary, it's structured, timestamped, and trivially searchable.

What one ordinary month says about you

One ordinary month, read back to you

what each line discloses

Pharmacy, four timesan ongoing condition
₹3,240.00
Same restaurant, Thursdaysa relationship, and roughly when it started
₹8,900.00
Temple donation, monthlyyour religion
₹1,100.00
Rentyour neighbourhood and your income bracket
₹19,200.00
Salary credited on the 1styour employer and your pay
₹86,000.00
None of these are secrets on their own. Together they're a profile no one asked you to consent to.

This is why financial data is regulated more tightly than almost anything else, and why "we only share aggregated, anonymised data with partners" deserves more suspicion than it usually gets. Spending patterns are close to a fingerprint: a handful of dated transactions is generally enough to pick one person out of a large supposedly-anonymous dataset.

Follow the money, not the marketing

Every app has to be paid for by someone. There are broadly three models, and they produce very different products:

  • You pay. A subscription or a one-off price. Your interests and the app's are aligned; it just has to be worth the money.
  • Advertisers pay. Now the app's customer is the advertiser and you're the inventory. Ad SDKs typically ship device identifiers off to networks whose privacy policies you never see, and the incentive is to know as much about you as possible.
  • Data buyers pay. The quiet one. Nobody advertises this on the app-store page; you find it in a clause about "partners" and "research".

You can usually tell which one you're in by reading two things: the privacy policy's sharing section, and the app-store data-safety label. If the label lists financial info collected and linked to your identity for advertising, that's the whole answer.

Six questions worth asking

  1. Does it show ads? If yes, assume some data goes to an ad network.
  2. Who else gets the data? Look for "partners", "affiliates", "service providers" and read what they're allowed to do with it.
  3. Is there third-party analytics? Session recording and behavioural analytics in a finance app means a third party holds a copy of your screens.
  4. Does your data train an AI model? An increasingly common clause, often buried in an "improving our services" sentence.
  5. Can you export everything? An app that can't emit a CSV is an app that's holding your history hostage.
  6. Can you actually delete it? Deleting the app is not deleting the account.

Where SplitPocket stands

We'd rather state it plainly than make you find it in a policy:

  • No ads. Not now, not as a free tier, not as "relevant offers".
  • No selling your data, to anyone, for anything.
  • No third-party analytics or trackers, not in the app, and not on this website. The page you're reading loads no third-party scripts at all.
  • No feeding your ledger to AI systems.
  • Your data leaves with you as CSV, whenever you want it.
  • Access rules live in the database, enforced per-row, not just checked in the app. Group members see the group's expenses; your personal expenses are yours.

The full text is on the privacy page, and it's short enough to read in one sitting, which is itself a design decision.

The test that matters

Imagine your expense history printed out and handed to a stranger who wanted to sell you something. That's the actual risk model for ad-supported finance apps. Not a dramatic breach, just the ordinary daily business of the thing. Pick an app whose business doesn't require it.