Behavioral analysis goldmine

Want to know intimate details about someone's life? Just peek into their financial tracker data.

Stranger experiment

I decided to open many years of my own financial tracking data and look at it as if I were a stranger skimming through someone else's records. The results were very surprising. I could answer many questions about my own life, which I'm obviously not going to share here. You can try that yourself on your own data and see what you can learn from that.

Knowing you better than yourself

The annotated financial transactions, such as "Eating out with Alice" or "Trip to Dublin for Bob's wedding", that are present in financial trackers are much more descriptive than the transactions your bank can see. The actual processing of transactions is siloed too, whereas finance tracking shows the global state, including cash and cryptocurrency transactions. Moreover, the temporal aspect along with user-defined tags can give out the precise history.

After all, the transactions represent someone's choices under certain constraints - financial resources. The behavior is almost ingrained into the data. You just need to point the right tools at it. Examples of what can be inferred include: habits, lifestyle patterns, health information, relationship status, employment details, political or social views, times of struggles, and even future plans.

It's quite interesting, yet scary at the same time the patterns you can extract from these records alone. Now, what if that stranger is a neural network? The pattern recognition on thousands of transactions can be very precise, plus it may as well predict certain patterns based on history. The neural network will not just see isolated events - it would identify a sequence. Someone might think they are private about their weight loss goal, but the algorithm can see something like: groceries up 40%, fast food down 80%, gym membership, clustered doctor visits in a 4-week period, followed by a nutritionist payment.

Similar can be applied to many life changes. Career transition can show up as: professional development courses, books, networking events expenses, followed by moving expenses. Relationship troubles could include: couples therapy payments, increased alcohol spending, separate restaurant charges instead of shared ones, then divorce attorney fees.

I bet it can also craft a biography too.

The real risks

Financial information can be used for advertising. You may think that "Who cares, it's just ads?". Sure, but what if the same data is analyzed by insurance companies? This may as well affect your life directly. The insurance company can estimate whether the individual has risky behavior, and therefore deny coverage, or make it more expensive.

Of course, the data has to be shared in the first place, and it's not that easy. However, acquisitions do happen, and I believe most of the time they happen for the data, not the service itself. Take WhatsApp / Facebook as an example - what once was a privacy-friendly messenger, became integrated into Facebook's ad network. Acquisitions happen frequently enough to notice that dynamic.

Another risk factor is fraud. Insider threats are real - I believe there is a dark market for precious data where financial information can be sold for future targeted attempts to extract money from an individual. Data breaches happen too, and services storing financial information become attractive to the attackers.

A more sensitive subject is changes in the political landscape. What once was perfectly legal, may become illegal later on, and discrimination can easily happen based on past behavior patterns. Someone might purchase books on political theory, record the transactions as "Amazon - Specific Book Title That Could Get Me In Trouble" and then face scrutiny if they move to a country where those ideas are considered extremist.

I'm not trying to spread panic here or anything, it's just how things are I believe. Financial information is sensitive, and should be protected on all layers, beyond just regulation. Rules can be broken after all, and it's hard to trace that down. Plus digital records are easily replicated, one can reasonably assume that once digital data is somewhere in plain text, it can be anywhere else really.

Convenience, right?

I suspect most people use services on the Internet for convenience. Those who understand the privacy implications of sharing financial data resort to plain text accounting, spreadsheets, or self-hosted solutions, sacrificing the convenience of multi-device data synchronization and reliability of cloud storage solutions.

I believe that financial trackers require beyond-bank-level-security for the reasons described. Privacy and data reliability first. Not just based on policies, but with technical safeguards in place in addition to them. I think that's what is missing on the market, so I decided to build that kind of thing. The app that I would trust myself, adjusted to my standards for both data leak and loss of data protection.