Quantified self, but for what you buy

Every successful tracker shares one property: it removed the logging. Nobody sustained a paper sleep diary, but everybody wearing a band gets a sleep score without doing anything. Step counts, recovery scores, screen time, listening history: all of them work because the measurement is passive and the feedback is a little uncomfortable.

That is also the boundary of the field. The behaviors that got tracked are the ones where a sensor already existed. Body, phone, music. Money and consumption were left out, not because they matter less, but because manually logging purchases is exactly the kind of chore that nobody keeps up for more than a week. Expense apps have been asking people to categorize transactions for fifteen years, and the retention numbers speak for themselves.

Music is taste. Steps are effort. Purchases are life: where you spend your evenings, what you eat, what you wear, what you rely on, what you gave up. A year of order confirmations describes a person more precisely than a year of listening history, because money is a stronger commitment than a play.

It is also the signal you have never been shown. Your bank flattens everything into "Shopping". A retailer shows you your history with them and nowhere else. Nobody has handed you the cross-brand picture, which is why the first look tends to be genuinely surprising rather than merely informative.

Every purchase, on every site, triggers an email: a confirmation, a shipping notice, a receipt, a booking reference. That is a passive, dated, cross-brand event stream, generated by the sellers themselves, with no logging required from you. It is structurally the same thing a wrist sensor is for sleep.

And it can be read without reading anything private. Sender, subject, date. Enough to say "a purchase happened, at this brand, on this day". Not enough to know what was in the box. The pattern survives without the details, which is the design constraint that makes this kind of tracking acceptable at all.

A budgeting app is a discipline product. It starts from a target, tells you where you failed against it, and depends on you maintaining categories. Most people abandon it, then feel bad, which is a poor foundation for a daily habit.

A tracker is a mirror. It makes no plan for you and issues no verdict. It shows you the shape of what you already did and lets the pattern do the arguing. That is why people keep opening Whoop or Strava for years and stop opening budget apps within a month, and it is the model consumption should have had from the start.

Yearly recaps are the hook, not the product. Purchases keep arriving all year, so the picture can keep updating instead of freezing in December: a new brand entering your top five, a category creeping up, a subscription that renewed while you were not using it.

There is a second layer too. Once you can see how a brand treated you over three years rather than one transaction, a single tap after each purchase becomes meaningful. Enough of those taps, across enough people, produce a live picture of how brands behave right now, instead of a star average built in 2019.