Spending that is seen gets taken seriously
While my car charges I often walk over to the snack shop next door and pick up something to eat. It never felt like much, until the months added up. What follows from that: why a month-end total cannot tell you where money went, why spending that looks reasonable one entry at a time only gives itself away in aggregate, and why the small stuff ended up with an index of its own.
While my car charges, I often walk over to the snack shop next door and pick up something to eat. It is never much, and logging it never feels like anything. Over the months, it turned out to be quite a lot.
That made me think again about what bookkeeping is actually for. My bank statement and my payment apps have always handed me the month-end total, and that total cannot answer my question. I want to know at which unguarded moment the money left. It will only tell me how much left in a month.
Building PFEx, one sentence slowly became the test for every feature: spending that is seen gets taken seriously. A sum that goes out and is never seen was, psychologically, never spent. Next month it happens again.
A total cannot see the single entry
A month-end total is where aggregation ends, not where seeing begins. It compresses thirty small purchases and one large one into a single figure, and staring at that figure tells you only that the month ran over. It never tells you which days, or which things, pushed it there. What aggregation erases is attribution, and without attribution there is nothing to take seriously.
That is not to say totals are useless. A total can only audit; it cannot locate. A bookkeeping tool that stops at the total is doing accounting, not management. Accounting asks whether the books are right. Management asks whether anything will be done differently next time.
So I refuse to simplify the ledger layer. In PFEx every entry is grouped in time order and carries its own category and account, precisely so the total can always be taken apart again, down to one day and one kind of thing. Reading the total means the books are closed. Being able to take it back apart is what seeing means.
A single entry cannot see the aggregate either
Laying the entries out one by one does not solve it from the other side. Those snack-shop purchases all look perfectly reasonable in a list. Read them row by row and the only verdict available is that there is nothing wrong here. Reasonable one at a time, out of control in aggregate: that is what makes this kind of spending so awkward. Every entry survives questioning. Only the pile gives it away.
This is also why I stopped believing that recording everything was enough. A complete record means the facts are in, and facts being in is not a judgment being made. One entry is a fact; gathered, they become a judgment. What sits between them is an aggregation layer, and if that layer only ever produces a monthly total, we are back in the dead end of the first section.
The way out is to choose the grain yourself: read it by day, by category, accumulated over a period. The same money answers differently at each grain. That is when seeing becomes three-dimensional.
The small stuff needs a band of its own
Once those two are clear, a third problem surfaces: inside total spending, the small stuff is permanently hidden behind the large.
Rent, fuel and flights come out of decisions. The moment you decide, you have already seen the money. Snacks, coffee, whatever you pick up on the way come out of habit. The amount is too small to warrant a decision of its own, so even the moment of deciding to spend it never exists. Put both kinds in one table and the second is only ever seen in passing, carried along by the first. Being seen in passing is barely different from not being seen at all.
So I gather everything below one band into its own index, the odd-lot index. The threshold is rounded to each currency's denominations: 100 CNY, 15 USD, 15 EUR, 2000 JPY, 20000 KRW, with no per-trade conversion. What matters is not an exact exchange rate but that this band is worth about the same in that currency too. It brackets the price range you spend without thinking, not a precise figure.
What the index reads out is three things: how much the odd lots add up to this period, how many fills there were, and what share of total spending they took. Its value is not in talking you out of one snack. It is turning "it adds up over the months" into a number you can read on the spot. A single entry is not evidence; gathered they are an indicator. That sentence is written into the app's own help text.
Taking something seriously is not the same as giving it up from today. Eating something while the car charges is fine. What I wanted was to admit that this money is moving all the time, and moving more than it feels like.
Why keep books at all
Asked now why I keep books, my answer is not what it was at the start. It is not so the total reconciles at month end. It is so that spending which happens without attention gets placed in front of me at least once.
A total is for auditing. A single entry is for locating. The aggregate of the small stuff is for spotting a habit. Spending that is seen gets taken seriously, and only what is taken seriously ever gets a chance to change.