Expense tracking has one failure mode and everything else is a detail. If logging a transaction takes more than about ten seconds, you will stop doing it, and it usually takes about three weeks.
Everything below is in service of that one constraint.
Log it when it happens
The alternative is reconstructing later, and reconstruction does not work. By the time you sit down on Sunday you have forgotten a third of the week, and the third you forget is not random. It is the small, frequent, unremarkable spending, which is precisely the category most likely to be the thing you needed to see.
Logging in the moment takes seconds and produces a record you can trust. Logging on Sunday takes half an hour and produces a record that is confidently wrong.
If you are somewhere without signal, this matters more than it sounds: an app that queues the entry offline and syncs later means you can log on a plane or in a basement and not lose the habit to circumstance.
Do not convert foreign spending by hand
If you spent 45 euros, record 45 euros.
Converting in your head at the till, or worse, looking up a rate later and writing down an approximation, breaks your own history in two ways. You have frozen a rate you can never verify, and you have made your categories unreadable, because a rate move now looks identical to a change in your behaviour. Two identical grocery shops a month apart will appear to be different amounts.
Record what you paid, in what you paid it, and let the tool convert when it produces reports. In Spendly transactions inherit the currency of the account they belong to, so a euro card records euros without you deciding anything.
Cash in a foreign currency deserves one extra note: record it against a cash account in that currency rather than against your main one. A wallet of leftover forints is a real balance, and treating it as an untracked hole is how a trip’s spending ends up unaccounted for.
Categories that match your life, not a template
Generic categories are where tracking goes to die. “Miscellaneous” and “Other” absorb everything you did not want to think about, which is exactly the spending worth examining.
Use categories that describe how you actually spend. If your problem is coffee, “coffee” is a category. If you have never once wondered how much you spend on household supplies, do not create the category.
Keep the total small. Every category is a decision at the moment of logging, and decisions are what push you past the ten-second budget. Eight to twelve categories is enough for most people, and you can always split one later when you have a specific question it cannot answer.
The first month is measurement, not improvement
Track everything for a full month and change nothing.
This is harder than it sounds, because the moment you start measuring, the instinct is to start behaving better. Resist it. A month where you were unusually careful tells you what a careful month costs, which is not the number you need. You need to know what a normal month costs, because that is what your budget has to survive.
At the end of the month, look for four things:
- A category you underestimated. There is almost always one, and it is rarely the one you expected.
- Recurring charges you had stopped noticing. These compound quietly and are the easiest thing on the list to fix.
- The gap between what you thought and what happened. The size of that gap tells you how much you were previously deciding by feel.
- Any pattern in when you overspend rather than what you overspend on. Time of day, day of week, mood, being in a hurry. The pattern is more actionable than the category.
When you miss a fortnight
You will miss a fortnight. It is the normal way this goes, and how you handle it decides whether you are still tracking in six months.
Do not reconstruct. Mark the gap as missing data, start again today, and accept that one month of your history is unreliable. Trying to rebuild three weeks from memory takes an evening, produces figures you cannot trust, and is the point at which most people quit, not because they failed to track but because catching up felt like a punishment.
One incomplete month in a year of good data costs you almost nothing. Abandoning the habit costs you the year.
What tracking is actually for
The point is not the record. It is that after two or three months you stop guessing.
You know roughly what your life costs, you notice a new recurring charge because it is unfamiliar rather than because it broke something, and you can answer questions about your own money with a number instead of an impression. That is what makes a budget buildable, because a budget is just a set of limits and limits set from guesswork are guesses.
Once you have the data, reports turn it into the answer to a question you can only ask once the logging is honest: is this getting better or worse?
Common questions
What is the best way to track daily expenses?
Log each transaction at the moment it happens rather than reconstructing later. If logging takes more than about ten seconds you will stop doing it within a month, so speed matters more than detail. Reconstructing a week from memory produces a record that is wrong in exactly the categories you most need to be right about.
How should I record a purchase made in a foreign currency?
Record the amount you actually paid, in the currency you paid it. Do not convert by hand. Converting at entry freezes a rate you cannot check afterwards, and it makes month-to-month comparisons meaningless because a change in the exchange rate becomes indistinguishable from a change in your spending.
How long should I track before changing my spending?
One full month, with no changes and no judgment. The point of the first month is to see what your life actually costs, and cutting back while you measure gives you a picture of a month you do not normally have. Set limits from the second month onwards, using the real figures.
What do I do if I stop logging for a while?
Do not try to reconstruct the gap. Mark it as missing, start again today, and accept one bad month in your history. Attempting to rebuild three weeks from memory produces bad data and takes long enough that most people give up during the attempt rather than after it.
Related reading
- How to create a personal budget that actually works for turning a month of data into limits
- How to find and cancel forgotten subscriptions for the recurring charges tracking surfaces
- Managing personal finances in multiple currencies if you are logging in more than one currency