Most couples can tell you what they earn to the dollar. Ask what they spent last month on groceries, takeout and everything in between, and the numbers get vague fast. That gap is where arguments start, because both people are sure they know.
A bank statement settles it. One month of actual transactions, sorted into piles, compared with what you both assumed. It takes about 40 minutes and replaces a surprising amount of guesswork.
Why the statement beats the memory
Recurring bills are easy to remember because they arrive on a schedule. What people forget is everything else: the fourth grocery run, the second delivery order, the annual charge that renews quietly in the background. Shared spending is usually a long tail of small repeated decisions, not one big one.
So the review has three questions, in order. What did we spend by category? Which of these charges did neither of us consciously choose? Did the split we agreed on actually happen?
A worked month, line by line
Suppose one partner takes home $5,800 a month and the other takes home $3,700. Combined that is $9,500, so a proportional split puts the shares at about 61.1% and 38.9%. Here is one month of shared spending, categorized from a single joint account and one shared card. The figures are illustrative, not a national average.
| Category | Spent | Share of shared total |
|---|---|---|
| Rent | $2,350 | 57.9% |
| Groceries | $842 | 20.7% |
| Shared dining and takeout | $212 | 5.2% |
| Utilities | $186 | 4.6% |
| Transport and fuel | $164 | 4.0% |
| Household supplies | $96 | 2.4% |
| Internet | $75 | 1.8% |
| Streaming and subscriptions | $63 | 1.6% |
| Pharmacy and health | $44 | 1.1% |
| Renter's insurance | $28 | 0.7% |
| Total shared spending | $4,060 | 100% |
Once the category totals exist, two things become visible that memory hides. First, rent is 57.9% of everything shared, which means any conversation about the other 42% is a conversation about a fairly small pile. Cutting takeout in half saves $106, not $400. Second, the three smallest categories together, subscriptions, pharmacy and insurance, came to $135 and almost none of them get discussed in a normal month.
The charges nobody chose
Go back through the statement and mark every line that repeats something from a previous month. In this example, three subscriptions renewed inside the shared account: cloud storage at $17.99, a fitness app at $11.99, and a news site at $9.99. That is $39.97 of the month's $63 of subscriptions, or $479.64 a year, for things neither partner could remember deciding to keep.
This is the most common finding in a first review, and the least emotional to fix. It is a list of standing orders that outlived their reason, not a fight about who is careless.
Did the split actually happen?
The plan said the higher earner covers 61.1% of $4,060, or $2,479, and the lower earner covers 38.9%, or $1,581. Now look at who paid what. In this month, rent, insurance, transport, subscriptions and pharmacy came out of the higher earner's account: $2,649. Groceries, dining, utilities, internet and supplies went on the lower earner's card: $1,411.
The higher earner paid $170 more than the agreed share and the lower earner paid $170 less. Nothing went wrong and nobody was unfair. The categories landed unevenly, which is what happens when rent is one payment and groceries are nine. The fix is a $170 transfer, not a difficult conversation.
Here is the question worth asking out loud: which number in that table would you have guessed wrong about before reading it? Most people miss groceries by more than they miss rent.
Three lines to ignore before you start
If you would rather not build the category totals by hand, the statement analyzer will produce them from your bank's CSV export in a couple of minutes. Either way, statements contain movements that look like spending but are not. Transfers between your own accounts inflate the total. A refund credits a category it never belonged to, making the month look cheaper than it was. Credit card payments double count if you also count the purchases on that card. Strip all three out, then categorize what remains.
Run the same review on your own month
Sorting transactions by hand is fine for one month and tedious by the third. You can analyze a bank statement in the browser by uploading the CSV your bank exports, which produces the same category totals without the spreadsheet. The analysis runs locally, so nothing leaves the device. Use it to check the shape of the month, then come back to the split and ask who actually paid what.
Two habits make the review stick. Do it on the same day each month so the comparison is like for like, and write down the two or three numbers you expected beforehand, because the value comes from the gap between the guess and the statement. A month you can describe accurately is a month you can plan from.
