Should you go back and fill in the expenses you forgot to record
A fortnight of missing entries is uncomfortable to look at. Reconstructed data is not neutral, though — it is wrong in a predictable direction, and knowing which direction tells you when backfilling is worth doing.
You went away for a long weekend, came back to a busy fortnight, and the record has a hole in it. The hole is visible every time you open the book and it is faintly annoying, in the way an unfinished thing is annoying. The obvious response is to sit down with a card statement and reconstruct it.
Sometimes that is worth doing. Often it produces something worse than the gap, because a reconstructed fortnight is not a slightly rougher version of a recorded fortnight. It is a different shape entirely, and the difference runs in one direction.
What reconstruction reliably captures, and what it reliably loses
Anything with a paper trail comes back easily. Direct debits, standing orders, the annual car insurance, the nursery invoice, a hotel booking, a large supermarket shop, a train ticket bought online. These are large, they are individually memorable, and they appear on a statement with a name attached.
Everything else does not come back. The two coffees, the bakery on the way to the station, the top-up shop for milk and bread, the parking, the cash handed over at a market stall, the small contactless taps that make up an ordinary week. On a statement these appear as a merchant name and an amount, which is often enough to know that something happened and not enough to know what it was. From memory they do not appear at all.
So the reconstructed fortnight looks like this: fixed costs complete, large purchases complete, groceries mostly there, and the small frequent categories somewhere between half and mostly absent. Which means the reconstruction does not merely understate the total. It changes the proportions. Food looks like a bigger share of your spending than it is, because the coffees and the snacks that would sit beside it have vanished. The categories people most want visibility into are exactly the ones reconstruction is worst at.
That is the argument against backfilling by default, and it is not an argument about honesty or discipline. It is an argument about what the resulting figures can be used for.
When backfilling is genuinely worth it
There are cases where reconstruction is not just acceptable but clearly better than the gap.
- Large one-off costs. A boiler repair, a flight, a new washing machine, a school trip. These dominate a month, they are perfectly recoverable, and leaving them out distorts the year far more than adding them from a statement does.
- Annual and irregular obligations. Insurance renewals, road tax, professional memberships, the annual subscription that renewed in the middle of the gap. These matter mainly for planning, and planning needs the date and the amount, which are both on the statement.
- A trip. Holiday spending is unusual, concentrated and easy to reconstruct from a card statement plus a memory of where you were. It also tends to be recorded separately anyway, so contaminating it with estimates does less damage to the household baseline.
- A gap in a period you are about to make a decision about. If you are working out what childcare and term-time lessons cost you last year, filling the recoverable parts of the gap serves the specific question you are asking.
The common thread is that all four are large, dated, single-source and recoverable. Backfilling them adds real information at no cost to accuracy.
When backfilling makes the book less true
The opposite case is reconstructing ordinary daily life. Two weeks of groceries, transport, coffees and small household purchases, assembled from a statement and a vague memory of the fortnight.
What you get is a set of figures that look like data and behave like a guess. Six months later, nobody remembers that the second half of March was estimated. Somebody looks at the year and concludes that the household spends less on eating out in spring, and the conclusion is entirely an artefact of the reconstruction. An estimate that is not labelled becomes a fact with the passage of time, which is the real cost.
There is a second and smaller reason to leave ordinary gaps alone: the reconstruction session is long, joyless and typically happens in the evening after a busy fortnight, which is precisely the situation in which people stop keeping records altogether. The gap is not the thing that ends a household ledger. The two-hour repair job is.
If you do backfill, mark it
The whole problem with estimated data is that it is indistinguishable from recorded data after a few weeks. That is fixable cheaply.
- Decide the boundary of the gap explicitly — a start date and an end date — before entering anything.
- Enter only the recoverable items: statement lines, invoices, anything with a date and an amount you can read rather than recall.
- Put a note on every reconstructed entry saying it was reconstructed. A single word is enough, and it is the difference between an estimate and a false memory.
- Do not invent the small stuff. If you know you bought coffees and cannot recall how many, one honest line described as an estimate for the period is better than nine invented lines that look precise.
- When you next read the year, look for the marker before drawing any conclusion about that period.
The habit of marking estimates is worth more than the backfill itself. It generalises: rounded amounts, split bills you never settled exactly, the cash you handed to a relative, the market stall with no receipt. A household record with a handful of honestly labelled estimates is much more usable than one that pretends to a precision it never had.
The gap is not a failure
It is worth saying directly, because the impulse to backfill is usually less about the data than about the way the hole looks. A missing fortnight in a household record means there was a fortnight in which recording was not the most important thing, which is normal and will happen several times a year — around Christmas, around a move, around illness, around anything involving a small child.
The property that actually determines whether a record survives is whether you can return to it after two weeks away and simply carry on from today. Anything that makes returning feel like the start of a repair job works against that. If you open the book after a gap and the only available next action is a reconstruction session, the honest choice is to skip the reconstruction and record today’s shopping instead.
Frequently asked questions
Should I fill in expenses I forgot to record?
Fill in the large, dated, recoverable ones — a repair, a flight, an annual insurance renewal, a school trip — because leaving those out distorts the year more than a statement-based entry does. Reconstructing ordinary daily spending is usually not worth it, because memory and statements lose the small frequent purchases and keep the large ones, which changes the proportions rather than just the total.
Is it accurate to reconstruct spending from a bank statement?
It is accurate for anything that appears as a distinct line with a date, and poor for anything cash-like or split. A statement gives you a merchant and an amount, not what was bought, so a single supermarket line may contain groceries, a birthday card and cleaning products. Treat it as a source for large items rather than a full reconstruction.
How should I mark an estimated entry?
Put a note on the entry saying it was reconstructed or estimated, and record the start and end date of the gap somewhere you will see it later. Unlabelled estimates become facts after a few months, and a conclusion drawn from them can be entirely an artefact of the reconstruction.
What if I have missed a whole month?
Enter the fixed obligations and any large one-off costs from statements and invoices, mark them as reconstructed, and leave the daily spending out. Then resume from today. A month that is honestly partial is far more usable than one assembled from guesses, and much easier to face than a long repair session.
Does a gap ruin the usefulness of a year of records?
No. Large, regular costs are unaffected because they are recorded once from a single source. What a gap affects is comparisons involving small frequent categories over that period, so the practical answer is to note where the gaps are and avoid drawing conclusions about coffee or takeaways from months that contain one.
Try it for one month
Fambook gives a household one shared ledger: anyone can add an entry in seconds, every entry says who spent it, and the month adds up in one place instead of two. Records with no signal and syncs afterwards. Recording, categories, budgets, statistics, CSV import and export, sync and sharing for two people are free — the subscription only buys you less typing.