Sooner or later a receipt will have more than one plausible home. Two coffee payments on the same morning, or two grocery trips at similar totals, can leave a receipt hovering between candidates. Spendix handles this by presenting a ranked list rather than forcing a guess, and the skill lies in reading it well.
What the ranking is telling you
The candidates sheet lists transactions that could belong to the receipt, with the likeliest at the top. The ordering is a suggestion built from how closely the transaction resembles the receipt. It is useful, but it is a starting point. A ranking can place the right answer second, particularly when two payments are nearly identical.
The sheet also has an action to attach to the best candidate in one go. That is the right choice when the top entry clearly matches and the rest are visibly different. If two entries look alike, slow down.
A three-point check
When faced with similar candidates, compare them on three points, in this order.
- **Amount.** Does the receipt total equal the transaction exactly? A mismatch of a few units can point to rounding or a tip; a large gap points to a different purchase.
- **Date.** Receipts carry the date of purchase. Card and UPI entries usually share it, though a card transaction may post a day or two later.
- **Merchant.** Bank descriptions are often abbreviated. Use the merchant as a tiebreaker rather than the main test.
Most ties break on the second point. Two payments of the same size on different days rarely survive a glance at the receipt date.
When the candidates are truly identical
Occasionally two transactions share amount, date and merchant. This happens with repeat purchases, such as two identical metro top-ups or two tea stall payments. In that case the receipt cannot tell them apart and it hardly matters which one carries the attachment, since both are the same kind of spend.
If it does matter, say for a reimbursement claim, check the time of day on the receipt against any time information you have, or look at the order of the entries. Then attach to one and mention the other in a note so the pairing is deliberate.
What if neither is right?
The candidate list is built from transactions that already exist. If none fits, the sheet may say that no matching transactions were found for the amount yet. That is a signal, not a failure. Either the payment has not appeared, or it never will because it was cash.
In the first case, wait. In the second, create a transaction from the receipt instead of attaching it to something unrelated. Forcing a match to the nearest candidate quietly corrupts two records at once: the real transaction gets a stray attachment, and the true purchase never shows.
| Signal on the receipt | What it suggests |
|---|---|
| Exact amount, same day | Strong match; attach |
| Amount a little higher on the transaction | Tip, fee or rounding; check, then attach |
| Same amount, different day | Likely a separate purchase; look further |
| Several equal candidates | Choose one and add a short note |
Using notes to settle doubt later
Whenever a pairing needed real thought, a one-line note on the transaction saves the same thought later. Something like matched to the second of two equal payments is enough. Months on, when a charge is queried or a claim is prepared, the note shows the choice was deliberate and what it was based on, rather than leaving you to rebuild the reasoning from scratch.
Making future choices easier
Short gaps between scanning and matching mean fewer lookalikes in the list. Weekly sweeps keep each candidate list small, and a clear photo of the total means the amount you compare against is reliable. With those two habits, the ranked list becomes a confirmation tool more often than a puzzle.
Useful ideas start with a clearer picture of your own spending.
Explore your reports


