Instead of reading sixty rows top to bottom, group the month by merchant and decide once per merchant. One decision then covers every row from that shop, and you turn the repeat ones into rules so they never need your attention again. The same sixty rows that take half an hour line by line can take a few minutes this way.
Row by row versus merchant first
Take a month with 60 transactions. Suppose they come from 14 different merchants: a supermarket seen 9 times, two delivery apps seen 12 times between them, a fuel pump 4 times, and so on. This is hypothetical, but typical shapes look like this.
| Method | Decisions you make | What you carry away |
|---|---|---|
| Row by row | 60, one per row | A corrected month, nothing remembered |
| Merchant first | 14, one per merchant | A corrected month, plus rules for repeats |
The difference is not only speed. When you read in date order, you see the same supermarket in different moods: once under Groceries, once under Shopping, once left blank. You fix each separately and the inconsistency creeps back next month. Grouped together, the inconsistency is obvious and you fix it once.
How to group the month
In the Report builder in SpendixAI you can filter and group transactions, so you can group the month by merchant and read the totals beside each name. If your tracker, or the view you are using, cannot group, sorting by description does most of the same job. Start with the merchants that have the biggest totals, because a wrong category on those distorts your picture the most.
- Open the month and group or sort by merchant.
- Read the biggest merchants first, along with the category each row currently has.
- For each, decide the one category that fits, then correct any row that disagrees.
- Skim the small merchants last; many of them are fine as they are.
When should a decision become a rule?
If you would give the same answer every time, make it a rule. A rule applies before history and AI, so it is the firmest way to keep a repeat merchant in the right category. Good candidates are your electricity board, your internet provider, your regular pharmacy, the milk vendor and anything with a fixed purpose.
Leave genuinely mixed merchants alone. A department store may be groceries one month and clothing the next, and a rigid rule would be wrong half the time. Correct those by hand when they come up.
Aliases for messy names
Some merchants arrive under several descriptions: a payment handle one week, a truncated name the next. Add a merchant alias so every variant maps to one clean name. Then grouping works properly, because the same merchant is no longer split across five labels, and rules and history can recognise it.
A fifteen-minute monthly routine
- Import the statement and wait for the batch to finish.
- Group by merchant and work through the top ten by amount.
- Make a rule for anything you corrected that will repeat.
- Add an alias for any merchant that appears under more than one name.
- Clear what is left in review, then look at the month's summary.
Each month, the pile of unfamiliar merchants shrinks, because your decisions are stored. The review becomes mostly a check that nothing odd slipped in, and that is exactly what you want it to become.
If you share a household and one person pays most of the small bills, review that person's merchants together once a quarter. Two people often categorise the same shop differently, and a rule settles the argument without another conversation.
Useful ideas start with a clearer picture of your own spending.
Explore your reports


