When you have hundreds of rows to categorise, count how often each description appears and fix the most frequent ones first. A few merchants usually account for most rows, so five or six edits can cover well over half the list before you handle the long tail.
How do you count repeats?
With descriptions in column B and 200 rows, put this helper formula in D2 and fill down:
=COUNTIF($B$2:$B$201, B2)
Each row now shows how many times its description appears. Sort the sheet by column D, largest first, and the repeating merchants rise to the top. Delete the helper column when you finish.
What does the payoff look like?
Say the counts come out like this. The numbers are invented.
| Merchant group | Rows | Running total of rows | Share of 200 |
|---|---|---|---|
| Supermarket | 34 | 34 | 17 percent |
| Fuel stop | 22 | 56 | 28 percent |
| Coffee | 19 | 75 | 38 percent |
| Takeaway | 17 | 92 | 46 percent |
| Streaming | 12 | 104 | 52 percent |
The running total adds 34 + 22 + 19 + 17 + 12, which is 104 rows. Five edits cover 104 of 200 rows, or 52 percent. The other 96 rows are spread across smaller groups and one-offs.
What order should the rest go in?
- Anything with 5 or more rows: one edit each.
- Rows with 2 to 4 rows: group by type, such as all rail and bus lines, and edit together.
- True one-offs: assign by hand or leave as Other.
How do you apply a category to a group at once?
Sort so the group is together, select the category cells for those rows, type the category in the first one and copy it down the selection. Because the rows are contiguous after sorting, filling down is safe. Sort again by date afterwards if you want the original order back, and keep a date column intact so you can.
How do you check the job?
The sum of amounts and the number of rows must match the start. Suppose the file held 200 rows totalling A$7,915.20; after editing it must still be exactly that. Also scan the largest amounts in each category, since a wrong category on a big row matters most.
What about rows that are almost identical?
Descriptions often carry a reference number or a date, so the same shop appears as several different strings and COUNTIF treats them separately. Clean the text first: use a helper column with =LEFT(B2, 12) to keep the first twelve characters, or use Find and replace to remove the trailing numbers. Count again on the cleaned column, and the real groups appear.
Keep the counting method for next month. New rows can be appended under the old ones, the helper formula filled down, and the same top merchants will show up first. After two or three months most of your work is checking a short list of new names rather than starting again.
When two groups seem to overlap, such as a takeaway chain that also sells coffee, decide a rule once and write it down: for example, anything under a certain amount is Coffee and anything above is Takeaway. A written rule keeps the next batch consistent with this one, and it takes ten seconds to apply.
A note on tools
SpendixAI lets you bulk edit and delete transactions, and corrections become history that helps future categorisation. It is built around India and INR, so Australian users should treat the spreadsheet steps above as the main route.
Useful ideas start with a clearer picture of your own spending.
Explore your reports


