Similar ideas
Split transactions across categories. Not sure this solves the issue as with the current chosen categories it’s not obvious what % belongs where. Plus from a UX perspective I can’t conceive of an elegant, fast and unobtrusive solution.
Custom categories. This I guess would be ideal but it isn’t really enough as the defaults ought not to be left broken. It shouldn’t be hard to come up with a set of categories that are actually unique to each transaction. Users shouldn’t have to do the work themselves either.
The issue
@yen puts it really well:
I too as a Monzo user have had the thought that Categories need improving! Is paying my rent “Finances,” “Bills,” “Expenses,” “Family,” or “General”? I have no idea!
Since Aristotle and Callimachus, the model of categories we use is one where each option needs to be clearly defined, mutually exclusive and collectively exhaustive.
More recently, multi-label classification is a well-trodden problem in machine learning that anyone designing such a system would be familiar with.
This way, any entity in the given classification universe belongs unequivocally to one, and only one, of the proposed categories.
This is how databases work. This is how drop-down menus work. This is how bullet choice (as opposed to check boxes) work.
It’s a well trodden design paradigm. It allows us to make useful insights. If the same item can be in one category one day but a different one the next day it destroys the data integrity and makes analyses useless.
Following universally-recognised design standards, when the user has exactly one choice, it’s vital to make sure that the options are both comprehensive and clearly distinct.
Otherwise it wastes time because it effectively requires the user to run a evaluation metrics analysis each time they categorise or, as bad, a lookup of the last use for a similar transaction. The goal should be to make the user’s job as simple as possible.
Current situation
The categories are:
- Bills
- Charity
- Eating Out
- Entertainment
- Expenses
- Family
- Finances
- General
- Groceries
- Holidays
- Personal Care
- Shopping
- Transport
There is far too much crossover (and therefore probably far too many categories). It complete unfathomable how we ended up with this mess in the first place!
If it’s not obvious to everyone, you have to ask what’s useful to users when using categories to analyse their finances. I’m not really sure what the point of doubling a lot of these up are. e.g. Expenses vs Bills. And what’s Finances for?!
So for example is your train season ticket Expenses, Bills or Transport? This might even depend on how you organise other transactions.
Is a big Tesco shop Groceries or Shopping? What if you also picked up a DVD? Or some kids clothes? Or a EU converter plug? Or some plants and compost? Or a phone top up? Or a bottle of brandy for granny?
Example
Also realise that tags can be used as subcategories. So it could make more sense to reduce it to, say,
- Entertainment (includes eating out, gigs, festivals, drinks, shows, cinema)
- Expenses (includes utility bills, tax, etc.)
- Shopping (includes groceries, clothing, gifts, etc.)
- Transport (includes flights, car repairs, inner tubes, trains, petrol, etc.)
- Personal
- Other
This is obviously not ideal and definitely not a suggestion for the actual categories, but is already a vast improvement.
The very first thing should be making exclusive categories where it’s obvious where each item goes and it can only go in one place.
Market research
Theory is all well and good but how do other company’s group financial transactions? There must be a lot of prior art devising a sensible list?
Sources
10 Budget Categories That Belong in Your Plan | Quicken
Categorize Spending with Common Budget & Expense Categories | Mint
Budgeting 101: Personal Budget Categories | First Bank
Also worth looking at these for inspiration:
Obviously these aren’t perfect but with a little tweaking a vast improvement on the current shambles is easy. Just start from a well-informed and planned beginning considering the use-cases and the purpose of the feature as well as the relevant technical literature.