Why expense trackers fail in India

Almost every expense tracker in India fails at the same point, and it is not the point you would guess. It is not the bank connection, or the charts, or the budgets. It is that the app cannot tell you what your money was actually for.

We tested this on ourselves before building anything. One of us tracked 96 transactions across a full month on a normal Indian bank account, ran them through the most popular tracker in this category, and counted how many came back with the right category and the right amount against the right person.

Thirty nine. Out of ninety six.

The other fifty seven were blank, wrong, or technically the right category with the wrong amount, because the payment was split between people and the app had no idea how to divide it. That is the problem this page is about.

Why India breaks them

An American budgeting app has an easy job. Nearly every payment goes to a registered merchant with a clean name on it, and there is a card network behind it handing over a merchant category code. The app looks up the code and it is done.

India does not work like that. Your statement is full of lines like UPI/DR/4318726/RAJESH K/YESB/rajeshk@ok. There is no merchant category code. There is barely a merchant. It could be the vegetable seller downstairs, your maid's salary, your share of last weekend, or your brother repaying you. The string looks identical in all four cases.

So the app does what it can, which is guess from a name, and it guesses badly. Then it hands you a screen full of "Miscellaneous" and asks you to fix it yourself. People do that for four or five days, get bored, and go back to not tracking anything. That is the real reason most Indians who want to track their money end up in a Google Sheet.

More than 30 million urban Indians track their money every month, and most of them do it in a spreadsheet or the Notes app. Not because they like spreadsheets. Because no app has been accurate enough to replace one.

A third go to people

Roughly a third of a young Indian's payments are person to person. A big slice of that is just settling up with friends.

There is no dataset anywhere that can tell you what your ₹840 to Rajesh was for. No model trained on merchant names solves it, because Rajesh is not a merchant. The only thing that answers the question is knowing who Rajesh is to you and what the two of you were doing together.

Splitify knows, because it is also the free app you split bills on. When your flatmate transfers you ₹6,200, we do not see a mysterious credit that might be income. We see the settlement of the electricity bill and two grocery runs that are already sitting in your shared ledger. It is your own money coming back, and we book it that way.

Every other tracker in the country counts that as income. Which means their monthly "you saved this much" number is fiction for anyone who lives with flatmates.

Rent split is not rent spent

This one is worth its own section because of how much it distorts the numbers.

Say rent is ₹36,000 and it leaves your account every month because you are the one with the landlord's account details. You split it three ways. Your actual rent expense is ₹12,000. The other ₹24,000 is money you fronted and get back within a week.

A normal tracker sees ₹36,000 leave and books ₹36,000 of rent against you. Then it sees ₹24,000 arrive and calls it income. Your rent looks 3x too high, your income looks inflated, and your savings rate is meaningless.

Splitify handles this because the split and the transaction live in the same app. The bank alert for the ₹36,000 gets matched to the split expense you already created, and only your ₹12,000 share lands in your spending chart. Nothing to fix by hand.

It learns from you

Most trackers treat categorisation as a fixed problem. A model runs once, produces a label, and if it is wrong that is your problem to fix, over and over, forever, for the same merchant.

We treat a correction as a lesson. When you tell us a particular UPI handle is your gym, you are not editing one row. You are teaching the app something it remembers, and it holds for every payment to that handle from then on.

What this means in practice is that the app you use in month three is noticeably better than the one you installed in month one, and you did nothing except confirm a handful of categories along the way.

Where we actually stand

On that same 96 transaction test, we roughly doubled what the leading app managed once we could see the context behind the split expenses. That gap is the whole reason this app exists.

We are still not where we want to be. The bar we hold ourselves to is getting it right almost every time with no effort from you, and nobody in this category is there yet, us included. We measure it every week, and we would rather say that plainly than quote you a number that flatters us.

Setting it up

1. Install Splitify and sign in with your phone number. Splitting is free, so you can stop here if that is all you want.
2. Connect your bank alerts. In India this works over the SMS alerts your bank already sends you and the transaction emails in your inbox. You are not handing over net banking credentials and nothing can move money out.
3. Let it backfill. The app reads the alerts already sitting there, so you get a populated month or two immediately rather than a blank screen you have to feed for four weeks.
4. Fix the first ten it gets wrong. This is the part that pays off. Each correction sticks for that payee permanently.
5. Set two or three budgets. Not fifteen. Pick the categories you actually overspend on and leave the rest alone.

Picking an app

Genuinely useful advice, not a ranking of ourselves at number one.

• If most of your spending is on cards at established merchants and you never split anything, almost any of the SMS based trackers will do a reasonable job. The problem we solve barely applies to you.
• If you live with flatmates or travel with friends a lot, the split problem is most of your inaccuracy, and no app that lacks a splitting ledger can fix it. That is the case for us.
• If you run several accounts and cards, check that the app pools them into one view rather than making you read each bank's own app. That is most of the value for anyone with a salary account, a UPI account and two credit cards.
• If you are not comfortable connecting email or SMS to any app, that is a reasonable position. Use a manual tracker, or use Splitify only for splitting, which needs none of that access.

Whatever you pick, judge it on one thing after four weeks: open the spending chart and see whether you believe the number. If you don't, the app failed, no matter how nice the chart looks.

Common questions