The Indian finance function has been improvising for decades.
Data lives in five places, none of them talking to each other, so the close arrives late and the forecast is guesswork. This is not a discipline problem. It is a plumbing problem.
Reconciliation becomes a full-time job disguised as a monthly task, and the person doing it is usually the person you hired to think.
The board reads yesterday's newspaper. By the time a signal is visible, the moment it described has already moved on.
Working capital drifts for two quarters before anyone notices, and by then it is a crisis to manage rather than a decision to make.
Why it stayed this way for twenty years
Nobody chose this. It is what happens when a market gets excellent bookkeeping software and almost no layer above it.
It is genuinely good at what it does, and most of India runs on it for that reason. But a voucher store was never meant to answer "what is my runway", and asking it to has always meant exporting to Excel.
SAP and NetSuite answer these questions well, at implementation costs sized for companies an order of magnitude larger. The mid-market was left to improvise with spreadsheets.
Since 2017 every invoice above the threshold has existed as structured, timestamped, machine-readable data. Almost none of it gets used for anything beyond filing the return.
Skilled people spending the first fortnight of every month rebuilding the same workbook, because no software sat between the books and the board.
What changes in the first month
Not a transformation programme. Three specific things stop being true.
Entries are read nightly, so nothing waits for a month-end scramble.
Everyone reads the same figure at the same time, and the argument about whose number is right ends.
Cash, collections and margin projected forward from your own history rather than a template.