Ten years ago, the finance function I ran inside a large multinational had capabilities a ₹50 crore Indian business could not have bought at any price: a three-day close, margins by product and customer on a screen, a cash forecast that updated itself. Today that capability costs less than one junior accountant a month. Almost nobody at ₹10–200 crore has it. This article is about what becomes possible when you do.
I spent twenty-five years in finance leadership at Shell, Schlumberger, Weatherford and Equifax, and now work directly with promoters and founders in India. The large companies were not better run because their people were smarter. Every decision-maker had the same current, trustworthy numbers, and routine work was done by systems rather than memory. Both are now within reach of a business a hundredth of the size.
What "transformation" means here, and what it does not
Software vendors have worn the phrase out, so let me be precise. Finance transformation is the move from a finance function that records what happened to one that shapes what happens next. It has three components, and all three must change together.
- Process. A documented rhythm: when the books close, who approves what, which numbers are reviewed each week and month, and what happens when they are off plan.
- Systems. Connected tools that do the routine work, so a bank transaction, an invoice and an approval each happen once and flow everywhere they are needed.
- People. A team that spends its time on analysis and control rather than typing and chasing, led by someone who has seen what good looks like.
It is not a new ERP on its own. I have watched businesses spend eighteen months implementing a system that reproduced their old spreadsheets, badly. It is not a dashboard nobody opens after the second month. And it is not headcount. Most businesses I work with end up with a smaller finance team doing more valuable work.
The eight shifts, side by side
This is the table I draw on a whiteboard in the first meeting with a promoter. The left column is the finance function I find in most ₹10–200 crore businesses, including well-run ones. The right column is what the same business can have within a year, with tools that already exist and a team roughly the size it has today.
| Where most businesses are | What a modern finance function looks like |
|---|---|
| Books closed in 10–15 days, in spreadsheets, by heroic effort | Month-end in three working days or fewer, on a close calendar, with automated bank reconciliation |
| Cash position known when the bank statement arrives | 13-week cash forecast refreshed weekly, with alerts before a shortfall |
| One annual budget, prepared in October, forgotten by March | Rolling 12-month forecast and three scenarios, updated every month |
| Margins known only at company level, once a year | Contribution margin by product, customer and channel on a live dashboard |
| Approvals over WhatsApp, on the promoter's memory | Approval matrix and maker-checker controls inside the ERP, on mobile |
| Collections chased when cash gets tight | Automated reminders, enforced credit limits and an ageing review every week |
| Tally for compliance; everything else in Excel | Cloud ERP, bank feeds, e-invoicing and dashboards in one connected stack |
| Your CA tells you what happened last quarter | A CFO tells you what to do next week, with the data to back it |
None of the eight is exotic; each is standard practice in any listed company. The gap between the columns is no longer money or technology. It is knowing what to build, in what order, and having someone accountable for building it.
Shift one: run the business on the numbers, not after them
Everything else depends on the close. A business that closes on the 12th is reviewing March in the second half of April, when the decisions March should have prompted are three weeks late. The slow close is rarely a systems problem. It is bank statements typed by hand, vendor bills that arrive when the vendor remembers, accruals nobody books until the auditor asks, and one person who holds the whole picture in their head.
The three-day close is built in order. A close calendar, each task owned by a named person on a named day. Bank feeds, so reconciliation becomes exception review rather than data entry. Standing accruals booked on day one rather than reconstructed on day ten. A short list of balance-sheet accounts reconciled every month, so the audit finds nothing new. Where I have led this, the close has typically come down from over ten days to under three within two quarters, and the team has spent less time at month-end, not more.
Once the close is fast, the monthly review becomes useful. I keep it to one page: six to eight numbers that matter for that business, against plan and last year, with a line of commentary each. Revenue, gross margin, operating cash flow, debtor days, inventory days, net debt, and one or two operational measures. It lands in the promoter's inbox before the review meeting, so the meeting is spent on decisions rather than on finding out what happened.
Shift two: steer cash, do not just count it
Every promoter knows the bank balance. Few can say what it will be in six weeks. That is where the stress lives: a GST payment, a salary run and a slow customer landing in the same fortnight, and a profitable business asking its bank for an enhancement it did not expect to need.
The 13-week cash forecast changes this more than any other single tool. It lists, week by week, what is expected in and out from today's balance. The first version takes an afternoon; maintaining it takes twenty minutes on a Monday. It does not change the cash. It changes the warning time, and six weeks is long enough to pull a collection forward, defer a purchase or speak to the bank before the problem rather than during it.
Behind it sits the rolling 12-month forecast, re-cut every month with the latest actuals and three scenarios while the annual budget stays as the commitment to the board or the bank. When a large customer delays or a raw material spikes, the business knows the consequence that week, not at the next budget.
The cash itself usually improves alongside, because the forecast makes the working-capital levers visible. In a ₹30–50 crore manufacturer, tightening collections, cutting dead stock and renegotiating supplier terms has typically released ₹2–4 crore, with no new sale. A trading business of the same size has typically freed 15–25% of its working capital. These are what happens when somebody owns the number every week.
Shift three: controls that live in the system, not in the promoter's head
In most ₹10–200 crore businesses the real control system is the promoter. Every payment above a threshold gets a WhatsApp message and a thumbs-up. This works at ₹10 crore. It fails at ₹60 crore, usually not through fraud but through volume: the promoter approves two hundred items a week without reading them, and the thumbs-up stops meaning anything.
The replacement is an approval matrix: one page saying who may approve what, up to what amount, in which system. It is then configured in the ERP so that a purchase order, a payment and a credit note each need a maker and a checker, prompted on their phone. The promoter approves fewer items and reads each one, because the routine ones no longer reach them.
Collections work the same way. Credit limits move from a figure in a file into the system, where the next order is held when the last one is overdue, and reminders go out without anyone drafting them. The weekly ageing review then deals with exceptions rather than the whole book.
What the stack looks like, and what it costs
The stack is now well established: a cloud ERP, usually ERPNext or Odoo, or Tally on cloud with the right add-ons for businesses not ready to move; bank feeds and GST e-invoicing connected; a dashboard the promoter can open on a phone; approvals inside the ERP; payroll and expenses feeding the ledger rather than re-typed into it.
The software typically costs a ₹50 crore business ₹1.5–4 lakh a year. Implementation, done properly, is a one-time cost of the same order, spread over six to nine months. One junior accountant costs ₹3–5 lakh a year, and most businesses of this size employ three or four. The return is not mainly headcount, although the team usually shrinks by one as retyping disappears. It comes from cash released from working capital, from borrowing cost that falls by 100–250 basis points once lenders see clean, current numbers, and from decisions made on time because the information is there.
How to sequence it: 90, 180 and 365 days
Order matters more than speed. Businesses that start with the ERP tend to automate their existing confusion. The sequence I follow starts with what the business already has.
- First 90 days. Chart of accounts cleaned up to carry the reporting the business needs. Close calendar in place and the close down to five days. 13-week cash forecast live. One-page monthly MIS before the review meeting. Approval matrix agreed on paper.
- Days 90 to 180. Bank feeds and e-invoicing connected. Close under three days. Rolling forecast with scenarios. Credit limits, automated reminders and the weekly ageing review switched on. ERP decision taken, if needed, against a written list of requirements.
- Days 180 to 365. ERP implemented or upgraded, with the approval matrix inside it. Contribution margin by product and customer on the dashboard. Quarterly lender pack sent without being asked. Finance team moved from data entry to analysis and control.
By the end of the year the business has the right-hand column of the table, and a finance function that keeps improving on its own, because the rhythm and the ownership are now part of how the business runs rather than a project that finished.
Frequently asked questions
Do we need to replace Tally to do any of this?
Not to start. The close calendar, the 13-week cash forecast, the one-page MIS and the approval matrix can all be built around Tally, and I usually do the first six months that way. The ERP decision comes later, against a written list of what the business needs. For many businesses under ₹50 crore, Tally on cloud with the right add-ons remains the sensible answer.
How long before we see a return?
Usually cash, within the first quarter, because the forecast and the ageing review surface collections that have simply been left alone. The borrowing-cost benefit follows once a lender has seen two or three quarters of clean, timely numbers. The full stack, including the ERP, typically pays for itself inside eighteen months, and the working-capital release alone has often covered it in the first year.
Who leads this if we do not have a CFO?
This is the work a fractional CFO exists for. A full-time CFO is rarely justified below ₹200 crore, but the transformation needs somebody who has built a modern finance function before and is accountable to the promoter for the result. A fractional CFO working a few days a month leads the design, sequences the work, manages the implementation partners and runs the monthly rhythm until the team can run it alone.
Closing thought
The businesses I admire most at ₹10–200 crore are not the ones with the biggest finance teams or the most expensive software. They are the ones where the promoter opens one page on a Monday morning and knows where the cash is, where the margin is and what needs a decision this week. That is what a modern finance function makes possible, and it is now available to any business willing to build it in the right order.
If you would like to know which of the eight shifts your business should start with, book a confidential 30-minute strategy call. We will look at your close, your cash visibility and your controls, and map out the first ninety days. You can also read how we run finance transformation and ERP and systems advisory engagements.