Thirteen weeks of cash, week by week.

A monthly cash forecast averages the month. Cash doesn't work in averages. Salaries go out on one day, GST on the 20th, PF on the 15th, and two large customers pay whenever their systems get round to it. A month that ends comfortably can still have a week where the balance is lower than anyone would like to explain.

Thirteen weeks is the useful horizon because most of it is already known. The invoices are raised, the bills have arrived, payroll is committed, the statutory dates don't move. And it's long enough to act: chase a debtor, move a payment, delay a start date, bring a renewal forward.

This template is the version that works for Indian companies, which mostly means the statutory calendar is built in and receipts are entered net of the TDS your customers deduct.

The summary shows closing cash twice. One line assumes every receipt arrives as entered. The other counts only the receipts you marked Certain.

The certain-only line is the one to plan against. Cash forecasts almost never fail because costs surprised anyone. They fail because receipts marked as likely turned out to be hopeful.

In the example that ships with the file, the company looks comfortable on all receipts, never dropping below ₹45 lakh. On certain receipts only it falls to about ₹11 lakh, below its ₹15 lakh minimum, in three of the thirteen weeks. Both lines are true. Only one of them tells the founder to act, and it's the one most spreadsheets don't show.

1. 01Start from a reconciled opening balance. Add every bank account, the payment gateway and any wallets, and check the total against statements rather than the banking app.
2. 02Set the minimum cash line deliberately. One full payroll plus the statutory dues that follow it is a sensible floor for most companies.
3. 03Enter open invoices customer by customer, in the week that customer really pays, not the week the invoice is due. Use their payment history rather than your terms.
4. 04Enter receipts net of TDS. A ₹10 lakh professional services invoice with 10% TDS deducted brings in about ₹9 lakh plus the GST, and the TDS returns later as a tax credit.
5. 05Fill in payroll and the statutory dates from the calendar sheet.
6. 06Add vendor payments from open bills, and the one-offs people forget: annual software renewals, insurance, advance tax, deposits, equipment.
7. 07Read the summary. If the certain-only line goes below the minimum, you have found the thing the forecast was built to find.

Setting it up properly takes a couple of hours the first time. After that, the weekly update is usually under an hour, and most of that hour is spent on the receipts rather than anything else.

The forecast earns its keep as a habit rather than a document. A simple Monday routine:

• ·Update the opening balance from the bank.
• ·Move anything that didn't happen last week into the variance log, with a one-line reason.
• ·Shift the weeks along: delete the week that has gone, add a new week thirteen out.
• ·Re-mark confidence on receipts in light of what you learned.
• ·Look at the certain-only line and the flags, and decide what needs doing this week.

After two months of variance notes, patterns appear. This customer always pays a fortnight after terms. That vendor invoices late. Marketing spend runs above plan in the first week of a campaign. The forecast stops being a guess and starts being a model of how your business actually moves money.

Seeing a tight week six weeks out is the whole point. There are usually more options than founders expect, and they're worth taking in order.

1. 01Collect. Call the finance contact at the two or three largest debtors. Often the delay is a missing purchase order reference or an invoice stuck in a portal.
2. 02Bill. Anything delivered but not invoiced, plus milestones and change requests.
3. 03Move. Ask a vendor to shift one payment by two weeks, which is an ordinary request made a month out and a distress signal made the day before.
4. 04Stop. Discretionary spend: new campaigns, travel, tools, hires not yet started.
5. 05Bring forward. Offer an annual prepaid option to a customer renewing soon, priced honestly.
6. 06Escalate. If the gap can't be closed, your investors should hear about it now, with the forecast attached, rather than in three weeks with a request.

Services firms and agencies

One receipt row per client invoice, entered net of TDS, in the week that client actually pays. Add a row for payroll of contractors, who often need paying before your own clients settle.

Software companies

Gateway and subscription settlements as one recurring row, annual prepaid renewals as their own rows in the week they're expected, and a row for any refund cycle you run.

Restaurants and cloud kitchens

Aggregator settlements land weekly, so they fit this format neatly. Enter them net of commission, and keep supplier payments split between daily fresh produce and monthly accounts.

Healthcare services

Split receipts by payer: walk-in cash arrives the same day, corporate and insurance receipts land weeks later and often short of the amount billed. Model the deduction rate rather than the invoice value.

In every case the structure stays the same. What changes is how many receipt rows you need and how confident you can be about the timing of each.

The calendar sheet carries the ones that catch people out, with their usual dates as of September 2026. Your CA should confirm which apply to you.

• ·TDS you deducted is deposited by the 7th of the following month, with March deductions due by 30 April.
• ·PF and ESIC contributions go out by the 15th, employer and employee shares together.
• ·GST for monthly filers is paid with GSTR-3B by the 20th. Quarterly filers under QRMP pay by challan on different dates.
• ·Professional tax depends on the state: Karnataka runs monthly, Kerala half-yearly through local bodies.
• ·Advance tax instalments fall on 15 June, 15 September, 15 December and 15 March for companies expecting to be profitable.
• ·Salaries are due on your own payroll date, with wage payment deadlines set under the Code on Wages.

Between them these typically account for a third or more of a month's outflows, and they are the easiest part of the forecast to get exactly right.

This forecast answers whether you have enough money each week. It doesn't tell you whether the business works, and it isn't a runway calculation. Runway is a monthly question answered from burn; the weekly forecast is about timing.

Most companies need three things that agree with each other: this weekly view for the next quarter, a monthly plan for the year, and a monthly MIS that reports what happened. When the weekly and monthly views disagree, it's almost always the monthly one that's optimistic about collections.

It's worth reconciling them once a month, in about ten minutes. Take the closing balance this forecast predicted for the month just gone, compare it with the actual bank balance, and find the difference. If it's small, both models are working. If it's large and in the same direction every month, one assumption is wrong and it's usually receivable timing, which is exactly the assumption the variance log is there to correct.