Know which week the cash gets tight.

The P&L for the month looks fine. Revenue is on plan, costs are under control, and the monthly forecast shows cash ending September comfortably above where it started. Then payroll goes out on the last working day, two large clients who were invoiced on 45-day terms haven't paid, GST and PF went out earlier in the month, and for about five days the account is lower than anyone would like to explain to the team.

Nothing went wrong, exactly. The monthly view was accurate. It just averaged the month into a single number, and cash doesn't move in averages. It moves on dates.

That's the gap a weekly cash forecast closes. It won't tell you whether the business is healthy. The P&L and the unit economics do that. It tells you whether there'll be enough money in the account on each specific week, early enough to do something about it when there won't.

Thirteen weeks is a quarter. It's short enough that most of what goes into it is already known: invoices you've raised, bills you've received, salaries you've committed to, statutory dues with fixed dates. And it's long enough to act. You can chase a debtor, move a vendor payment, push a start date back or bring a renewal conversation forward. Discover the gap two days out, and none of those are available any more.

Beyond thirteen weeks, weekly detail becomes guesswork, so it's the wrong tool. The longer view belongs in a monthly forecast built from the plan, which answers a different question: how many months of runway do we have, and what does the next raise need to cover?

Much of the cash forecasting advice online was written for American companies. In India a big share of the outflows land on statutory dates that don't move, which is useful: they're the easiest part of the forecast to get exactly right.

Receipts have their own Indian wrinkle. When a business customer pays your invoice for professional services, they will usually deduct TDS, often 10% of the fee, under what is now section 393 of the Income Tax Act 2025. That money isn't lost, but it isn't arriving in your bank either. It sits as a credit against your own tax. A forecast that expects the full invoice value overstates receipts from every such customer.

The same goes in the other direction for GST. The GST on your sales invoices comes in with the receipt and goes out on the 20th. Treating it as available cash is one of the most common ways a founder ends up short in the third week of the month.

An illustrative example. A digital agency with around ₹2.5 crore of annual billing, 28 people, salaries paid on the last working day of the month, and a minimum cash line of ₹15 lakh that the founder has decided never to go below. Figures in ₹ lakh.

A monthly view would eventually show September ending around ₹11 lakh, but only once the month had closed, by which point the breach has already happened. The weekly view shows it five weeks ahead, and exactly when: week five, the payroll week, is ₹4.3 lakh under the minimum. And it shows why. A ₹16 lakh receipt the founder had pencilled in for week four is actually due in week six, because that client has paid on day 58 of 45-day terms for the last three invoices.

Seen in week one, there are options. Ask that client for part payment against the milestone already delivered. Move a ₹3 lakh software renewal from week five to week seven, which the vendor agrees to without fuss. Hold a new joiner's start date back by two weeks, which saves one partial month of salary. None of those is dramatic, and together they keep week five above the line.

Seen on the Monday of week five, the only option left is an awkward conversation.

1. 01Start from cash reconciled to the bank, across every account, gateway and wallet. A forecast that starts from the wrong number is wrong every week after.
2. 02Forecast receipts from open invoices customer by customer, using when each customer actually pays, not the terms on the invoice. Net of TDS where they deduct it.
3. 03Add receipts from billing you haven't raised yet, kept separate and conservative, so it's obvious how much of the forecast depends on things that haven't happened.
4. 04Lay in payroll and the statutory calendar, which are the most predictable lines.
5. 05Add vendor payments from open bills and committed contracts, and the one-offs: deposits, annual renewals, equipment.
6. 06Set a minimum cash line and flag every week that falls below it.
7. 07Each week, replace the week just gone with actuals, add a new week at the end, and write down why anything differed by more than a set amount.

Step seven is the one teams drop, and it's the one that makes the forecast get better. After a couple of months of variance notes, you know which customers to trust and which assumptions to haircut.

When a new hire can start

Not whether you can afford the role, which is a plan question, but which week the first salary can land without pushing a month below the line. Sometimes a two-week shift in start date is the entire difference.

Which invoices to chase first

Collections effort goes where it moves the tightest week, not to the oldest invoice or the most annoying customer. The forecast tells you which receipt the next month actually depends on.

What to ask vendors for

Asking a supplier to move one payment by two weeks is an ordinary request when it's made a month early. It becomes a signal of distress when it's made the day before.

When to talk to investors

If the forecast shows a gap that operating changes can't close, your investors should hear about it with a quarter's notice and a plan, not with three weeks' notice and a request.

Whether a big order is good news

A large order that needs inventory, contractors or media spend up front can make a healthy month very tight. The forecast shows the dip before you commit, so you can negotiate an advance or staged billing.

• ·A 13-week forecast file set up for your business, fed from your receivables and payables exports.
• ·A customer payment table showing how late each significant customer really pays.
• ·Your statutory calendar, filled in for the dues that apply to you.
• ·A variance log and a short weekly review routine for whoever owns the forecast.
• ·A monthly runway view linked to the plan, so the short-term and long-term numbers agree.
• ·A written guide to running it without us.

The forecast is built in a spreadsheet on purpose. Cash tools inside accounting software are useful for looking back, but they rarely handle the judgement a good forecast needs: which customer to haircut, which one-off to leave out until it's certain, what the minimum line should be. Once your team has run the spreadsheet for a few months and knows which inputs matter, moving it into a tool is easy. Doing it the other way round tends to automate the wrong assumptions.