A 13-week forecast tells you the date you run short, while there is still time to act. Here is the line-item structure, where SMEs get it wrong, and how to keep it honest.
A 13-week cash flow forecast tells you the week your bank balance goes short, far enough ahead that you still have options. It works on money in and money out, not invoices raised, and you rebuild it every week. For most Indian SMEs it is the single most useful financial document they do not currently produce.
Thirteen weeks is one quarter, counted the way cash actually moves.
Why the annual budget doesn't cover this
A budget answers whether the year works. It is built on accruals, agreed once, and revisited quarterly if you are disciplined. None of that helps on the Tuesday you discover that a ₹40 lakh receipt slipped to next month and salaries go out on the 30th.
Cash and profit come apart constantly, and the gap is where businesses fail:
- You invoice in March, the customer pays in June. Profitable in March, short in April.
- You buy inventory in advance of a season. Cash leaves months before revenue arrives.
- You pay GST on invoices raised, including the ones nobody has paid you for yet.
A 13-week forecast is the only routine document that catches these before they become a phone call to your bank.
The structure
Weeks across the top, line items down the side. Every figure is money hitting or leaving the bank in that week — not what you earned, what settles.
| Line | What goes in it |
|---|---|
| Opening bank balance | Actual balance, all accounts, start of week |
| Receipts — customer collections | By customer, on their real payment behaviour |
| Receipts — other | Grants, refunds, interest, loan drawdowns |
| Payments — payroll | Salaries, contractor payments, reimbursements |
| Payments — statutory | GST, TDS, advance tax, PF, ESI, on due dates |
| Payments — suppliers | By supplier, on actual terms you are honouring |
| Payments — rent & utilities | Fixed monthly commitments |
| Payments — loan servicing | EMIs, interest, any scheduled repayment |
| Payments — capex & one-offs | Equipment, deposits, annual renewals |
| Net movement | Receipts minus payments |
| Closing bank balance | Opening plus net movement |
| Headroom vs minimum balance | Closing minus the floor you refuse to go below |
That last row is the one people leave out, and it is the one that matters. Pick a minimum operating balance — one month of fixed costs is a reasonable starting point — and track the gap to it. A forecast that shows ₹3 lakh in the bank looks fine until you remember that ₹18 lakh leaves on the 7th.
Building the first one
- Take the actual bank balance today. All accounts, combined, real cleared balance. Not the ledger, not what the accounting software says. This is the only number in the model you know for certain.
- List every open receivable with a realistic date. Not the invoice due date — the date that customer actually pays. Check their last few invoices. Aged debtors go in the week you honestly expect them, or not at all.
- Put the fixed outflows in first. Payroll, rent, EMIs, statutory dues. These are dated and non-negotiable, so they anchor the model.
- Add variable payments by supplier. What you owe, when you intend to pay it. If you are already stretching a supplier, model what you are actually doing, not what the terms say.
- Calculate forward, week by week. Closing balance becomes next week's opening.
- Mark the first week headroom goes negative. That date is the output of the whole exercise.
The first build takes half a day and is uncomfortable, because it makes stretched payables and slow collections explicit. That discomfort is the deliverable.
Running it weekly
Same day every week. Update three things: actual closing balance, what really got collected, what really got paid. Then roll the window forward a week.
Also compare last week's forecast to what happened. If you predicted ₹22 lakh of collections and got ₹9 lakh, the variance is more informative than the forecast — it tells you your collection assumptions are optimistic, and by how much. After a month of this you will forecast collections far better than you do today.
Where SMEs get it wrong
Forecasting revenue instead of receipts. The most common error by a distance. Sales closed this week is not cash this week. Model the collection date.
Leaving out statutory payments. GST, TDS, advance tax and PF get forgotten because they are not vendor invoices. They are also the ones you cannot delay.
Aggregating customers. "Collections: ₹35 lakh" is not a forecast, it is a hope. Name the customers. Five named receipts you can chase beat one blended number you cannot.
Optimism about your own behaviour. If you have been paying a supplier at 75 days, do not model 45 because that is the agreement. Model what you will do.
Building it once. A forecast from six weeks ago is a historical document. Its entire value is being current.
When it shows a shortfall
That is the forecast working. You now have weeks of notice instead of days, and the options are all easier the earlier you use them:
- Accelerate collections. Chase the specific named invoices landing near the shortfall week. Consider a discount for early settlement on the largest one.
- Reschedule discretionary outflows. Capex, annual renewals, non-critical projects — move them past the pinch point.
- Talk to suppliers before you miss. A supplier told in advance is managing a plan. A supplier who finds out by not being paid is managing a problem.
- Arrange facilities early. An overdraft or working-capital line negotiated eight weeks out, from a position of visibility, prices very differently from one requested the week you need it.
- Time statutory payments deliberately. Fund them on schedule. These are the wrong ones to slip, and the penalties compound.
The businesses that get into trouble are rarely the ones with a shortfall in week nine. They are the ones that found out in week nine.
Where CapEasy fits
We build and run the 13-week forecast as part of our Virtual CFO work — set up on your actual bank data and collection history, updated weekly, reviewed with you monthly alongside the MIS pack.
If you would rather build it yourself, the table above is the whole model. The discipline of updating it every week is the part that makes it worth anything.
Frequently asked questions
Why 13 weeks and not 12 months?
Thirteen weeks is one quarter, expressed in the unit cash actually moves in. An annual budget tells you whether the year works; a 13-week forecast tells you whether Tuesday works. It is short enough that you can name every large receipt and payment, which is what makes it accurate enough to act on.
How is this different from a P&L projection?
A P&L records revenue when you invoice. A cash forecast records money when it actually lands. A profitable month can still be a month you cannot make payroll, and only the cash view shows that. Build it on receipts and payments, not accruals.
How often should it be updated?
Weekly, same day, without exception. Roll the window forward one week each time so you always see a full quarter ahead. A forecast updated monthly is a report; updated weekly it becomes an early-warning system.
What if our customers pay unpredictably?
Forecast by customer, not in aggregate, and use each one's actual behaviour rather than your payment terms. A customer whose last six invoices cleared at 60 days is a 60-day customer, whatever the contract says. Unpredictability is a reason to forecast more carefully, not less.
Who should own the forecast?
Whoever can see both the bank account and the sales pipeline — usually the finance lead, with the founder reviewing it. It should not sit with someone who has to ask permission to find out what got collected this week.
What GST and TDS outflows should we include?
Every statutory payment with its actual due date: GST, TDS, advance tax, PF and ESI. These are the payments SMEs most often leave out, and they are the least negotiable ones in the list. Missing them is what turns a forecast that looked fine into a shortfall.