Financial projections a bank will actually believe

A credit officer is not checking whether your projections are ambitious. They are checking whether the ambition has been costed — whether the growth you have projected has the capacity, the working capital and the receivables to match it, and whether anything similar has happened in your business before. Growth that appears in the sales line and nowhere else is the thing that gets noticed.

Which means credible projections are usually less about being conservative than about being consistent.

What does a projection have to be consistent with?

Four things, and a file is judged against all of them.

Your own history. If you have grown at twelve per cent for three years and project forty, the file needs a reason — a new line, a new market, a machine that lifts capacity. Without one, the projection reads as a number chosen to make the application work.

Your capacity. Manufacturing more requires the ability to manufacture more. A sales projection that exceeds installed capacity, with no capex to expand it, contradicts itself.

Your working capital. Growth consumes cash before it produces it — more inventory, more receivables, and suppliers who want paying on the same terms as before. If the facility you are asking for does not fund the growth you are projecting, one of the two is wrong.

Your sector. Officers who lend into building materials know roughly what margins and collection periods look like there. A projection outside those norms is not disqualifying, but it needs saying why.

Which assumptions get scrutinised hardest?

Assumption	What it is checked against	Common overreach
Sales growth	Your history, capacity, market size	A step change with no stated cause
Gross margin	Your last three years	Improving every year for no reason
Collection period	Your actual debtor days	Assuming customers suddenly pay faster
Inventory holding	Your actual stock turns	Holding less stock while selling more
Creditor days	Supplier terms	Assuming suppliers fund the growth
Overheads	Your cost base	Sales doubling while salaries stay flat

The collection period is the one most often adjusted quietly, because shortening it reduces the working capital requirement and improves the ratios. It is also the easiest to check — the bank has your last three years of debtor days. If they were 75 and your projection says 45, you are asserting that customers who have never paid in 45 days are about to start.

Is it better to project conservatively?

Not necessarily, and understating has its own costs.

Project too low and the working capital calculation returns a facility smaller than you actually need. You will be back within the year asking for an enhancement, which is a slower process than getting it right first time — and asking twice in twelve months raises questions about whether you understand your own business.

Project too high and you either get declined, or get sanctioned against numbers you then miss. Missing your own projections matters at renewal, because next year’s file is read against this year’s promises.

The target is your realistic case, stated with its reasoning. Not your best case, and not a deliberately safe number.

How much detail should the assumptions carry?

Enough that someone else could rebuild the projection from them.

“Sales growth of 25%” is a number. “Sales growth of 25%, of which 15% from the new dealer appointments in Gujarat signed in March and 10% from the existing base” is an argument. The second can be tested and believed; the first can only be accepted or doubted.

Three things worth stating explicitly, because they carry most of the model:

• Where the additional sales come from — customers, geography, product, or price.
• What capacity supports them, and whether that needs investment.
• What the working capital cycle does as volume rises, in days rather than rupees.

What about a new business with no history?

The standard of evidence rises, because there is nothing to check the projection against.

What substitutes for history is the project report — demand for the product, comparable businesses, quoted machinery costs, identified suppliers, and a realistic view of how long it takes to reach capacity. A first-year projection at full capacity utilisation is the commonest error in new-project files, and an experienced officer will discount it immediately.

Ramp-up matters more than the eventual number. A projection that reaches 40% of capacity in year one, 65% in year two and 85% in year three reads as considered. One that starts at 90% reads as though nobody has run a factory.

StartYourIndustry exists partly for this — project profiles with investment breakdowns and demand data, so the assumptions in a new-project file come from somewhere rather than from hope.

Last verified: 22 August 2026. General guidance on a bank process, not financial advice. Confirm with your bank or a chartered accountant before relying on it.

Common questions

How many years should projections cover?

Usually two to five, depending on the facility. Term loans generally need projections covering the repayment period, so a seven-year loan needs a longer view than a working capital renewal.

What if we miss the projections after sanction?

It gets discussed at renewal. Missing by a margin with a clear explanation is survivable and common. Missing badly and repeatedly affects how the next file is read, which is the practical argument against optimistic numbers.

Should we show a downside case?

Rarely asked for, and it can help on a larger or riskier proposal — it demonstrates you have thought about what happens if the growth does not arrive. Keep it brief and do not let it become the number the bank sanctions against.

Does the bank verify our assumptions independently?

To a degree. GST returns, bank statements and your audited accounts are all cross-checked, and for larger proposals a site visit is normal. Assume anything checkable will be checked.

More on this: why CMA files get sent back and preparing a DPR for a manufacturing project. See also StartYourIndustry.