How to Prepare a Bankable DPR: What Credit Committees Actually Check

A Detailed Project Report has two audiences and they want different things. The promoter wants a document that makes the project look fundable. The credit committee wants to know what happens if the project underperforms, and whether the promoter has thought about it.

DPRs written for the first audience fail. Here is how to write for the second.

What a credit officer reads first

Not the market study. In our experience of sitting in these meetings, the order is: promoter profile and credit history, means of finance and promoter contribution, debt service coverage ratio in the stress case, and then whether the capital cost numbers are traceable.

Everything else — market sizing, technical configuration, competitor analysis — exists to make those four believable. A DPR that buries the DSCR on page 94 is making the reader work, and readers who have to work get suspicious.

The structure that works

1. Executive summary — one page, with project cost, means of finance, DSCR, IRR and payback stated plainly.
2. Promoter profile — background, relevant experience, existing businesses, credit history, and net worth statements.
3. Project description — what is being made, at what capacity, by what process, on what site.
4. Market study — demand evidence for your district and state, not a national market size, plus competitor supply and realistic price realisation.
5. Technical configuration — mass and energy balance, equipment schedule with specifications, layout, utility loads and manpower plan.
6. Cost of project — line by line, with every number traceable to a quotation, estimate or tariff order.
7. Means of finance — promoter contribution, term loan, subsidy, working capital, with the subsidy timing treated conservatively.
8. Financial projections — ten years, driver-based, with the assumption sheet visible rather than buried in formulas.
9. Sensitivity and risk — base, upside and a genuine stress case, with named mitigations.
10. Annexures — quotations, licences, land documents, LOIs, CVs, and anything you referenced.

Where DPRs fail

• Projected margins above the industry norm with no explanation. A credit officer who processes twenty food files a year knows your sector's margin band.
• Capacity utilisation at 85% in year one. Nothing achieves that. Model 55–65% in year one and ramp it.
• Capital costs from a single supplier quotation with no comparison, which reads as an un-tendered project.
• Working capital sized at an annual average when the business buys a season's raw material in ten weeks.
• A subsidy receipt built into year-one cash flow. Disbursement takes six to fourteen months after commissioning.
• No stress case, or a stress case that still comfortably services debt — which signals the base case is padded.

Building projections a lender believes

Projections should be driver-based, not grown by a percentage. Volume comes from installed capacity times realistic utilisation times operating days. Revenue comes from volume times a price you can evidence from actual invoices or published market rates. Cost comes from a bill of materials, a tariff order, a lane rate and a manpower schedule.

When it is built this way, a credit officer can change one assumption and see what happens — and will. Models where the numbers are hard-coded rather than driven invite the suspicion that the answer was decided before the arithmetic.

Keep the assumption sheet as the first tab, not the last. Every number a reader might question should have its source named on the same line.

A credit officer is not underwriting your best case. Show them the bad case and how you survive it, and the good case becomes credible.

CMA data is not the DPR

CMA data — the forms your lender uses to assess working capital and analyse your financials — is a separate deliverable in the lender's own format, covering past performance, the current year's estimate and projections, fund flow and ratio analysis. Submitting a DPR without CMA data is the most common reason a first submission comes straight back.

The numbers in both must reconcile exactly. Where they do not, the file stalls while someone works out which document is wrong, and that someone is rarely in a hurry.

After submission is half the work

Files are decided in the queries and the site visit as much as in the document. Expect two or three rounds of written queries. Answer them in writing, on the record, quickly, and without changing earlier numbers — a revised projection between rounds is the fastest way to lose a credit officer's confidence.

Prepare the promoter for the site visit. The questions are usually simple and practical: where does the raw material come from, who will run the plant, what happens in the lean season, what is your experience of this product. A promoter who answers these fluently moves the file; one who defers every question to the consultant does not.

What it costs and how long it takes

For a project up to ₹5 crore, a competent DPR with CMA data runs ₹55,000 to ₹1.4 lakh and takes three to five weeks. For ₹5–50 crore projects with multi-lender or consortium funding, ₹2–6 lakh and five to eight weeks is normal.

Add three to four months from submission to sanction, assuming the file is complete. Promoters who compress this by starting construction before sanction take a risk that has ended several otherwise viable projects.