How to Write a Commercial Credit Memo: A Nine-Section Framework for Lenders

Write a commercial credit memo in nine sections, in this order: recommendation, borrower and structure, purpose, business and industry, financial analysis, repayment capacity, security, risks and mitigants, and terms. Lead with the decision. Every sentence after it must either support that decision or qualify it — and every figure must be traceable to a source document.

That last point is not stylistic. YuSight cites 100% of figures back to source document and page precisely because an uncited number in a memo is an assertion, and committees approve on evidence.

Key facts

• Nine sections, written in decision order — not chronological order. The most common structural failure in commercial credit memos is burying the recommendation on page nine.
• The OCC expects credit analysis to establish trends, not snapshots: examiners are directed to "analyze balance sheet and profit and loss items in current and preceding financial statements, and determine the existence of any favorable or adverse trends" (OCC, *Comptroller's Handbook: Commercial Loans*, Section 206).
• Guarantor analysis is a global-condition test, not a signature check. The OCC's Commercial Real Estate Lending booklet, version 2.0 (March 2022), frames it as evaluating "borrower and guarantor creditworthiness and global financial condition, including assets (type, amount, and liquidity), global cash flow, direct and contingent liabilities" (OCC).
• Sensitivity analysis is expected, not optional. The same booklet directs banks to set "requirements for feasibility studies and sensitivity and risk analyses (e.g., sensitivity of income projections to changes in economic variables such as interest rates, vacancy rates, or operating expenses)."
• YuSight cites 100% of figures, each clickable to the source page, in a ~30-minute end-to-end CAM draft.

Why does section order matter more than section content?

Because of who reads it and how.

A credit committee member reads twelve memos before a Thursday meeting. They read your first page properly, skim the middle, and stop at anything that contradicts what page one told them. If your recommendation appears on page nine, they have already formed a view from the financial tables — and your careful reasoning arrives after the verdict.

Decision order means: state the conclusion, then justify it, then qualify it, then bind it in terms. A memo that follows the analytical order — how you did the work — reads like a diary. A memo that follows decision order reads like advice.

The nine sections below assume a commercial facility to an operating company. For the definitional treatment of what each section contains, see what a credit assessment memo is. This page is about writing them well.

Section 1: Recommendation and executive summary

What it does: gives the decision, the amount, the price, the tenor, the security and the single reason a reader should be comfortable — in under 200 words.

What good looks like: the reader can stop here and vote. Amount recommended, and whether it differs from the amount requested. Facility type and tenor. Pricing and the basis for it. Risk grade, and whether it has moved. The one sentence naming the primary repayment source. The one sentence naming the material risk. Any deviation from policy, flagged here and not on page eleven.

The test: delete every other section. Does what remains still constitute a proposal a sanctioning authority could act on? If not, section 1 is under-written.

Section 2: Borrower, group structure and facility request

What it does: establishes who is legally on the hook.

What good looks like: the exact legal name as it appears on the constitutional document, the registration identifier, the date of incorporation, and — for anything with more than one entity — a group chart naming every entity whose financials feed the spread, marked obligor or non-obligor. Guarantors are named with guarantee type and cap, not listed as "promoters".

The failure mode here is silent aggregation. A memo that says "the group had revenue of $84m" without naming which entities were consolidated, on what basis, and whether any of them are outside the obligor perimeter is describing a number the lender cannot lend against.

Section 3: Purpose and use of proceeds

What it does: ties the money to a business event you can later test.

What good looks like: "Working capital gap arising from a 22-day increase in receivable days following the shift of two distributors from advance payment to 60-day credit in Q2 FY26" is a purpose. "Business expansion" is not. The difference is that the first can be monitored — receivable days become a covenant, and the memo has told the monitoring team what to watch.

For term facilities, tie the drawdown to the asset: invoice, quotation, contract value, and the borrower's contribution percentage.

Section 4: Business and industry analysis

What it does: explains why this borrower earns money and what would stop it.

What good looks like: specificity about the revenue model and the customer base. Who pays, on what terms, under what contract, with what notice period. Where the borrower sits in its buyer's supply chain and how easily it could be replaced. Two or three named competitors. The regulatory or input-price variable that moves margin most.

Industry commentary lifted from a sector report and pasted in is filler. If a paragraph would read identically in a memo about a different borrower in the same sector, delete it.

Section 5: Financial analysis

What it does: converts statements into a trend and a judgement.

What good looks like: three years plus an interim, with each adjustment named and justified. Say which figures are audited, which are provisional, and which are management-prepared. Name the accounting policy that changed. Explain every movement above your materiality threshold — and set that threshold explicitly.

Two disciplines separate strong financial sections from weak ones. First, every ratio has a stated definition, because DSCR has several variants and they disagree. Second, every figure carries a source: statement, page, line. Our 24-ratio guide covers which metrics earn their place.

Section 6: Repayment capacity and cash flow

What it does: answers the only question that matters — will the cash be there when the instalment is due?

What good looks like: a debt service calculation shown line by line, on the borrower's actual obligations including the proposed facility, with the adjustments visible. Here is that arithmetic for a mid-market manufacturer requesting a new ₹4.00 Cr term loan on top of existing debt.

Step 1 — cash available for debt service, FY26 audited

Step 2 — remove the non-recurring item

Insurance claim settlement of ₹0.62 Cr (Note 22, p. 35) is a one-off and does not recur.

5.41 − 0.62 = ₹4.79 Cr adjusted CADS

Step 3 — existing debt service

Term loan principal ₹1.90 Cr + interest ₹0.86 Cr = ₹2.76 Cr

DSCR on existing debt: 4.79 ÷ 2.76 = 1.74x

Step 4 — add the proposed facility

New term loan ₹4.00 Cr, five years, straight-line principal ₹0.80 Cr; interest at 10.25% on an average year-one outstanding of ₹3.60 Cr = ₹0.36 Cr. Incremental debt service ₹1.16 Cr.

Total debt service: 2.76 + 1.16 = ₹3.92 Cr

DSCR post-facility: 4.79 ÷ 3.92 = 1.22x

Step 5 — the obligation the borrower did not mention

Promoter loan repayment of ₹0.30 Cr per annum is contractual and ranks alongside bank debt.

Total: 3.92 + 0.30 = ₹4.22 Cr. DSCR: 4.79 ÷ 4.22 = 1.14x

The unadjusted headline was 1.96x. The number a committee should decide on is 1.14x. Both are arithmetically correct. Only one is honest, and the memo's job is to show the reader how you got from the first to the second.

Section 7: Security and collateral

What it does: describes the fallback, and how long it would take to reach it.

What good looks like: each item of security with valuation, valuation date, valuer, basis (market, forced sale, book), charge type, charge rank, and registration status. Then the sentence most memos omit: what enforcement would realistically take in months, and what the recovery would be after costs.

A memo that says "adequately secured by hypothecation of stock and book debts" without a drawing power calculation, an ageing of those book debts, and a view on the stock's realisable value has described a legal right, not an asset.

Section 8: Risk factors and mitigants

What it does: names what could go wrong, quantifies it, and says honestly which mitigants are real.

This is where memos most often collapse into comfort language. Two before-and-after pairs.

Risk factors: weak versus strong

Weak:

The company operates in a competitive industry and is exposed to raw material price volatility. There is some customer concentration. Management is experienced and has successfully navigated similar conditions in the past.

Three sentences, no numbers, no consequence. Every clause would be true of almost any manufacturer. A committee member learns nothing and cannot challenge anything.

Strong:

Risk 1 — Customer concentration (material). Alpha Components accounted for 61% of FY26 revenue (₹38.4 Cr of ₹63.0 Cr), up from 44% in FY24. The relationship runs on rolling 12-month purchase orders with 90 days' notice and no minimum volume commitment. If Alpha exits, ₹38.4 Cr of revenue and, at the FY26 contribution margin of 14%, ₹5.4 Cr of contribution disappear — against total debt service of ₹4.22 Cr. On the residual book alone the facility does not service.

Risk 2 — Working capital elongation (moderate, worsening). Receivable days moved 68 → 74 → 91 across FY24–FY26. Each additional 10 days at FY26 revenue absorbs ₹1.7 Cr of cash. The trend, not the level, is the concern.

The strong version quantifies the loss, states the trigger, and names the threshold at which the credit fails. It gives the committee something to disagree with, which is what a risk section is for.

Mitigants: weak versus strong

Weak:

These risks are mitigated by the long-standing promoter experience, the personal guarantees of the directors, and the hypothecation of stock and book debts.

This lists comfort items. It does not say what any of them would actually do in the scenario described.

Strong:

Concentration is partially mitigated. It is not eliminated, and the memo should not imply otherwise.

Real mitigant: the Alpha relationship is tooled. Alpha's dies sit on the borrower's shop floor, and Alpha's own published vendor standard sets requalification of an alternate supplier at 9–14 months. That raises Alpha's switching cost materially; it creates no contractual obligation and should not be scored as one.

Real mitigant: the borrower has added two customers in FY26 contributing ₹6.1 Cr combined, taking non-Alpha revenue from ₹19.2 Cr to ₹24.6 Cr. On present trajectory concentration falls below 50% during FY28. That is a direction of travel, not present cover.

Not a mitigant: the directors' personal guarantees. Combined declared net worth is ₹6.2 Cr, of which 78% is residential property and unlisted holdings. Against a ₹38.4 Cr revenue dependency and ₹16.0 Cr of exposure, a guarantee that cannot be realised inside a workout window is a behavioural instrument, not security. We recommend it be taken and not counted.

That last paragraph is the one that earns a memo its credibility. A mitigant section that only lists things in your favour will be read as advocacy, and an experienced committee discounts advocacy automatically.

Section 9: Terms, covenants and conditions

What it does: converts the analysis into enforceable obligations.

What good looks like: every covenant has a formula, a test date, a frequency, a source document and a consequence. "Maintain DSCR above 1.25x" is not a covenant — it is a wish, because it does not say whose DSCR definition, tested on what statements, on what date, or what happens on breach. The definitions that cause covenant disputes are worth reading before you draft this section.

Conditions precedent and conditions subsequent are separated, each with an owner and a date.

Recommendation: weak versus strong

Weak:

In view of the above, we recommend sanction of the facility as proposed, subject to the usual terms and conditions.

Strong:

Recommend sanction of a ₹4.00 Cr term loan, five-year tenor, at 10.25% floating over the internal benchmark, against the ₹5.00 Cr requested.

The ₹1.00 Cr reduction is deliberate. At ₹5.00 Cr, post-facility DSCR including the promoter loan falls to 1.05x — inside the noise band of a single quarter's collection slippage. At ₹4.00 Cr it holds at 1.14x, which is below our 1.25x policy floor and therefore a documented deviation requiring Level 2 approval, recorded here rather than in an annexure.

The recommendation rests on the FY27 receivable improvement being real. Accordingly it is conditioned on: receivable days tested quarterly against a 91-day ceiling, stepping to 85 days from Q3 FY27; the ₹0.30 Cr promoter loan subordinated in writing before first disbursement; and a review trigger — not an event of default — if Alpha's share of trailing twelve-month revenue exceeds 65%.

If receivable days exceed 100 at any quarterly test, the correct action is to reduce the working capital limit, not to waive.

The strong version tells the committee what the analyst actually thinks, where the credit is fragile, and what to do when it moves. That is advice. The weak version is a signature block with a paragraph attached.

What separates a memo that passes from one that gets returned?

Before you send it, run it against our 47-item credit memo checklist — eighteen of those items are blocking, and a single blocking fail returns the file regardless of how well the rest is written.

FAQ

How long should a credit memo be?

Long enough to support the decision and no longer — typically 8 to 15 pages for a mid-market commercial facility, and two to four for a small ticket. Length is a symptom, not a target. A twenty-page memo usually means nobody decided what mattered.

What should be included in a commercial credit memo?

Nine things: the recommendation, the borrower and group structure, the purpose, the business and industry, the financial analysis, the repayment capacity, the security, the risks and mitigants, and the terms. If a paragraph does not belong to one of those nine, it probably belongs in the file rather than the memo.

How can lenders make credit memo writing faster?

By removing the parts that carry no judgement. Keying, footing, ratio computation and first-draft narrative are mechanical; deciding whether a shareholder loan is debt or equity is not. Automate the first set, protect analyst time for the second — and insist that every automated figure remains traceable to its source page.

Should the recommendation really come first?

Yes. Credit committees read in triage order, not narrative order, so a conclusion that arrives last arrives after the reader has formed one. Leading with the recommendation also disciplines the writer: if you cannot state it in 200 words, the analysis is not finished.

What is the most common mistake in the risk section?

Writing risks that would be true of any borrower. "Competitive industry" and "input price volatility" are not risks, they are conditions. A risk names the trigger, sizes the loss, and states the level at which the credit stops working.

How do you write about a guarantee you do not think is worth much?

Say so plainly, and recommend taking it anyway. A guarantee with limited realisable value still changes behaviour and preserves a claim. What damages a memo is scoring it as cover — because the one time it is tested, the file shows the analyst knew and wrote otherwise.

Do I need to show the arithmetic, or just the ratio?

Show it whenever an adjustment was made. A DSCR of 1.14x that the reader can reconstruct is worth more than a 1.74x they must take on trust, and the adjustments are exactly where reviewers and examiners will probe.

How do you handle a deviation from credit policy?

Name it in the recommendation, quantify the gap, state the approval level it requires, and say what compensating condition you have attached. Deviations buried in annexures are the single fastest way to lose a committee's trust in the rest of the memo.

Key takeaways

• Write in decision order: recommendation first, justification second, qualification third, terms last.
• Every figure gets a source. An uncited number is an assertion.
• Show the debt service arithmetic line by line, including the obligations the borrower did not volunteer.
• Quantify risks and name the threshold at which the credit fails.
• Separate real mitigants from decorative ones, and say which is which.
• Covenants need a formula, a test date, a frequency, a source and a consequence.
• Deviations belong in the recommendation, never in an annexure.

See your first CAM in 30 minutes — [book a live demo](https://yuverse.ai/yusight). Bring a file you have already written by hand and compare the citations.