Comparative Statement (Quotation Comparison) Format for Construction Procurement in India

Every construction company has a version of this story. Three quotations arrive for 8 MT of TMT steel. Vendor B's basic rate is Rs 300 per tonne cheaper than Vendor A's, so the purchase order goes to Vendor B. Then the material lands at site with a freight bill nobody priced in, the invoice shows GST on a higher assessable value than anyone checked, and the "cheaper" vendor turns out to have cost Rs 1,600 more per tonne, delivered. Multiply that by every cement, steel, aggregate and electrical purchase across a project, and the leak is larger than the margin on most line items.

The document that stops this is the comparative statement — the CS. It is one sheet of paper (or one Excel tab) that sits between the material indent and the purchase order, and it does exactly one job: it puts every vendor's offer on identical footing, computes the true delivered cost of each, and records why the order went where it went. Public works departments have required it for decades. Most private contractors either skip it or fill it in after the PO is already decided, which is worse than skipping it, because a backdated CS is the first thing an auditor learns to smell.

This guide covers the format column by column, the landed-cost arithmetic that makes a CS worth preparing, the L1 evaluation rules that public procurement has already debugged for you, the frauds a CS catches and — just as important — the frauds it structurally cannot catch, and the surprising number of statutory contexts where a dated CS is the document that saves you: income-tax scrutiny of related-party purchases, bogus-purchase proceedings, GST input tax credit disputes, and RERA project audits. A free Comparative Statement (CS) Format in Excel — the exact sheet described below — is in the templates library.

What a comparative statement is#

A comparative statement is a tabular comparison of all quotations received against one enquiry, prepared before the purchase decision, showing for each vendor the complete delivered cost of each item plus the commercial terms, ending in a ranking (L1, L2, L3 — lowest to highest) and a signed recommendation.

The format comes from Indian public procurement. Government purchase manuals require every tender evaluation to pass through a comparative statement checked by someone other than its preparer, and the General Financial Rules require open or limited competitive bidding for purchases above small thresholds, with a Local Purchase Committee certifying reasonableness of rates in the middle band. The exact rupee thresholds have been revised repeatedly — what matters for a private company is not the numbers but the architecture: below a small value, buy directly; in the middle band, at least three quotations and a committee; above it, formal tendering. Almost every construction company that adopts a procurement policy ends up rebuilding the same three tiers, because the alternative at each tier has already been shown to fail.

The Central Vigilance Commission, which has spent five decades cataloguing how public tenders go wrong, has left two observations that every private CS should be built around. First, on what happens when rates are compared against nothing: works are awarded "without preparing any market rate justification", with comparison made against works awarded years earlier — a procedure the CVC calls neither objective nor appropriate, because justification "should be based on realistic prevailing rates". Second, on what happens when the comparison basis is left vague: the CVC has described a tender where quantities were never specified in the bid documents, leaving the evaluation committee to compare bids "on surmises and conjectures". A CS built on unclear quantities or missing cost heads is arithmetic without meaning.

One more piece of procurement architecture worth copying: the anti-splitting rule. Government rules prohibit dividing a requirement into smaller lots to bring each below the threshold that would trigger quotations or higher-level sanction — and CVC inspection reports record exactly this being done, orders "split in order to bring it within the powers of junior officers". In a private company the same fraud looks like four POs of Rs 45,000 each to the same vendor in the same week, each below the Rs 50,000 mark where your policy requires three quotations. The fix is a policy line: aggregation is by requirement, not by PO — if the month's need for an item crosses the threshold, the CS requirement applies to the whole of it.

Why basic-rate comparison is wrong: the landed cost principle#

The single most common CS mistake is ranking vendors on basic rate. Basic rate is the beginning of a price, not the price. The number that belongs in the comparison is the landed rate — what one unit of material actually costs by the time it is usable at your site:

Landed rate per unit = (basic rate − discount + packing and forwarding + freight to site + loading and unloading + transit insurance) + GST on the taxable value

Every element in that bracket varies between vendors, and vendors know buyers compare basic rates — which is precisely why quotes are structured to look cheap at the top line. An ex-works quote shifts freight to you. A "freight extra at actuals" line is an open cheque. A vendor quoting FOR-site (freight paid, delivered) at a slightly higher basic rate is routinely cheaper, delivered, than the low basic rate that hasn't priced the truck.

There is a harder reason to compare landed totals, and it comes from a real cartel. In a 2012 Indian Railways tender for feed valves, three manufacturers — in Haryana, Tamil Nadu and West Bengal, with entirely different cost structures — quoted basic prices of Rs 17,147.54, Rs 14,534.52 and Rs 14,674.28. An 18 per cent spread; a healthy-looking competition. But each firm had worked backwards through different excise, cess and CST figures so that all three landed at a total unit price of exactly Rs 17,147.54 — identical to the paisa, and about a third above the last purchase rate. The Competition Commission of India penalised all three, and the CAG's later analysis of the case made the point that matters for CS design: a comparison of basic rates showed nothing, and only the delivered totals exposed the arrangement. The tender committee that caught it caught it because their comparative statement computed the landed figure. A CS that compares basic rates is blind to the exact technique.

Getting the GST column right#

GST is the largest single add-on in the landed computation, and the rates moved substantially in the September 2025 rate rationalisation — cement dropped from 28 per cent to 18 per cent with effect from 22 September 2025, which by itself changed the landed mathematics of every concrete-heavy comparison. The working set for a CS:

The full HSN-wise table, the works-contract rates and the input tax credit rules are in our GST on construction materials guide. Three CS-specific rules: compute GST on the taxable value after discount but including freight where the supplier bills freight (a composite supply takes the rate of the principal goods); confirm every vendor has quoted the same basis (inclusive quotes hide the assessable value and must be broken open before comparison); and where your entity cannot claim input tax credit on the purchase — a developer building for sale is largely blocked from ITC on construction inputs, while a contractor supplying works contract services generally is not — the GST is a real cost and belongs inside the comparison, not below it.

The credit-period column#

A 30-day credit period is worth money — at 12 per cent annual working-capital cost, roughly 1 per cent of the invoice. A vendor at Rs 100 with 30 days' credit beats a vendor at Rs 99.5 against advance payment. The CS format below carries payment terms in the terms matrix rather than converting them to rupees, because the conversion assumptions (your actual cost of funds, whether you would really pay on day 30) are argument-prone. But when two landed rates are within about 1 per cent of each other, the credit terms are the tiebreak, and the recommendation note should say so explicitly.

The format, column by column#

The downloadable sheet has four zones. This structure is deliberately close to what a government tender evaluation uses, because that format has survived decades of audit.

Zone 1 — header. CS number and date, project name, indent reference, enquiry/RFQ reference with the date quotations were invited, the due date, and the number of quotations received against invited. The indent reference matters more than it looks: it proves the requirement existed before the quotes did. A CS with no indent behind it is a purchase looking for a justification. (If your indent discipline is weak, fix that first — our material indent management guide covers the workflow.)

Zone 2 — the rate comparison. One row per item, with the specification written in full — "TMT Fe 500D, 12 mm, IS 1786" — not "steel". Then, for each vendor, a column block: basic rate, discount, packing and forwarding, freight (with the term stated: ex-works, FOR site, freight extra), loading and unloading, taxable value, GST rate and amount, landed rate per unit, and the extended total for the indent quantity. Where a vendor has quoted ex-works, you insert a realistic freight estimate into their column and mark it as your estimate — the comparison must be on identical delivery terms even when the quotes were not.

Zone 3 — the terms matrix. One row per commercial term, one column per vendor: payment terms, delivery period, quote validity, warranty or test-certificate commitment, and brand or make offered. This zone is where the "cheapest" vendor's quote usually falls apart — a 21-day delivery on a slab-casting item, or a validity that expires before your approval cycle finishes, is a cost the rate columns never see.

Zone 4 — ranking and sign-off. L1, L2, L3 on the landed extended total; the recommendation with its reason; and three signatures — prepared by, checked by, approved by. If the recommendation is anyone other than L1, the reason is written here, in words, at the time. Not because a rule forces a private company to buy from L1 — no law does — but because an undocumented deviation from L1 is indistinguishable, two years later, from a kickback. The person who checks the CS should not be the person who prepared it, and the person who approves should be neither; this separation of preparer, checker and approver is the core of what an auditor calls segregation of duties, and it is the cheapest fraud control that exists.

A worked example, from the template's sample rows. Vendor A quotes TMT at Rs 54,200 per MT less 2 per cent, freight Rs 850 and loading Rs 150 per MT: taxable value Rs 54,116, landed Rs 63,857 with GST. Vendor B quotes Rs 53,900 net — Rs 300 per MT cheaper on the face — but ex-works, and the realistic freight for the lot is Rs 1,600 per MT: taxable value Rs 55,500, landed Rs 65,490. Vendor A is L1 by Rs 1,633 per MT despite the higher basic rate. On the 8 MT indent that is Rs 13,000 — found by one Excel formula.

The L1 rules public procurement has already debugged#

Private companies are free to buy from anyone. But four rules from the government side are worth adopting wholesale, because each one closes a hole that someone has already exploited.

Rule 1 — negotiate with L1 only. The CVC's standing instruction since 1998, restated in its circular of 3 March 2007, is that post-tender negotiation is banned except with L1, and even that only in exceptional situations — proprietary items, limited supply sources, or suspicion of a cartel. The reasoning transfers perfectly to private buying: the moment vendors learn that L2 can win by matching a price after quotes open, every vendor's first quote becomes a placeholder, and your quotation process stops producing real prices. Asking all vendors to submit "revised best offers" after seeing where they stand — a common private-sector habit — is precisely the practice the CVC calls a violation on the government side, and it corrodes a private company's quote quality the same way.

Rule 2 — if L1 backs out, re-enquire; don't walk up the ladder. Quietly awarding to L2 at L2's rate rewards a vendor pair where one quotes low with no intention to supply and the other collects the order. The CVC's answer is a re-tender. Yours can be a fast re-enquiry — but the principle is the same.

Rule 3 — if you split the quantity, pre-disclose the ratio. Splitting an order between L1 and L2 (with L2 matching L1's rate) is legitimate for delivery security on critical materials — but only if the enquiry said so in advance. A split invented after quotes open is an invitation for vendors to coordinate, and the CCI's procurement guidance warns that routine splitting removes the incentive to compete at all, because everyone is assured a share.

Rule 4 — three quotations minimum, and no splitting below the threshold. Three is the floor at which a comparison means anything. One quotation is a price dictated; two is a coin toss; three begins to be a market. Where a genuine market has fewer suppliers — a proprietary admixture, a sole authorised dealer — the CS is replaced by a single-source justification note stating why no comparison was possible, signed at the same approval level. What is never acceptable is requirement-splitting to duck below the quotation threshold: it is prohibited in the government rules for exactly the reason your policy should prohibit it.

What a CS catches — and what it cannot#

A comparative statement is a strong control against unilateral overpricing, arithmetic errors, and non-comparable quotes. It is worth being honest about the rest, because the most expensive procurement frauds in the Indian record are the ones engineered to pass through a CS looking clean.

What it catches. Freight and GST games between vendors. Quotes that skip a cost head. A single vendor drifting above market across months — if you file CS sheets by material, the trend is visible. Naive identical pricing: quotes matching to the rupee are, in the CCI's own words to procurement officers, "a very strong indicator towards a possible collusion", and a CS computing landed totals is exactly where identical totals surface.

Ghost quotations. The oldest trick: two of the three quotes are procured by the same person — a cooperative vendor supplies "competitor" letterheads, or the site engineer fills all three himself. On paper, a competitive purchase; in fact, a dictated price plus a paper trail. The scale this reaches is documented: a CAG audit of Chhattisgarh's e-procurement system found 74 common computers used by bidders and government officials across 1,921 tenders worth Rs 4,601 crore, and 79 vendors operating with two PANs. Every one of those tenders had a comparison sheet that looked competitive. The CS cannot see this — but the metadata can, and the CCI's October 2024 Diagnostic Toolkit for procurement officers lists the tells: bids from the same IP address, the same email ID, identical spelling mistakes across "competing" quotes, the same handwriting, sequential demand drafts from the same bank branch, one bidder's letterhead carrying another's fax number. The private-company translation: insist on quotes emailed directly from the vendor's domain, call the vendor's listed landline once to confirm the quote exists, check two GSTINs are not registered at the same address, and rotate who invites the quotes. Ten minutes of verification defeats most ghost-quote schemes, because they are built on the assumption nobody checks.

Cover pricing. The designated winner asks friendly competitors to submit deliberately high quotes — the appearance of competition concealing an inflated price. The CAG's research paper on procurement collusion calls complementary bidding the most frequently occurring form of bid rigging, and records a common pattern of the cover price being set roughly 10 per cent above the intended winner's. The CS-level defence is the one the CVC keeps repeating: an independent benchmark. A CS that only compares the three quotes against each other can be satisfied by three coordinated quotes; a CS with a last-purchase-rate column and a current market-rate estimate cannot. That is also the bridge between the CS and rate analysis — the estimate side of the same discipline.

What no CS can catch. Quantity manipulation: the CS governs the rate; if site measurement books over-record quantities, a flawless L1 award leaks money anyway — a 2025 CAG audit in Karnataka found Rs 1.48 crore of excess payment across 13 divisions from a single repeated measurement error, with the rate never in dispute. Post-award change orders: award at a genuinely low price, then recover it through a series of variations — a scheme that requires a clean CS as its entry ticket. Kickbacks structured as a percentage of every bill cleared, which attach at the payment stage the CS never touches. The lesson is not that the CS is weak; it is that the CS is one link, and it only holds as part of the chain: indent → CS → purchase order → GRN at site → invoice matched against both → monthly reconciliation. Auditors call the PO-GRN-invoice check the three-way match, and it is the control that picks up where the CS stops.

The fraud-survey numbers say this chain deserves more attention than it gets: PwC's 2024 economic crime survey found procurement fraud the most-reported economic crime among Indian organisations, and the ACFE's 2024 Report to the Nations puts construction's median fraud loss at USD 250,000 per case — fourth-highest of any industry — with corruption schemes (bribery, kickbacks, conflicts of interest) featuring in 52 per cent of construction cases, the highest share of any scheme type.

Where the CS saves you with the tax man, the auditor and RERA#

This is the part most contractors discover too late: the comparative statement is not just a cost control. It is admissible, contemporaneous evidence in at least four proceedings where the burden sits on you.

Income tax — related-party purchases. Section 40A(2) of the Income-tax Act lets the Assessing Officer disallow expenditure he considers excessive against fair market value where the payment goes to a specified person — a director's relative, a sister concern, any entity where 20 per cent ownership crosses over. Construction groups run on exactly such entities: the promoter's trading arm supplying steel, the brother-in-law's transport company. The case law is buyer-friendly but evidence-hungry: the Supreme Court has held the section cannot apply unless excessiveness is first established as a fact, and tribunals have deleted disallowances precisely because the AO "had not brought on record any comparable case" showing the payment exceeded market rate. Read that from the other side: a dated CS showing the related party's landed rate at or below the lowest independent quote is the comparable case, produced by you, on the day of purchase. CBDT's own 1968 circular on the section names purchases of raw materials and stores as squarely covered and directs officers to judge reasonableness fairly — a contemporaneous CS is what makes "fair" easy to reach. The auditor's guidance under CARO runs the same direction: a company claiming its related-party purchases are at arm's length must hold documentary proof at the time of entering the transaction, not reconstruct it at year-end.

Income tax — bogus purchase proceedings. Where a supplier appears in a hawala list or vanishes when the department issues notices, the onus of proving the purchase genuine falls on the buyer. The recent tribunal record shows exactly which files survive: a listed infrastructure company defeated bogus-purchase additions in 2025 because it produced purchase orders, e-way bills, goods receipt notes, weight slips, ledgers and bank statements — and the taxpayers who lose are the ones with invoices and cheque payments but no delivery evidence, no vehicle numbers, no GRN. The CS sits at the head of that winning document set: it demonstrates a real procurement process selected the vendor, which is the opposite of what an accommodation-entry file looks like.

GST — input tax credit. Section 155 of the CGST Act puts the burden of proving ITC eligibility on the person claiming it, and the department's standing scrutiny instructions target credit taken on invoices from suppliers who never filed returns or turned out not to exist. The courts that have protected buyers have protected diligent buyers — ones who verified registration, held proper invoices, paid through banking channels and could show receipt of goods. A CS plus GSTIN verification at vendor-selection, plus the GRN proving receipt, is that diligence in documentary form; it is also direct evidence against the collusion allegation that converts an ITC dispute into a penalty case.

RERA — the 70 per cent account. Promoters withdraw from the designated project account against certificates of cost incurred, and the CA's certificate is capped by what the books support. Haryana RERA's audit directions spell out the risk in one sentence: if the project cost incurred is inflated at the time a certificate is issued, the percentage of completion is certified too high, and the promoter withdraws more than the rules permit. Inflated or unsupported purchase costs are the easiest way that happens — and the CS-PO-GRN chain is what lets the certifying CA, and the annual RERA auditor behind him, tie every cost claim to a real, market-priced purchase.

Statutory audit. Auditors reporting on internal financial controls treat the procurement cycle as a standard test area, and the controls they test for are the ones this guide describes: a documented policy, an approval matrix, segregation of preparer-checker-approver, and the three-way match. A CS regime is not gold-plating; it is the textbook design of an adequate control, and its absence in a company above the audit thresholds is the kind of gap that surfaces in management letters first and qualified reports later.

The workflow around the sheet#

The CS works when it sits inside a fixed sequence — the same sequence covered end-to-end in our construction procurement process guide:

1. Indent from site, approved, with required-by date — the demand evidence.
2. Enquiry/RFQ to at least three vendors, identical specification and delivery terms to all, with a quote deadline. In writing — email is fine; WhatsApp quotes get re-confirmed on email.
3. Quotations in, logged with received dates. Late quotes are recorded as late.
4. CS prepared by purchase, checked by someone independent, within the quotes' validity.
5. Approval per the delegation matrix — site engineer to a limit, project manager above it, director above that. The approval is on the CS itself, not on a verbal summary of it.
6. PO released to the selected vendor, referencing the CS number, at the CS rate and terms.
7. GRN at delivery, quantity and quality checked at site, feeding the three-way match before the invoice is paid.

Two operating disciplines keep the system honest. First, timestamps: the indent predates the enquiry, the quotes predate the CS, the CS predates the PO. Any file where those run backwards was decided first and papered later. Second, filing: CS sheets filed by material as well as by project, so that rate trends across months are visible — that cross-file view is where a slowly drifting cartel or a favoured vendor's creep shows up.

In Excel this runs adequately up to a point — one project, one purchaser, modest volume. It strains where companies grow: version chaos, approval by phone call, the CS-to-PO rate link breaking silently, and no cross-project rate memory. That last one matters most: the strongest single benchmark for any new CS is your own landed rate for the same item last month on another site. Procurement software holds the indent, enquiry, CS, approval and PO in one chain with the rate history attached — SiteSetu's procurement module is built around exactly this indent-to-PO flow, with the CS stage enforced before a PO can issue. But the discipline is the point, not the tool; a company that runs a clean Excel CS regime is ninety per cent of the way there.

Using the free template#

The Comparative Statement (CS) Format in the templates library is a ready Excel implementation of everything above: header zone with indent and RFQ references, per-vendor landed-cost columns with the GST computation built in, the commercial terms matrix, L1/L2/L3 ranking, a written recommendation block, and the three-signature strip. The sample rows carry the TMT example from this guide so the arithmetic is visible before you overwrite it with your own enquiry. Pair it with the indent, purchase order and GRN formats from the same library and the whole documentary chain this guide describes is in place by the end of the week — at zero software cost.

FAQs#

What is a comparative statement in construction procurement? A comparative statement (CS) is a tabular comparison of all vendor quotations received against one enquiry, prepared before the purchase decision. It computes each vendor's landed rate — basic rate plus packing, freight, loading and GST — on identical delivery terms, compares commercial terms like payment and delivery period, ranks vendors L1/L2/L3 on the landed total, and records the signed recommendation and approval.

How many quotations are required for a comparative statement? Three is the accepted minimum for the comparison to mean anything, and it is the floor built into government purchase rules that most company policies mirror. For proprietary or single-source items where three quotes are genuinely impossible, replace the CS with a signed single-source justification note approved at the same level.

What is landed rate and how is it calculated? Landed rate is the true delivered cost of one unit of material at site: basic rate minus discount, plus packing and forwarding, freight to site, loading/unloading and transit insurance, plus GST on that taxable value. Two vendors must always be compared on landed rate under identical delivery terms — a low basic rate quoted ex-works is routinely costlier, delivered, than a higher basic rate quoted FOR site.

What does L1 mean, and is a company required to buy from L1? L1 is the vendor with the lowest landed total; L2 and L3 follow. No law compels a private company to buy from L1. But any award to a vendor other than L1 should carry a written, contemporaneous reason — delivery capability, quality history, approved brand — on the CS itself, because an undocumented L1 deviation is indefensible in front of any auditor, tax officer or investor later.

Can we negotiate with L2 or ask all vendors for revised offers? Best practice — mandatory on the government side under CVC instructions — is to negotiate with L1 only, and sparingly. Asking all vendors for revised offers after quotes open teaches the market that first quotes are placeholders, and every future enquiry gets worse prices. If L1 refuses to supply, re-issue the enquiry rather than walking up to L2 at L2's rate.

What if vendors quote identical or near-identical rates? Identical totals to the rupee are a strong red flag for coordination — the CCI's guidance to procurement officers treats identical quotes as a leading collusion indicator. Do not simply split the order among them. Check the quotes for shared tells (same errors, same formatting, sequential DDs), bring in quotations from outside the usual circle, benchmark against your last purchase rate, and if the pattern persists on significant value, take advice — bid rigging is presumed anti-competitive under Section 3(3) of the Competition Act.

Is a comparative statement legally required for private companies? No statute mandates a CS for private purchases. It becomes commercially mandatory through the back door: income-tax officers scrutinising related-party purchase rates, ITC disputes putting the burden of proof on the buyer, bogus-purchase proceedings demanding evidence of a real procurement process, RERA auditors testing certified project costs, and statutory auditors testing procurement controls. In each, a dated CS is the document that answers the question.

What is the difference between a comparative statement and rate analysis? Rate analysis builds up what an item of work should cost from materials, labour, machinery and overheads — it produces the estimate. A comparative statement compares what vendors are actually offering against each other, ideally with that estimate as a benchmark column. Rate analysis sets the yardstick; the CS holds the market against it.

Who should sign a comparative statement? Three different people: the preparer (usually purchase or the site engineer who floated the enquiry), a checker who verifies rates against the quotes and re-runs the arithmetic, and the approver per the company's delegation matrix. Preparer, checker and approver should never be the same person — that segregation is the control auditors test first in any procurement cycle.

How long should comparative statements be kept? Keep the CS with its quotations, indent, PO, GRN and invoice as one purchase file for at least eight years, matching the books-of-account retention under the Companies Act and comfortably covering income-tax reassessment windows and GST adjudication timelines. For RERA-registered projects, retain procurement files for the project's full audit trail — the certified cost of the project is only as defensible as the purchase files behind it.