How GeM Bid Evaluation Works: The Rules That Decide Who Wins

GeM bid evaluation is the set of rules the buyer's procurement officer applies after bid opening to decide which seller gets the order. Most sellers think of it as a black box — you submit, you wait, someone wins. It isn't a black box. It's a documented sequence: technical eligibility gate first, then price or weighted-score evaluation, then statutory preferences, then award. If you understand the sequence, you can usually tell who's going to win a GeM bid before the result is declared — and more importantly, figure out why your last three close losses actually lost.

This guide walks through how the GeM bid evaluation process actually plays out across Direct Purchase, Bid, and Reverse Auction. We'll cover when buyers pick L1 versus QCBS, what the scoring formulas look like with worked numbers, how MSE and Make-in-India preferences bend the math, and the preventable reasons bids get thrown out before evaluation even starts.

At Arched, we monitor 500+ procurement portals including GeM for infrastructure firms and government suppliers. Patterns in how buyers evaluate are a constant in our matching work. What follows is the practitioner view of how GeM actually decides.

The three modes of buying on GeM — and why it matters

Before evaluation rules come into play, the buyer has to pick a procurement mode. The mode determines whether evaluation happens at all, and if so, how it happens. The GeM General Terms and Conditions give buyers three standard options for catalogue-listed goods and services, plus custom bid flows for non-catalogue items.

Mode	Value limit (typical)	What the buyer does	What the seller should know
Direct Purchase	Up to Rs. 25,000	Picks any listed seller without comparison	No evaluation. Your catalogue quality and price matter, not your bid
L1 Purchase (3-seller comparison)	Rs. 25,000 to Rs. 5 lakh	Compares prices of minimum 3 sellers meeting spec	Silent competition — you don't even know you're being compared
Bid / Reverse Auction	Above Rs. 5 lakh	Runs a published bid, receives technical + financial responses	This is where evaluation rules apply in full

The thresholds get revised through GeM's General Terms, and individual departments can set lower thresholds in their own procurement manuals. The practical implication: most of what a seller experiences as "GeM evaluation" happens in the Bid or Reverse Auction mode, because everything below Rs. 5 lakh is transacted against catalogue listings with minimal evaluation logic.

Reverse Auction — the one that's frequently misread

Reverse Auctions on GeM are often bundled into "L1 buying" in seller conversations, but they're a distinct flow. After the initial technical qualification, qualifying sellers are invited to a live auction window (commonly 30 minutes, extended if bids come in near close) where they can keep dropping their price against the current L1. The final L1 at auction close wins.

A few things sellers consistently miss about reverse auctions:

• The auction only opens to technically qualified sellers. If you fail the technical gate, the final price doesn't matter.
• You see the current L1 price, not who's holding it. Bidding against a phantom is part of the design.
• Minimum decrement is set by the buyer (typically 0.5% or Rs. 1,000, whichever is higher). You can't drop by a token Re. 1 to claim L1.
• Anti-snipe extensions add 3-5 minutes every time someone bids in the final minutes. A reverse auction advertised as 30 minutes can realistically run 60-90.

If you've ever "won" a reverse auction at a price you regretted the next morning, you know why we flag this: the design of the auction creates decision pressure that a disciplined bidder has to prepare for, not improvise through.

Technical eligibility vs financial evaluation — the gate and the race

Every GeM bid above the direct-purchase threshold runs a two-stage evaluation: a technical eligibility gate followed by a financial (price) evaluation. These aren't the same thing, and confusing them is the single most common reason sellers misread their own losses.

The technical eligibility gate

The technical evaluation on GeM is binary: you either meet the stated qualification criteria for that bid or you don't. There is no partial credit at this stage. Buyers specify requirements in the ATC (Additional Terms and Conditions) section of the bid document — turnover thresholds, years of experience, similar-work experience, OEM authorisation, BIS/ISO certifications, specific product specifications.

Typical technical-gate criteria on a mid-value GeM bid look like:

• Average annual turnover (last 3 financial years) ≥ 30-50% of estimated bid value
• Similar works executed — one project at 80%, or two at 50%, or three at 40% of estimated value, within the last 7 years
• OEM authorisation letter for branded product supply
• GST registration active and compliant
• Relevant certifications — BIS licence, ISO 9001, product-specific test reports

Fail any single one that the bid marks "essential" and the system rejects your bid before evaluators even look at your price. Your competitive price doesn't get evaluated. This is why the pre-qualification section of a bid document deserves more time than the commercial offer.

The financial evaluation

Only bids that clear the technical gate move into financial evaluation. Here GeM applies one of two logics depending on what the buyer selected when publishing the bid:

• L1 (Lowest Price): Of qualified bidders, the one quoting the lowest total price wins
• QCBS (Quality and Cost Based Selection): Technical score and financial score are weighted together into a composite

Most GeM bids by volume are L1 — it's the default for goods, commoditised services, and standard specifications where "qualified" means "qualified enough" and the buyer just wants the best price. QCBS starts showing up as you move into consultancy, design work, specialised engineering, and procurement where quality materially varies across qualified bidders.

Related: QCBS vs QBS selection methods, the QCBS deep dive.

L1 evaluation on GeM — how the lowest-price math actually runs

L1 is simpler than sellers assume and also stricter than they remember in the middle of bidding.

The formula

Among the N bidders who passed the technical gate, rank by total landed price including GST, packing, freight, and any other line items the buyer asked you to quote. The lowest total wins. That's it.

Complications that trip people up:

1. Total price, not base price. If you quoted a lower base but heavier packing-and-forwarding, your landed total can exceed a bidder with higher base and nil P&F.
2. All line items considered. If the BOQ has 20 line items, the comparison is on the sum. A seller cannot strategically under-quote one line to appear cheaper overall and then expect the evaluator to ignore a different line that went high.
3. Arithmetical corrections get applied. If your unit price × quantity doesn't equal your line total, evaluators correct to unit price × quantity. This has pushed sellers from L1 to L2 more than once.
4. Statutory preferences apply AFTER L1 is identified, not as part of L1 math. We'll get to that shortly.

Worked example — L1 on a 500-unit office chair bid

Four sellers clear the technical gate for a bid estimated at Rs. 12 lakh for 500 office chairs:

Seller	Unit price	Pack & forwarding	GST (18%)	Total landed
A	Rs. 1,850	2%	18% on sub-total	Rs. 11,12,790
B	Rs. 1,790	3%	18%	Rs. 10,88,214
C	Rs. 1,820	Nil (FOR destination)	18%	Rs. 10,73,800
D	Rs. 1,760	4%	18%	Rs. 10,79,770

Naive read: D looks cheapest at Rs. 1,760/unit. Actual L1: Seller C at Rs. 10,73,800 because "FOR destination" includes freight in the unit price and carries nil P&F. Seller D loses despite the lowest quoted unit price.

This is the kind of arithmetic that decides GeM bids every day. The quoted unit price is not the scoring variable — landed total is.

QCBS on GeM — when quality actually counts toward your score

QCBS (Quality and Cost Based Selection) is the weighted-evaluation mode used when the buyer can't reduce the decision to price alone. On GeM, you'll see QCBS on:

• Consultancy and advisory services
• Design, engineering, and architectural work
• Specialised IT and software development procurement
• Some training and capacity-building bids
• Turnkey projects where technical approach varies meaningfully across bidders

The standard QCBS formula

QCBS composite score for each bidder is:

Composite Score = (Technical Score × Tw) + (Financial Score × Fw)

Where:
  Technical Score = marks awarded on a defined evaluation matrix (out of 100)
  Financial Score = (Lowest Financial Bid / This Bidder's Financial Bid) × 100
  Tw + Fw = 1.0 (commonly 0.70 technical + 0.30 financial, or 0.80 + 0.20)

The financial score inversion is the critical piece. The bidder with the lowest price gets 100 financial marks. Every other bidder gets scaled down proportionally. So if you bid Rs. 50 lakh and L1 is Rs. 40 lakh, your financial score is (40/50) × 100 = 80.

Worked example — QCBS on a consultancy bid (70:30 split)

Three bidders pass the technical gate on a project management consultancy bid:

Bidder	Technical marks (of 100)	Financial bid	Financial score	Composite (0.70 T + 0.30 F)
A	88	Rs. 48 lakh	(42/48) × 100 = 87.5	(88 × 0.7) + (87.5 × 0.3) = 87.85
B	82	Rs. 42 lakh	(42/42) × 100 = 100.0	(82 × 0.7) + (100 × 0.3) = 87.40
C	76	Rs. 44 lakh	(42/44) × 100 = 95.5	(76 × 0.7) + (95.5 × 0.3) = 81.85

Winner: Bidder A at 87.85, even though Bidder B was cheapest. This is exactly why QCBS exists — under L1, B wins on price alone; under QCBS, A's 6-mark technical edge outweighs a Rs. 6 lakh price delta.

The weights are the lever. At 50:50 the same numbers give Bidder B a win. At 80:20 Bidder A pulls further ahead. Always check the weightage in the bid document before you price — it tells you whether to optimise for quality or price.

The technical evaluation matrix — what buyers actually score

When a bid runs QCBS, the technical marks aren't assigned by gut feel. Every QCBS bid on GeM carries a published evaluation matrix that breaks the 100 technical marks across defined sub-criteria. Typical breakdowns we see on consultancy and services bids:

Sub-criterion	Typical marks	What's being measured
Firm's past experience (similar assignments)	20-30	Number, value, and relevance of past projects
Key personnel — qualifications and experience	25-35	CVs of proposed team lead, specialists
Methodology and approach	20-30	Technical proposal quality, work plan
Client feedback / past performance	10-20	GeM seller rating, reference checks
Local presence, financial strength, etc.	5-15	Varies by bid

The methodology section is the one most bidders treat as boilerplate and most evaluators read most carefully. A sharp, bid-specific methodology with a clear work plan, risk register, and deliverable schedule routinely scores 5-10 marks above a copy-paste submission. In a 70:30 QCBS, 5 technical marks is equivalent to dropping your price by ~17% — and it costs you nothing except preparation time.

Past performance rolls into technical scoring

GeM maintains a seller rating visible on your seller dashboard, based on past order fulfilment, on-time delivery, quality rejections, and buyer feedback. Most QCBS matrices now include a sub-criterion for past performance that plugs this rating directly into your technical score. Ratings below 3.5/5 start costing marks. Below 3.0/5 you start getting technically disqualified on certain ministry bids regardless of other credentials.

This is an area where a lot of sellers quietly bleed score without realising it. A single delayed delivery a year ago is still on your rating when you're bidding today.

Statutory preferences — how the math bends for MSE, Startups, and Make-in-India

GeM bid evaluation isn't a pure merit contest. Several statutory preference policies layer on after the baseline L1 or QCBS ranking, and they genuinely change who wins on a large share of bids.

MSE (Micro and Small Enterprise) preference

Under the Public Procurement Policy for MSEs, 2012, MSE sellers with valid Udyam registration get a purchase preference on bids up to defined limits:

• 25% of the tender quantity is reserved for MSEs across qualifying bids (across sub-reservations for SC/ST-owned and women-owned MSEs)
• L1 matching right: If an MSE's quoted price is within 15% of the L1 price, the MSE can match L1 and receive up to 25% of the order at the L1 price. Non-MSE L1 winner gets the remaining 75%.

Worked example: On a Rs. 1 crore bid, non-MSE Bidder A is L1 at Rs. 80 lakh. MSE Bidder B quoted Rs. 89 lakh — within 15% of L1 (89 ≤ 80 × 1.15 = 92). B gets the right to match Rs. 80 lakh and take 25% of quantity. A takes 75%.

Your Udyam registration is the gate to this preference — if it's not linked to your GeM seller account, you don't get invoked even when you qualify.

Make-in-India / Class-I, Class-II, Non-local Supplier

Under the Public Procurement (Preference to Make in India) Order, 2017 and subsequent amendments, bidders are classified by local content:

• Class-I local supplier: ≥ 50% local content — gets purchase preference on all procurement
• Class-II local supplier: 20-50% local content — gets preference on procurement above certain thresholds
• Non-local supplier: < 20% local content

In bids where Class-I preference is invoked, a Class-I bidder within 20% of L1 gets the right to match L1 and take the full order if they're willing to match. On some bids, non-local suppliers are completely excluded if sufficient Class-I capacity exists. The relevant Make-in-India category is declared in the bid document; read that section before you finalise your sourcing and bidding strategy.

Startup preference

Under DPIIT's startup recognition scheme, recognised startups can be exempted from prior-experience and prior-turnover requirements on GeM bids, subject to the buyer invoking the exemption. This doesn't change the evaluation math for startups — it changes whether they get past the technical gate at all. If you're a startup and the bid invites the exemption, use it deliberately; many buyers will disqualify a startup that ignored the exemption and submitted partial experience data anyway.

What separates winning bids from close losses — practitioner patterns

After watching a lot of GeM bid results, a few patterns hold across departments and categories:

Winners read the bid document twice before they price. The technical ATC, the preference invocations, and the evaluation weightage are all on page 4-8 of most bids. Everything downstream of bad reading of those pages is a waste.

Winners price the total landed cost, not the unit. They model GST, freight, P&F, warranty cost, and any buyer-side deductions before quoting. L1 is a total number, not a unit number.

Winners maintain a live, clean credential library. Turnover certificates from a CA, similar-work completion certificates from past clients, valid OEM authorisation letters, current BIS licences — all dated within validity, all matching the legal entity name on GeM. Losses on technical grounds are almost always an expired or mismatched document, not a missing one.

Winners compute residual bid capacity before they bid. Stretching thin across too many simultaneous obligations creates delivery failures that feed back into rating, which feeds back into the next bid's technical score.

Winners in QCBS write a bid-specific methodology. Not a template with the buyer's name swapped in. A plan that references the specific scope, sites, or deliverables in this bid.

Close losers in L1 almost always lose on a few percentage points of total cost they didn't have to concede — the MSE within 15% who didn't invoke the benefit, the bidder who quoted FOR-origin when they could have quoted FOR-destination, the 2% P&F that tipped them past a rival. Tight bid math is non-optional on GeM.

Common preventable disqualifications

These are the categories of loss that don't have to happen:

• Expired document uploads — BIS licence expired last month, turnover certificate from FY22 when FY23 was asked
• Entity name mismatch — CA certificate issued in proprietor's name when seller is registered on GeM as a Pvt Ltd
• Missing or wrong digital signature — bid signed with a DSC that isn't the authorised signatory's. The digital signature certificate has to match the entity
• BOQ structure not followed — quoting a line item in a different UoM than the buyer specified, or skipping a line
• GST not reconciled — GSTN declared as active but returns not filed for 2+ quarters
• No bid management checklist — submitting in the last 15 minutes, missing an upload the system didn't flag until after close

None of these are strategy problems. They're operating-discipline problems, and they're the cheapest ones to fix.

Where this leaves your GeM bidding strategy

Understanding GeM bid evaluation changes the work. It's no longer about hoping you're cheap enough — it's about reading each bid against the evaluation framework, identifying whether the game is L1 or QCBS, knowing which preferences are invoked, and deciding whether this bid is actually winnable at a price you can deliver at.

The operational version of this is:

1. Filter bids to ones your technical credentials actually clear — no wasted time on bids you'd fail the gate on
2. On L1 bids, model total landed cost and check against your walk-away margin before you commit
3. On QCBS bids, work the technical score first — every mark is worth more than the equivalent price reduction up to the weightage break-even
4. Always check MSE / Make-in-India invocations and invoke them where you qualify
5. Keep your seller rating above 4.0/5 — it's compounding interest on future bids

The manual version of running this across every GeM bid plus every state portal is the work that kills BD team productivity. This is exactly the problem we built Arched's bid intelligence platform for — it filters 500+ portals against your firm's actual credentials, flags bids where your technical gate is clean, and shows you the evaluation structure so you can walk in knowing whether the bid is an L1 price fight or a QCBS quality play before you spend a minute writing.

See our take on how bid management software compares for contractors managing GeM alongside other portals, and how contract models like EPC, HAM, BOT, and TOT interact with evaluation structures on larger infrastructure tenders.

Frequently asked questions

Does GeM always use L1 for bid evaluation?

No. L1 is the default for goods and standard services, but QCBS (70:30 or 80:20 weightage between technical and financial) is used for consultancy, design, and specialised services. The evaluation mode is declared in the bid document under "Evaluation Method" — always check this before pricing, because it changes whether you should optimise for lowest price or for technical score.

What's the difference between L1 and the lowest quoted unit price?

L1 is the lowest total landed cost including GST, packing and forwarding, freight, and all line-item totals summed across the BOQ. A bidder with the lowest quoted unit price can end up L2 or L3 if their packing, freight, or other line items push the total above another bidder's landed cost. Always compute the total, not the headline unit rate.

Does MSE preference automatically apply or do I need to invoke it?

Your Udyam registration must be valid and linked to your GeM seller profile for MSE preference to apply. On bids that invoke MSE purchase preference, a qualifying MSE within 15% of L1 gets the right to match L1 and take up to 25% of the quantity. If your Udyam isn't linked or is expired, the system won't invoke the preference even if you'd otherwise qualify — this is one of the most common unforced errors on GeM.

How does seller rating affect my GeM bid evaluation?

Seller rating feeds into the technical scoring matrix on QCBS bids (typically 10-20 marks out of 100) and can trigger technical disqualification below 3.0/5 on some ministry bids. It's cumulative from past order fulfilment, on-time delivery, quality rejections, and buyer feedback. A single delayed delivery or rejected consignment lingers in your rating for months and affects every bid you submit in that window.

Can a buyer reject all bids after evaluation?

Yes. GeM allows buyers to cancel or re-tender a bid for documented reasons — technical specifications needing revision, all bids above estimate, insufficient qualified bidders, or policy-level changes. This can't be used arbitrarily — it requires internal approval and a published reason — but it does happen, particularly on high-value or contested bids. If you see a bid re-tendered, read the revised specification carefully; the rules may have shifted in a way that changes who qualifies.

When is reverse auction used instead of a standard bid?

GeM buyers can opt for reverse auction after technical qualification to drive the final price down through live competitive bidding. It's most common on commoditised goods, high-volume standard services, and bids with 5+ qualified sellers where the buyer expects real price compression. You'll know at bid publication whether the bid flows into reverse auction — the bid document flags it in the commercial evaluation section.

Does Make-in-India preference override L1?

It can. Under the Public Procurement (Preference to Make in India) Order, a Class-I local supplier within 20% of L1 gets the right to match L1 and take the full order. On certain notified product categories, non-local suppliers are excluded entirely if sufficient Class-I capacity is available. Check the Make-in-India section of every bid document — it's where this gets invoked, and ignoring it has cost plenty of otherwise-cheaper bidders the order.

How do I know if a GeM bid is L1 or QCBS before I invest time?

The bid document under "Evaluation Method" or "Selection Criteria" explicitly states whether evaluation is L1, QCBS, or another method. For QCBS, the technical-financial weightage (e.g., 70:30, 80:20) and the technical evaluation matrix are also published up front. If you can't find this section, you're looking at the wrong document — ATC and evaluation method are always in the main bid document, not a supporting attachment.