Preparing an RFQ Package That Gets You Comparable Bids

Introduction

Every procurement engineer has lived this meeting: three bids on the table for the same package, spread across a factor of 1.4 in price. The cheapest bidder has excluded instruments, first fills, and factory testing. The most expensive one has included two years of spares nobody asked for. The middle bid is silent on delivery terms. The meeting runs for three hours and produces a list of clarification questions instead of a recommendation.

That outcome was decided weeks earlier, when the Request for Quotation went out. Bids diverge when the RFQ lets them diverge: a vague scope invites each bidder to price different assumptions; missing data gets replaced with qualifications; and without a mandatory structure for pricing and compliance, every bid arrives in its own format, measuring its own scope, on its own terms.

A good RFQ package is not paperwork. It is the instrument that converts a commercial contest into an engineering comparison. This post covers what goes into one, and the mechanics that make bids comparable by construction.

What Belongs in the Package

For an upstream equipment package — a separation skid, a dehydration unit, a well test package — the RFQ should carry five distinct parts:

1. Scope of supply matrix. A table, not prose: every element of the supply, with a column for vendor and a column for purchaser. Skid steelwork, instruments, first fills, special tools, spares, surface protection, transport, offloading, site support during commissioning, documentation. Ambiguity in this matrix becomes either a double-count or a gap — and the gap is always discovered on site.

2. Technical requisition. The engineering definition the bidder prices against: process datasheets and design cases, the design basis extract (feed envelope, utilities available at the battery limit, site and environmental conditions), the applicable specifications and standards — with editions stated — relevant PFD and P&ID extracts, footprint and weight limits where they bite (offshore, they always bite), electrical area classification, and painting and preservation requirements. The maturity of this section sets the ceiling on bid quality: bidders cannot price definition that does not exist, and an RFQ issued from a half-frozen design guarantees revision churn mid-tender.

3. Commercial and contractual terms. Delivery terms as Incoterms with a named edition and place, required schedule expressed against milestones, warranty, liquidated damages basis, payment milestones, and bid validity period long enough to survive the evaluation you actually intend to run.

4. Instructions to bidders. Submission format, deadline, the clarification protocol — questions in writing by a stated cut-off, answers circulated to all bidders — and the rule that qualifications are only recognised in the exceptions register, not scattered through cover letters.

5. Evaluation criteria. State how the bids will be judged — the split between technical compliance, price, and schedule — before the bids arrive. Criteria invented after opening are a governance finding waiting to happen.

The Comparability Mechanics

Comparability does not come from asking nicely. It comes from three mandatory structures:

• A clause-by-clause compliance statement. Every requirement in the requisition gets a response: comply, or deviate with a description. Silence defaults to comply and is contractually held there. This converts a 40-page proposal into an exceptions list you can price.
• A mandatory pricing breakdown. Not a single lump sum: engineering, materials, fabrication, testing, documentation, spares, and options as separate line items in a supplied template, single stated currency, price basis date, escalation basis fixed. The breakdown is what lets you move scope between bids during levelling — and it is also your first cost model for the next project.
• A single exceptions register. One table per bid holding every deviation, exclusion, and alternative offered. If it is not in the register, it is not part of the bid.

Levelling: A Worked Example

Three bids arrive for a glycol dehydration skid:

Bid A:  $1.00M — excludes instruments, first fill, and FAT
Bid B:  $1.18M — fully inclusive, 26-week delivery
Bid C:  $1.32M — fully inclusive, 20-week delivery

Bid A looks 15% cheaper than B. Level it: price A's exclusions from the breakdown data in the other bids — instruments $0.12M, first fill $0.03M, factory acceptance testing $0.05M:

Bid A levelled:  1.00 + 0.12 + 0.03 + 0.05  =  $1.20M

And purchaser-supplied instruments on a vendor skid are not free even after buying them — allow $0.05M for the purchaser-side specification, procurement, and integration engineering the exclusion transfers to you:

Bid A levelled + interface allowance:  1.20 + 0.05  =  $1.25M

Now level schedule. If the facility earns $15,000 per week of earlier availability, C's six-week advantage over B is worth $0.09M:

Bid C schedule-adjusted:  1.32 − 0.09  =  $1.23M   vs   Bid B:  $1.18M

On a levelled basis the ranking is B ($1.18M), then C ($1.23M), then A ($1.25M) — the reverse of the raw prices. That inversion is not unusual; it is the normal result of levelling done properly. The recommendation writes itself, and it survives audit because every adjustment traces to a bid document or a stated allowance.

Technical Bid Evaluation

Run the technical evaluation before commercial preferences contaminate it — on larger tenders, formally, with a two-envelope process. The sequence that works:

1. Screen for showstoppers — non-compliances that no price can fix (material selection outside the sour-service envelope, footprint that does not fit, unproven scale-up). Kill early, in writing.
2. Normalise the scope technically — identify each bid's deviations and gaps from the compliance statement and exceptions register, so every technical difference is named before any price is known.
3. Clarify in writing only. Verbal answers do not exist. Significant clarifications go to all bidders.
4. Open commercial and level the acceptable bids — the dollar levelling above, applied only to bids that survived the technical gate — then evaluate against the published criteria and record the basis. The tender file should let a reader reconstruct the decision two years later.

Common Failures

• Issuing the RFQ from an immature design. Every revision issued mid-tender invites every bidder to reprice everything. If the engineering definition is not stable enough to hold for the tender period, the package is not ready to tender.
• Ambiguous battery limits. "Skid complete with all necessary items" is a dispute, not a scope.
• Schedule as a duration from an undefined start. Tie deliveries to purchase-order date or drawing approvals — dates a contract can enforce.
• "Or equal" without an adjudication rule. State who judges equality and against which datasheet parameters.
• Letting the vendor list do the evaluation's job. An over-restricted bidder list produces single-bid outcomes and removes your negotiating position; an unqualified list wastes everyone's time. Pre-qualify on capability, then let the RFQ discriminate.
• Treating the estimate and the tender as strangers. The pricing breakdown you mandate is the same structure your cost estimate used — reconcile them, and the next estimate gets better.

Conclusion

An RFQ package earns its keep twice: once at evaluation, when three bids land in the same structure and the comparison takes days instead of months, and again at execution, when the scope matrix and exceptions register have already answered the arguments before they start.

The discipline is front-loaded. A stable technical definition, an explicit scope matrix, mandatory pricing and compliance structures, and published evaluation criteria — none of it is glamorous, and all of it is cheaper than discovering, six months after award, what the winning bidder actually meant by "included".