A manufacturing bid/no-bid decision should first rule out work the business cannot responsibly deliver, then compare feasible RFQs by expected contribution, estimating effort and constrained production capacity. Revenue size and an enthusiastic sales note are insufficient. AI can organize the RFQ and surface missing evidence; qualified commercial and technical owners must make the commitment.
This guide focuses on allocating scarce estimating attention before detailed quotation. The proposed decision framework gives engineering and commercial reviewers a shared basis for choosing what to estimate. All scored opportunities and calculations are synthetic. The aim is to spend quoting time where the business has a credible route to winning useful work.
Why a large quotation pipeline can hide weak demand
An RFQ list may contain funded requirements, budget exercises, approved-vendor checks, competitive price comparisons and speculative future projects. Treating them all as equally valuable fills the estimator's queue without creating equally valuable opportunities.
The hidden cost is displacement. A complex low-probability request can occupy the hours needed to answer a simpler repeat opportunity on time. Once production is constrained, the wrong win can also displace work with better contribution per bottleneck hour. Quotation throughput alone therefore says little about commercial quality.
Classify the buying situation from evidence: named project, required date, drawing maturity, decision process and customer relationship. Unknown is an acceptable classification. Inventing a probability because a CRM requires a number makes the pipeline look more precise without improving the decision.
Check whether the required part and process fit your capabilities.
Check capacity, dependencies and the customer's delivery requirement.
Assess terms and commercial fit before committing estimating time.
Read the visual explanation
A practical sequence for this decision. Apply your actual records, responsibilities and acceptance criteria.
Apply technical and commercial gates before scoring
Use gates for conditions that should not be compensated by a high score elsewhere. A missing required capability does not become acceptable because the order value is attractive. Each gate needs an owner who can determine whether it passes, fails or remains unresolved.
| Gate | Pass evidence | If unresolved |
|---|---|---|
| Process and material fit | Qualified process owner confirms scope | Hold detailed estimate |
| Drawing and specification maturity | Identified revision and requirements | Request clarification or price stated assumptions |
| Required quality evidence | Quality owner confirms applicable scope | Identify qualification work before commitment |
| Delivery feasibility | Planning checks constrained resources and supply | Propose a feasible date or decline |
| Commercial terms | Authorized review of payment and exposure | Escalate terms rather than burying them in price |
| Customer intent | Evidence of timing and buying process | Limit estimating effort until clearer |
This is an internal decision tool, not a universal compliance standard. Requirements vary by product, contract and sector. A regulated or safety-critical part needs appropriately qualified review; an AI extraction of a specification is not evidence that the business can manufacture it.
Technical fit
Capability, drawings, tolerances and required evidence.
Delivery fit
Capacity, material availability and a credible schedule.
Commercial fit
Cost, terms, risk and the value of scarce estimating time.
Conceptual guide by Looqus Editorial. Read the surrounding analysis for assumptions, exceptions and primary sources.
Use capacity evidence, not a free-space impression
Finite capacity planning considers limits on resources and can move a delivery date when capacity is unavailable. Microsoft documents this behavior and the configuration needed for its own planning system. The general operating implication is to check the relevant bottleneck and time window, rather than assuming the factory has capacity because total utilization looks moderate. Microsoft finite capacity planning.
An RFQ may fit machining capacity but fail heat treatment, inspection or a specific operator qualification. Identify the resources that constrain the requested delivery. Include setup and changeover assumptions, rather than dividing quantity by nominal cycle time alone.
Keep provisional opportunities separate from confirmed production commitments. A weighted pipeline is useful for planning scenarios but is not a reservation ledger. If several quotes promise the same available slot, define what happens when more than one wins: revised date, alternate route, overtime, subcontracting or a capacity commitment policy.
Score feasible RFQs with explicit evidence quality
Once gates pass, a compact score can support prioritization. It should expose judgment rather than pretend to remove it. The weights below are a proposed starting design, not an industry benchmark.
| Dimension | Illustrative weight | Strong evidence | Weak evidence |
|---|---|---|---|
| Expected contribution quality | 30% | Current cost basis and credible price range | Revenue target only |
| Win position | 25% | Known need and relevant differentiation | “Good relationship” without buying context |
| Capacity fit | 20% | Fits actual bottleneck window | Unchecked requested delivery |
| Repeat potential | 15% | Documented programme or repeat history | Verbal expansion possibility |
| Estimating efficiency | 10% | Reusable route and mature drawing | Extensive design work before qualification |
Score each dimension on a defined scale, such as one to five, and record the reason. Label evidence quality separately. A four-out-of-five score based on assumptions should not look identical to a four supported by current documents.
Avoid allowing the total to overrule a failed gate. Maintain a small exception path for strategic work, with a named sponsor, explicit learning objective and a limit on estimating hours. “Strategic” without a cost boundary is an easy way to avoid making the decision.
Establish contribution before celebrating order value
A workable early estimate separates material, external processes, direct conversion, setup and other relevant costs. Detailed BOM and route costing comes later, but the first pass must be credible enough to distinguish attractive work from a large unprofitable order.
Microsoft's BOM documentation explains that component requirements combined with a production route form a basis for estimated product cost. Its calculation documentation also makes quantity, date, site and applicable versions explicit. Those are useful reminders that a copied historic unit cost may describe a different job. Microsoft BOMs and formulas, Microsoft BOM calculations.
For bid triage, use a contribution definition agreed with finance and clearly identify included costs. Do not switch between gross margin, throughput contribution and operating profit inside one ranking. Where bottleneck capacity is full, compare contribution per constrained hour as well as total contribution. Where capacity is genuinely idle, the tradeoff may differ.
Fictional USD example. Monetary assumptions in this guide are illustrative US-dollar amounts, not quoted prices, measured results or exchange-rate conversions. The currency does not imply a US tax or legal setting.
Synthetic example · assumed win probabilities
Expected values are decision assumptions, not earned profit. If both win, B and C need 210 bottleneck hours and do not fit in 100 available hours.
Worked example: a smaller RFQ earns the first estimate
Consider three synthetic opportunities, all of which pass the technical gates. RFQ A has estimated contribution of USD 30,000, a judgmental 25% win probability and requires 20 estimating hours. RFQ B has USD 18,000 contribution, 50% win probability and requires eight hours. RFQ C has USD 12,000 contribution, 60% win probability and requires six hours.
Expected contribution before estimating cost is USD 7,500 for A, USD 9,000 for B and USD 7,200 for C. Per estimating hour, those figures are USD 375, USD 1,125 and USD 1,200 respectively. If the team has fourteen hours available before the deadlines, B and C together require fourteen hours and offer modeled expected contribution of USD 16,200. A cannot be completed within that window without moving other work or reducing scope.
At an assumed USD 120 per estimating hour, B and C cost USD 1,680 to estimate, leaving USD 14,520 of modeled expected contribution after estimating cost. This is a prioritization calculation, not predicted revenue. The probabilities are assumptions and the jobs may compete for production capacity if both win.
Now test a weaker win position: if B's probability is actually 20%, its expected contribution falls to USD 3,600, or USD 450 per estimating hour. C still leads the ranking, but B's advantage over alternatives shrinks. A short customer qualification conversation that changes the probability estimate can be worth more than polishing the cost workbook.
Assume B uses 90 bottleneck hours and C uses 120. Their full-job contribution per bottleneck hour is USD 200 and USD 100. If only 100 production hours remain in the required window, winning both is not feasible without an approved alternative. The bid decision must therefore pair estimating priorities with delivery conditions; expected values cannot manufacture factory capacity.
Bid
An authorized owner accepts the supported opportunity.
Clarify
Resolve a missing requirement before committing.
No bid
Record the reason and communicate the decision clearly.
Conceptual guide by Looqus Editorial. Read the surrounding analysis for assumptions, exceptions and primary sources.
Ask the questions that change a bid decision
Qualification should reduce specific uncertainty. Ask whether the drawing revision is final, what required delivery means, which criteria determine award, whether the buyer will accept a phased delivery, and what quantity is committed versus forecast. The goal is not a long intake form; it is a better allocation of estimating effort.
For a budgetary request, offer a bounded estimate with clear assumptions and an expiry condition instead of investing in a fully engineered quotation without a commercial reason. For an incumbent comparison, establish where the manufacturer can credibly differ: lead time, process capability, quality evidence or order flexibility. Do not invent differentiation from a generic “AI-powered” claim.
When information is missing, name the decision it prevents. “Material specification needed to choose a valid process route” is more useful than “incomplete RFQ.” The specialist's task is to obtain the missing evidence or record an explicit limit on the quotation, then verify that the estimator is no longer blocked.
Keep a disciplined no-bid and conditional-bid path
A no-bid should preserve commercial clarity. Record the reason, whether the issue is permanent or timing-specific, and any agreed follow-up condition. A delivery mismatch today may become a good opportunity next quarter; a process outside the business's capabilities is a different category.
Conditional bids need equally clear boundaries. State the revision, quantity, delivery assumptions and approvals on which the price depends. Avoid adding vague qualifications that the buyer cannot interpret. If the condition fundamentally changes the requirement, resolve it before treating the opportunity as a firm quotation.
NIST's supplier-selection guidance emphasizes evaluating total ownership cost and broader supplier performance rather than purchase price alone. For a manufacturer selling to that buyer, the useful inference is to explain the relevant delivery and quality tradeoffs with evidence; it does not imply that every buyer will pay a premium. NIST supplier selection.
Run the triage meeting around decisions and owners
Keep the review short enough to happen consistently. Present the technical gates, unresolved questions, estimating effort, capacity concern and recommended disposition. The outcome should be bid, conditional bid, clarify, defer or no-bid, with an owner and a due date.
Do not let “clarify” become permanent storage. Age the unanswered requests and set a rule for when the team stops spending time without buyer engagement. Likewise, a bid approval should release a defined estimating budget, not unlimited work. Escalate when drawings or scope change enough to invalidate that budget.
AI can assist with document classification, version comparison and preparation of the decision record. Any candidate automation needs testing against real RFQ formats and explicit permission boundaries. Engineering acceptance, commercial approval and external communication remain governed by the agreed operating process.
Measure whether better selection produces better work
Track estimating hours per qualified quote, response time for priority opportunities, win rate by decision category and contribution estimated versus realized. Segment by new versus repeat customer and by process family. A higher overall win rate can simply mean the business stopped quoting new accounts, which may or may not match its strategy.
Review declined opportunities selectively. If attractive work is repeatedly rejected because one approval is slow, improve that bottleneck. If accepted jobs consistently miss contribution estimates, revisit costing assumptions before tightening the bid score. Triage cannot repair systematically wrong cost data.
Pilot with a narrow product family and compare a defined period against the prior process, accounting for changes in demand and customer mix. Treat a small sample as directional evidence. Do not claim revenue uplift from a few wins whose outcomes were influenced by price, market conditions and sales effort as well as triage.
Decide what to improve first with Looqus
This framework fits manufacturers with more RFQs than estimating capacity, recurring technical families and a reasonably visible production constraint. It is less useful where every project is unique, cost data is unreliable or senior staff cannot agree what work the business wants. In those situations, the first improvement may be decision ownership and cost visibility rather than automation.
Looqus can organize RFQ evidence, obtain missing information and keep each opportunity moving toward an explicit bid decision. Qualified technical and commercial owners retain approval; specialists close the information gaps and verify that the estimating team has a usable decision. AI does the work. We own the outcome.
Sources
- Microsoft: finite capacity planning.
- Microsoft: bills of materials and formulas.
- Microsoft: BOM calculations.
- NIST: supplier selection process.
AI does the work.We own the outcome.
Direct limited estimating time toward feasible work, with technical and commercial questions resolved before quoting competes for the same production capacity.
- Agree the result
- Connect the work
- Own the follow-through