Industrial OEMs do not have unlimited bid capacity. Every serious RFQ pulls people away from another request: sales has to qualify the buyer, engineering has to interpret the duty, estimating has to build the cost, sourcing has to price bought-out parts and finance or legal may have to review the terms. A bid that looks free in CRM can consume a week of the company’s hardest-to-replace work.
Revenue alone tells you very little. A large package can have an impossible date, a weak competitive position and uncapped technical risk. A smaller repeat order can fit the product, the buyer and an open production slot. The framework has to answer two different questions: how likely are we to win, and what happens if we do?
Start with conditions a score cannot fix
Do not put a fatal condition in a weighted score. A preferred account and a large order should not cancel an unavailable certification, a design outside the company’s technical limits or a delivery date that no credible plan can meet.
Define the stop conditions before the team scores anything. Some conditions mean no-bid. Others mean “stop until the named owner approves an exception.” Make that difference explicit. A salesperson should not have to guess whether a missing hazardous-area certification can be escalated or whether it ends the pursuit.
| Stop condition | Evidence to check | Who may clear it |
|---|---|---|
| Legal or trade restriction | Customer, end use, country, sanctions and export-control result | Legal or compliance |
| Mandatory certification is unavailable | Customer specification, jurisdiction and current product certificates | Product or engineering authority |
| No technically feasible offer | Required duty against the approved product and engineering envelope | Chief engineer or product owner |
| No credible delivery plan | Engineering load, bottleneck capacity, supplier dates and test slot | Operations leader |
| Terms exceed delegated authority | Liability, damages, warranty, payment, bonds and insurance | Named commercial approver |
| Too little information to price responsibly | Missing inputs that control scope, cost or acceptance | Bid leader after the buyer answers or an assumption is approved |
Make the decision three times
One approval at receipt is too early to be permanent. The team learns the most expensive facts after engineering, suppliers and the customer answer questions. Run the decision at three points and spend more effort only when the opportunity earns it.
A PMI case study on proposal development puts a first risk assessment, delivery timing, assumptions and constraints ahead of the first bid decision. That sequence fits industrial OEM work: each gate needs enough evidence to decide whether the team spends the next block of time. It does not need every answer the final proposal will contain.
| Gate | Question | Minimum evidence | Decision |
|---|---|---|---|
| 1. Qualify | Should we spend technical time on this request? | Customer, project status, product fit, deadline, basic scope and stop conditions | Decline, hold for information or fund the first technical pass |
| 2. Commit | Should we fund the full estimate and proposal? | Technical path, major bought-out items, capacity, competitor position, cost range and material terms | No-bid, pursue with actions or commit the bid team |
| 3. Submit | Are we willing to make this exact promise? | Final scope, price, margin, delivery, exclusions, risk reserve and approvals | Approve, revise or withdraw |
Separate the chance of winning from the cost of winning
A bid can fail in two directions. The customer may never place the project or may choose a competitor. That is external risk. The OEM may win and then miss the cost, schedule or performance promise. That is internal risk. Mixing both into one “opportunity score” hides the difference.
PMI’s work on risk in the bidding phase makes this split directly: external risk concerns the competitive value of the bid, while internal risk concerns deviation from the delivery baseline. The response is different. Weak buyer access needs capture work. An unproven design needs engineering evidence, a risk allowance, a changed offer or a no-bid.
Ask sales to support the win case with facts: buyer access, project funding, decision process, incumbent position, specification influence and known competitors. Ask engineering and operations to support the delivery case with a technical path, cost range, schedule, supplier plan and acceptance basis. Neither side should score the other from intuition.
Score six questions with evidence
Score only the requests that clear the stop conditions. A short 0–3 scale works better than a theatrical 100-point model if each number has a written meaning. Record the evidence beside the score. If the reviewer cannot cite a customer message, system record, prior job, current plan or named judgment, the input is an opinion and should appear as one.
Research on bid/no-bid decisions covers resources, company fit, the chance that the project proceeds, competition and risk. For an industrial OEM, those ideas become six practical questions.
- A usable starting weight is 15% business fit, 15% project reality, 20% buyer position, 20% delivery fit, 20% economics and 10% risk position.
- Set a minimum score for delivery and economics. Do not let a high total cancel a 0 in either one.
- Use the total to rank the bids that clear those minimums. Test the weights and cutoffs against your own submitted bids and delivered orders before treating them as policy.
| Question | Strong evidence | Weak evidence |
|---|---|---|
| Does the project fit the business? | Named product, market, account and service priorities | The revenue is large |
| Is the project real? | Budget, buying process, decision date, site and internal sponsor | The buyer issued an RFQ |
| Why should this buyer choose us? | Installed base, access, specification position, past delivery and competitor gap | Sales feels good about it |
| Can we deliver the promise? | Engineering capacity, factory constraint, supplier plan, test slot and schedule logic | Operations will find a way |
| Will the order produce a good return? | Price range, cost range, risk reserve, cash terms, bid cost and aftermarket value | The target margin passes |
| How much can still move? | Known technical, cost, schedule and contract risks with owners and ranges | The scope looks normal |
Anchor the scale before people use it
Write what 0, 1, 2 and 3 mean for each question. Without anchors, one reviewer’s 2 is another reviewer’s 3 and the total looks more exact than the discussion behind it. Do not copy one set of anchors across every product family. A confirmed foundry slot may control one bid; test-cell capacity or a licensed engineer may control another.
| Delivery-fit score | Meaning |
|---|---|
| 0 — no plan | The requested date conflicts with a known constraint and no approved recovery plan exists |
| 1 — major exception | The plan depends on overtime, expedite, unconfirmed supplier capacity or displacement of committed work |
| 2 — workable | The current plan supports the date, with named risks and normal contingency |
| 3 — strong fit | Confirmed capacity and supplier dates support the request with room for expected variation |
These anchors are an example, not a universal policy. Build yours from actual late orders, expedited suppliers and missed test windows. Then calibrate the scores against what happened after the win.
Charge every pursuit for bid effort and constrained capacity
Estimate the hours required from sales, application engineering, estimating, sourcing, finance and legal. Use the effort this request will consume. A first-of-kind package with a 200-line compliance matrix should not carry the same bid cost as a repeat configuration with updated quantities.
A useful economic check is: probability of win × expected order contribution, less the cost of bidding. Use a risk-adjusted contribution range rather than a single optimistic margin. The GAO Cost Estimating and Assessment Guide explains why a cost range gives decision-makers more information than a point estimate and warns against arbitrary risk percentages. Industrial bids need the same discipline when incomplete requirements, new design work or supplier uncertainty can move cost.
Expected contribution alone is not enough when one resource limits throughput. If test engineering is full, compare the return per constrained test hour. If a bid will occupy the only large machining slot, show which committed or higher-value work it could displace. The bid competes with every other bid and order for the same people and equipment.
Do not turn missing facts into an average score
Unknown does not mean neutral. If nobody knows whether the project has funding, the “project is real” question is unanswered. If the buyer has not supplied the operating condition that controls equipment selection, delivery and economics are not ready to score. A default midpoint quietly rewards incomplete RFQs.
Show the missing fact, who must obtain it, when the gate must close and which decision depends on it. If the team has to proceed, record the assumption and the person who accepted it. That makes the exposure visible in costing and in the final proposal instead of burying it inside a score.
Example: two bids need the same engineering team
An OEM receives two requests in the same week. The first is a large compressor package for a new customer. The buyer wants a firm price in ten days, the gas composition is incomplete and the requested test date conflicts with the current shop plan. Sales has no access beyond procurement. The second is a smaller retrofit package for an installed-base customer. The site data is complete, the design reuses a released frame and the outage date matches an available test slot.
The large package clears the legal gate but stops at technical feasibility until the buyer supplies the gas composition. Its win case is also weak because the project status and decision process are unknown. The retrofit clears the gates, needs fewer engineering hours and carries better evidence on win probability and delivered cost. It enters full estimating first.
Two days later, the new customer supplies the missing composition and agrees to a later factory test. Engineering confirms a feasible design, but sourcing finds that the selected driver has a 38-week lead time. The team reopens the commit gate. It offers an alternate driver and a qualified delivery date instead of forcing the original promise through the score. The framework changed the bid, not just its status.
Record the decision, the dissent and the reason to reopen it
The record should show the gate, the evidence used at the time, scores, open questions, decision, approvers and any override. If engineering says the performance basis is weak and sales chooses to proceed, preserve both views. Do not replace the engineering concern with the final approval.
Reopen the gate when a fact that supported the decision changes: scope, requested date, technical path, supplier lead time, cost range, terms, project funding or competitive position. Shipley’s capture-planning guidance calls for regular decision-gate reviews that either advance or end the pursuit. A large industrial bid needs that discipline because the request can change several times before submission.
What Bourne does in the bid review
Bourne assembles the checked RFQ, customer and installed-base history, similar wins and losses, current technical exceptions, capacity signals, supplier responses and commercial terms in one review. It applies the stop conditions, calculates the agreed score and shows the source behind each input. Missing evidence becomes owned work instead of a blank cell.
The bid leader approves, holds or declines the pursuit. Any override preserves the original result and requires a reason. An approved request opens the technical clarification, costing and supplier work with the same assumptions and risk record. A no-bid gives sales a specific reason and becomes evidence for the next review.
Judge the framework by the orders it produces
A higher win rate does not prove the framework works. The team may win more by accepting bad terms or promising dates it cannot meet. Compare the original decision with both the sales result and the delivered result. Use the pattern to change gates, evidence requirements and score anchors.
| Measure | What it reveals |
|---|---|
| Engineering and estimating hours per submitted bid | Whether the gates reduce effort spent on weak pursuits |
| Time from receipt to each gate | Whether decisions happen early enough to protect the proposal deadline |
| Win rate by decision band | Whether the external-risk evidence separates stronger and weaker pursuits |
| Quoted margin versus delivered margin | Whether the internal-risk review finds cost and execution exposure |
| Schedule performance and change orders | Whether accepted assumptions turn into delivery problems |
| No-bid reason and later outcome | Whether the business declines the right work and learns from mistaken calls |
Review overrides separately. A sound exception can win a good strategic order. A repeated pattern of overrides followed by margin loss means the gate has no authority. Judgment still decides the bid. The record gives that judgment current evidence and a memory.
Buğra Gündüz
Buğra Gündüz is the co-founder and CEO of Bourne and co-founder of HockeyStack. He built HockeyStack into an eight-figure AI business. At Bourne, he works with entrepreneurs and established companies to create AI products and services.