A bid/no-bid scorecard is a promise a firm makes to itself: this is how we'll decide which projects deserve our estimating dollars, before enthusiasm for a particular pursuit gets a vote. Building one raises two practical questions—what to score, and how much each answer should count. This article works through both, along with the veto question that most scorecard templates dodge.
The criteria families that belong on the card
The specifics vary by firm and market, but strong scorecards draw from a recognizable set of families.
- Client quality and history. Does the owner pay on time, administer contracts reasonably, and resolve disputes without litigation? Prior experience with this client is worth more than any other single signal. A first-time client isn't disqualifying, but the unknowns should cost points.
- Project fit. Is this a building type, delivery method, and size your teams have done well recently? A project 40% larger than anything in your history is a different pursuit than one squarely in your lane, whatever the margin looks like.
- Backlog and team availability. A good project your best people can't staff is not a good project this year. Score the realistic availability of the PM and superintendent who would actually run it, not the org chart's theoretical capacity.
- Geography. Distance from your operating footprint costs money in supervision, labor relationships, and sub coverage, and the cost curve is steeper than most pursuit conversations admit.
- Competition and win probability. How many bidders, and who are they? Chasing a project with eight capable competitors is buying a lottery ticket with your precon budget.
- Margin potential and commercial terms. Expected fee, contingency posture, and the contract's risk allocation—onerous flow-downs, aggressive liquidated damages, thin retainage terms—belong here.
- Schedule and bond exposure. Compressed schedules and large bond requirements amplify every other risk on the list.
Twelve to eighteen questions across these families is usually enough. Past that, reviewers start pattern-matching instead of thinking.
Weighting without false precision
Weights should express your firm's actual priorities, and they should be embarrassingly simple. If client quality has burned you more than anything else, weight it heaviest and say so. A common structure: score each question 1–5, multiply by a per-question weight, and express the result as a percentage of the maximum. Resist decimal-place sophistication—a model that says 73.4 versus 71.9 implies a precision the inputs don't have. What you want from the arithmetic is consistency across pursuits and time, so that "we bid everything above 70" is a policy someone can audit, not a mood.
The discipline matters more than the calibration. Weights you revisit annually against real outcomes will converge on something useful; weights debated to perfection before launch will ship late and get revised anyway.
The veto row
Some conditions should end the conversation regardless of the composite score: an owner with unresolved claims against you, a contract your surety won't stand behind, a delivery date your operations lead calls impossible in writing. Put these on the scorecard as explicit deal-breakers—questions where a single critical "no" kills the pursuit even if every other row is glowing.
The veto row does two jobs. It prevents a high composite from steamrolling a fatal flaw, and it protects the dissenting reviewer. A risk manager who scores one cell "no" with a sentence of explanation is exercising the process; the same person arguing alone against an enthusiastic room is exercising courage. Scorecards should not require courage.
Thresholds, and what happens at the margin
Set the pass threshold per gate, and decide in advance what the margin means. A pursuit scoring far above threshold advances without discussion; far below, it dies without discussion—those cases are why the scorecard exists. The interesting band is the few points either side of the line, and the right treatment is usually a conversation with the narrative sections in front of the decision-maker, not a mechanical call. Deliberate overrides happen and should be recorded with their reasoning; a scorecard that logs its own exceptions gets smarter, because next year's weight review can examine whether the overrides outperformed the rule.
Consistency is a forms problem before it's a judgment problem
Two reviewers can hold the same opinion and produce different scores if the instrument lets them. Structured cells with fixed scales beat open text; a defined meaning for each point on the scale ("3 = client is new to us but well referenced") beats a bare number; a narrative field beside the score captures the judgment the number can't. This is also an argument for keeping the scorecard in Excel, where the structure is visible and the formulas are inspectable—a case we make at length in why spreadsheets are still right for risk assessment.
From rubric to running process
A scorecard only pays for itself when it's actually completed, by the right people, on every qualifying pursuit—and that's a coordination job no one enjoys. Bex Risk takes it over: your rubric, your weights, and your veto rules are configured once, and from then on a pursuit kicked off with a single email flows through the gates you've defined. Each reviewer receives an Excel scorecard locked to exactly the cells assigned to their role; Bex parses the returns, applies your scoring rules deterministically, chases the slow, routes around the absent, and hands the decision-maker one package per gate with the math done. The scoring logic stays yours and stays visible—Bex executes the rubric, it doesn't replace it, consistent with how we keep humans in the loop across every module.
If your scorecard lives in a template folder and runs on willpower, we can put it to work properly. Send us what you score today—the contact link is below—and we'll show you the same rubric running itself.