There are two failure modes for pursuit decisions. Shops with no meeting drift into bidding everything, because no forum exists to say no. Shops with a bad meeting drift into an hour of project description, war stories about the owner, and a decision made by whoever talks longest. Both produce the same result: estimating hours allocated by momentum instead of math.
The fix is a meeting with three properties: it is short because the analysis happens before it, structured because the scorecard is its agenda, and binding because decision rights are written down. Fifteen minutes per pursuit is not an aspiration — it's what the format below produces when the pre-work is done.
Before the meeting: the price of admission
A pursuit does not get on the agenda until three artifacts exist:
- The 25-question screen completed, with no automatic kills tripped. Killed pursuits never reach the meeting — the estimating lead declines them directly.
- The seven-question scorecard filled in, with the verdict, red flags, and levers visible.
- The cash exposure line: peak AR and its percentage of working capital, straight from the model in the Excel workbook.
This rule does most of the work. The meeting can only be fifteen minutes because it consumes analysis instead of producing it.
Who is in the room
| Role | Who | Why they're there |
|---|---|---|
| Scorecard owner | Estimating lead / chief estimator | Presents the score, defends the answers |
| Operations voice | GM or the PM who would run it | Reality-checks schedule fit and personnel |
| Money voice | CFO or controller | Owns the AR/bonding veto above the cash threshold |
| Principal | Owner/president | Only for strategic exceptions — not weekly |
Five people is the ceiling. Estimators who worked the screen attend only if they'll present; observers turn decisions into performances.
The 15-minute agenda
- Minutes 0-2 — The brief. Project, owner, value, due date, delivery method. No history, no anecdotes. It's on one screen.
- Minutes 2-7 — The scorecard walk. Seven questions, seven scores, called in order. Debate is allowed only where a score is contested — and the contest is about evidence ("their last three jobs paid at 54 days, that's a 2, not a 1"), not feelings.
- Minutes 7-10 — The cash check. Peak AR, percent of working capital, retention terms. The money voice speaks here and only here — but what they say is binding.
- Minutes 10-13 — The decision and its conditions. The verdict band suggests the answer; the room confirms or overrides it. A "bid loaded" verdict gets its load named in the room — points on the fee, a schedule contingency, an excluded scope. A strategic bid gets its strategic reason written down.
- Minutes 13-15 — Actions. Bid: who runs the takeoff, quote list owner, site-visit date. No-bid: who sends the letter (within 48 hours), what goes in the pursuit log.
Decision rights, written down
The meeting stays fast because nobody is negotiating for authority in it:
- The estimating lead decides inside the scorecard bands. A 48-point pursuit doesn't need permission to bid; a 15-point pursuit doesn't need permission to die.
- The money voice holds an absolute veto on cash grounds. AR over the line is not a debate — it's the hard-kill rule with a person attached.
- The principal can override to strategic — and every override is logged as one. After a year, you can audit them: our data shows strategic overrides win at roughly half the rate the owner remembers.
"We stopped arguing about jobs and started arguing about scores. That sounds like the same thing. It is not — an argument about a score ends with evidence."
Chief estimator, mid-Atlantic mechanical contractor
Skip thresholds: not everything earns a meeting
The weekly slot should carry four to six pursuits, not fifteen. Set two bypasses: pursuits under a size floor (say $250k) with a scorecard above the bid line are auto-approved by the estimating lead; anything tripping an automatic kill is auto-declined the same day. The meeting exists for the middle band and for anything above the cash threshold — the decisions where judgment actually adds something to the arithmetic.
The three anti-patterns
- The re-litigated no-bid. A GC calls the principal, the dead pursuit comes back. Rule: a decided pursuit reopens only with new information that changes a score — and the change is named before the meeting.
- The scorecard done live. Filling in answers during the meeting doubles its length and lets the room anchor each other. Scores come in written; the meeting contests them.
- The undocumented condition. "Bid it, but keep an eye on the schedule" is not a condition. "Bid it with two weeks of general conditions added for the September stack" is. If the load isn't written, the estimator can't price it.
Run this format weekly and the byproduct is as valuable as the decisions: a pursuit log where every entry has a score, a verdict, and an outcome — the dataset that turns next year's thresholds from defaults into your own empirical hit-rate curve.
And when the answer is bid: PILARS runs the takeoff in hours, not weeks — $100 per trade, per plan — so the decision you just made doesn't die waiting in the takeoff queue. See pricing here.