Brightline's first commercial bid, run through a pre-mortem: the risk that ranked first was not the one anyone in the room already worried about

The moment

Brightline Heating & Air, six vans, Dayton, Ohio — the same shop as the break-even, price-increase and ap-priority stories. Almost all residential service and replacement, almost no commercial, ever.

A local school district is building a new elementary school and put the HVAC installation and a three-year maintenance contract out to bid. Brightline's owner has known the district's facilities manager for years from smaller service calls and got invited to bid. The contract would be the largest job the company has ever taken — roughly a third of a year's current revenue in one signature — and nothing about it looks like the residential work the shop has done for eleven years.

It is the Tuesday before the bid is due. The owner, the office manager and the dispatcher sit down in the yard office with the draft bid on the table. Everyone in the room already has an opinion about what could go wrong. Nobody has written any of it down in a way that could be compared to anyone else's opinion.

What they did instead

Before the exercise, asked directly, the three of them named three things: the district could lose funding, a competitor could underbid the next one, and a warm winter could leave the shop short of cash regardless of the school job. All three are real. All three are also things that happen *to* Brightline, not things Brightline does.

Run through the tool exactly as stated — three causes, all marked as outside the shop's control — the result is not encouraging in the way a short list usually feels. Zero of the three qualify as something to act on this month; all three land on a watch list instead. The tool's own finding says it plainly: Nothing here is both actionable and top-ranked.

A second finding puts a name on the pattern: Every failure you imagined was somebody else's fault.

A third flags the list itself as too short to have found the real ones yet: Only 3 causes written down. Even this thin list produces one severe risk nobody would see coming in time: 1 severe risk you would not see coming, the recession-cuts-the-budget scenario, which scores as severe and hard to detect even though it is also outside anyone's control.

That is the cost of the fear-driven list: it feels like caution and it produces nothing to do on Monday morning.

The questions a CFO asks first

In this order, and the order matters.

1. Eighteen months from now, this failed. Why, in your own words, before anyone else in the room answers?
2. Of everything written down, how many can the shop actually act on this month?
3. Which one would nobody see coming until it had already happened?
4. Which of these does the shop control, which can it only influence, and which is purely something to watch?
5. If every reason on the list is something happening *to* the business, what got left out?

Question 5 is the one the first pass never answered, because a list of things nobody can control never has to name the mistake the shop itself might make.

The arithmetic, shown

The tool scores each cause on three 1-to-5 axes — how bad, how likely, how early you would know — then multiplies: severity × likelihood × (6 − detectability). A risk nobody would see coming scores worse than an equally bad one that shows up on a dashboard next week, because detectability is inverted into the formula, not just noted next to it. Anything outside the shop's control is then discounted, ×0.5; anything the shop can only influence is discounted ×0.85. Controllable risks are scored at full weight.

Scenario two — six risks, scored honestly, before anything changed. The same three people, the same Tuesday, six causes instead of three, each one including at least one the shop itself might cause:

3 of 6 justify doing something this month — the top three, all controllable or influenceable. The top-ranked cause is not the cash-flow risk everyone in the room already worried about out loud. It is the bid-scoping error nobody had mentioned, scored a 5 for severity and a 2 for detectability: bad, and the kind of bad nobody would notice until the job was already losing money. The result carries 2 severe risks you would not see coming — that bid-scoping error, and the district's budget getting cut, which is severe for an entirely different reason: it is outside the shop's control either way.

**Scenario three — the same six risks, after the Friday job-cost check started.** Nothing about the underlying causes changed. What changed is that the shop started reviewing labor and materials spend against the commercial-code-compliant scope every Friday, specifically for this contract — a control that exists for exactly one reason, to catch a scoping error early instead of at the final invoice. That moves the bid-scoping cause's detectability from 2 to 4. Its score drops to 30, and it falls to rank 3. The top three needing action this month are the same three causes as before, only reordered: the payment-terms risk and the staffing-pull risk now rank above it. The blind-spot count is cut in half — **1 severe risk you would not see coming** — and the one that remains is the district's capital budget, which no Friday review could ever have caught, because nothing the shop does moves that number either way.

What changed

The bid went in unchanged in price. What changed was what the owner is actually watching. Before the exercise, the loudest worry in the room was the 60-day payment terms — real, and still ranked second even after the Friday check started. The exercise moved the top item to something nobody had said out loud: whether the bid itself had been scoped correctly for a job with a different code, different inspections and different failure modes than anything the shop has priced before.

The Friday job-cost review did not remove that risk. It changed how long it could go unnoticed, which is the one axis the earlier three-item list never scored at all. The district's capital budget stayed exactly where it was — rank 6, uncontrollable, and now the only entry left on the blind-spot list — with one line written down next to it: the early signal is a delayed purchase order, and the response if it arrives is to pursue the district's smaller annual service contracts instead of the one large build.

What the tool will not do

It will not check whether the severity, likelihood or detectability scores entered are honest. Three people scoring the same bid-scoping risk a 2 out of 5 for likelihood because nobody wants to say 4 out loud produces a clean, confident, wrong ranking, and the tool has no way to see that happen.

It will not tell Brightline what the commercial building code actually requires, or whether the two technicians without certification can get it in time. Those are facts about a specific contract and a specific inspector, and this tool only ranks what someone already knows and is willing to write down.

It will not decide whether to bid on the school job at all. A severe, top-ranked, hard-to-see risk is a reason to build a control for it, not automatically a reason to walk away — that call rests on facts the tool never asked for, like what the shop's other work looks like for the next year.

It does not turn a defensive risk register back into a pre-mortem on its own. The technique only works if the failure is stated as already having happened, out loud, before anyone defends the plan — soften that one sentence and the scores that go in are the scores of a list built to justify the bid, not to test it, and the ranking on the other side is only as honest as what went in.