When two construction finalists both clear the brief, decide on evidence the interview cannot supply: how each handles the specific work of the first year, what references who watched them do that work say, and which risk you can manage. Who won the room is one score, not the decision.
The debrief splits two to two. Your CFO wants the finalist who ran the room for ninety minutes and left everyone a little energized. Your senior project executive wants the other one, who answered every question correctly and slightly slower. Two construction finalists for one VP of operations seat in Houston, both credible, and the decision is drifting toward whoever told the better story. Nobody has yet asked the question that would settle it.
How should you choose between two construction finalists?
Score both against the same written tests: the brief’s must-haves, a work sample built from year one’s problems, and references from people who watched the work. Score before anyone states a preference, then weigh the risk each carries. The interview impression stays on the sheet as one score among four.
Order matters. Once a senior leader has said “I liked her” out loud, every later piece of evidence gets read as confirming or threatening that view. So each panel member scores alone, in writing, before anyone talks.
Using the same tests for both is also the defensible route. EEOC guidance on selection procedures says a selection procedure must be job-related and evaluate the skills of the particular job in question.
Why the interview winner is the wrong tiebreaker for construction finalists
You think you want the candidate who impressed you. You need the one who will be impressive in month seven, on a Tuesday, when a GMP job in Nashville has burned its contingency and the owner’s rep wants a recovery schedule nobody has built.
Those can be the same person. Often they are not. An interview tests how someone performs in an interview, a skill a VP of operations uses perhaps twice a year, at the owner presentation and the board meeting.
The federal government’s own guidance on structured interviews sets out the fix: every candidate gets the same predetermined questions, and every answer is rated on the same scale and standards. Most contractor loops do neither. Each interviewer brings a favorite question, and the finalist who matches the panel’s mood wins on tone.
The finalist who sold you best is, by definition, the one most skilled at making you feel certain, and certainty is exactly what you should not feel yet.
What work will the construction finalists face in year one?
List the four or five specific problems the seat inherits, then ask each finalist to work one of them: a recovery plan on a named job, a forecast review, a superintendent bench gap. How they reason through your problem predicts more than how they describe their past ones.
The federal guidance on work samples makes the case: tasks that mirror the job are highly representative of it, and performance on them relates closely to performance in the role. A ninety-day plan asked of each finalist earns its place only when it is written against these problems.
One reliable sample is the forecast question. Hand each finalist a monthly cost report from a live job and ask what is still outstanding that has to go into the forecast. Site teams often cannot answer, because the item that will cost the money has not announced itself. A strong operations leader knows where to look: the verbal instruction nobody papered, the scope gap between two subcontracts. A weaker one reads the report back to you.
The rest of the list depends on your firm:
- The job that is behind, and who on the team will need replacing
- The gap between your commercial and field teams, who often run one project on different information
- The superintendent who expected this seat and did not get it
- The backlog your surety will ask about, and who is named to deliver it
A recovery plan explained at the trailer on a jobsite interview is a harder test than one explained in a conference room.
How deep should references go on two construction finalists?
Deep enough to reach people who watched each finalist do the year-one work, not people chosen to praise them. Aim for at least three references each, including one former subordinate and one person from the commercial side, all asked the same structured questions.
Federal guidance on reference checking recommends a minimum of three contacts and the same job-based questions for every reference. Most contractor reference checks are two friendly calls and a box ticked.
Ask what the person controlled, not whether they were liked: who owned the change-order position on the job that went sideways, and who called the owner when the schedule slipped. We set out that method in reference checks. OSHA’s establishment search adds inspection history by employer, which describes the operations a finalist ran, not the person alone.
Scoring two construction finalists side by side: a worked illustration
A worked illustration with assumed figures shows how the tie breaks. Score each finalist from 1 to 5 on four tests. Finalist A scores 5 in the interview room, 3 against the brief’s must-haves, 2 on the year-one work sample and 2 on reference depth, for a total of 12.
Finalist B scores 3 in the room, 4 against the must-haves, 5 on the work sample and 4 on references, for a total of 16. On impression alone, A wins easily. Across the four tests, B leads by 4 points, all of it in the evidence the job will draw on.
The numbers are illustrative; the shape is common.
| Test | What it measures | What it misses |
|---|---|---|
| Interview room | Presence, fluency, owner-facing polish | Decisions under pressure with incomplete numbers |
| Brief must-haves | Fit to the written seat | Problems the brief did not anticipate |
| Year-one work sample | Reasoning on your live jobs | Stamina across twelve months |
| Reference depth | What they controlled, seen by others | Their fit with your particular owner |
If the totals land within a point, the tie is real. Then decide on risk: which finalist’s weakness your organization can cover, and which would land on you personally.
What does picking the wrong construction finalist cost?
The wrong pick usually costs about a year. A weak operations leader is rarely obvious at ninety days; the cost shows up in the second WIP cycle, in fade on two jobs and in a project executive who leaves. By the time you reverse it, you are searching again with a weaker hand.
The median annual wage for construction managers was $114,990 in May 2025, and a VP of operations sits well above it. Twelve months of a senior salary, a second search and a second learning curve are the visible bill. The hidden one is the margin on the jobs that leader touched, and your surety’s view of who is delivering the backlog, which we covered in bench depth.
The wrong pick also costs you the right one. By month nine, the runner-up you declined has taken another seat, and the project executives reporting to the new VP have already decided whether to stay. You are the one who explains all of it to the surety.
What should you do with the runner-up?
Tell them the truth within a week: they were close, what decided it, and whether you would want to talk again. Then keep in touch on a real schedule, not as a silent backup. Treated well, the runner-up is often the next hire.
Honesty here is practical. A finalist kept on hold while you negotiate with the other usually finds out, and candidate withdrawal runs in both directions. BLS turnover data tracks hires and quits in construction every month, and a good runner-up is in that flow.
Keep the note specific. “We chose someone who has run a recovery on a negotiated healthcare job” tells them something useful; “we went another direction” tells them nothing. If the seat was confidential, say so again in the same message.
Before the offer goes out
The question that settles a tie is which finalist you would hand your worst job next Monday, and what evidence says so. Write the answer in one paragraph, cite the four tests, and file it with the hiring brief. If the paragraph leans on the interview, the decision is not finished. For a project executive or operations seat held in confidence, the process is confidential executive search for mandates that cannot be publicly posted, with the four tests built into the shortlist before your panel meets anyone. Take the two names and the year-one list to our employer page and make the choice on evidence.
Sources and further reading
- Selection procedures, EEOC guidance on employment tests and selection procedures
- Structured interviews, OPM assessment and selection guidance
- Work samples, OPM guidance on work samples and simulations
- Reference checking, OPM guidance on structured reference checks
- Establishment search, OSHA inspection records by employer
- Construction managers, BLS Occupational Outlook Handbook pay data
- Turnover data, BLS Job Openings and Labor Turnover Survey