What it looks like
when a match holds.
A search is worth judging on one thing: whether the relationship was still working a year later. These are real searches. Where a client and a candidate have agreed to be named, they are named. Where they have not, the case is anonymized and the details are still true.
Every case has two owners, so we tell it from both sides.
- The company's needWhat the business actually had to have, which is rarely what the job description said.
- The candidate's needWhat would make a move worth it for a person who was not looking.
- The tensionThe thing that surfaced mid-search and could have ended it.
- What changed before the offerThe scope, the terms, or the read that had to move before anyone signed.
- How it heldWhere the relationship stood at 90 days, or a year.
The two cases below are told from the search side, which is the side we can document in full. We publish a candidate's half of the story only with that person's permission, so some cases carry four beats rather than five until we have it.
When the process works.
Case 1 · Said no When the process said no.
Mountain West custom home builder · Operations Director
Read the case Close
How a multi-touch interview architecture stopped the wrong hire before he landed in an Operations Director role at a Mountain West custom home builder.
After the first interview the owner enthusiastically said he would like to move quickly with an offer. We encouraged him to stick with the process and continue due diligence. Four team interviewers met the candidate next. Their confidence scores came back 9, 8, 5, and 2 out of 10.
The 2 was right.
The candidate looked, on paper and in early stages, like a strong front-runner. Polished resume, regional credibility, clean screen and intro. By surface signals, he belonged in the final round. The interview process was deliberately designed not to trust surface signals.
Five touchpoints, structured to produce independent evidence. A CEO first round. Multi-stakeholder team interviews with structured Reflection forms, each interviewer scoring independently before any group debrief. A CEO deep-technical second round. A reference pass. A community signal sweep distinct from the background check.
The Reflections came back with extreme dispersion. Two interviewers saw a polished professional. One flagged a candidate who would resist coaching. The CEO, after a two-hour technical deep-dive with drawings, found the candidate could not read plans accurately even with assistance. Three independent observers, three different reads of the same person.
The dispersion itself was the diagnostic. The candidate was good at the parts of the interview that reward charisma and bad at the parts that require depth. That is the signature of someone optimizing for the appearance of fit.
Candidates who are hiding something tend to resist a thorough process. When the complaint is about the process itself, that's usually the signal, not a logistical concern to smooth over.
The candidate was disqualified before he landed. The role went unfilled for another month. That month was worth more than 18 months of cleanup.
Case 2 · Said yes When the process said yes.
General Contractor · Pacific Palisades · Superintendent
Read the case Close
VCI (Vallernaud Construction, Inc.) didn't come to Ambassador Group in a "we need more applicants" situation. They came in a "the last Superintendent hire is underperforming and we need to stop bleeding on this role" situation. President Damien Vallernaud's first frame was understandable: get me more candidates.
The pivot AG ran was: the volume isn't the problem. The filter is.
The search re-scoped before outreach began. VCI is a boutique, owner-led shop running ultra-high-end residential in Pacific Palisades. A Superintendent there has to carry the project, the trades, and the client relationship without the President triangulating every decision. That is not a smaller version of a big-GC superintendent role. It is a different job.
The interview architecture produced documented evidence at every stage. Damien's structured Reflection on the second interview came back at 9/10 with one specific open question, which got resolved before the offer. Rob Brant signed January 15, 2026 and started January 20.
The integration program ran from offer-acceptance forward, not from Day 1. AG's Assessment and Integration Advisor ran a Pre-Start call before Rob's first day, then the structured 2-Month check-in in early April. The integration program treats the first 90 days as a measurable system, not a hand-off.
Rob arrived ready to download rather than perform. The trades who had been in flux under the prior Superintendent began calling Rob directly within the first month. The 2-Month integration recap captured the pattern in the Advisor's own words: highly engaged, taking real ownership for outcomes.
If your last match didn't hold, the question isn't where to find more candidates. It's what filter let the wrong candidate through, and whether you've rebuilt it before opening the next search.
Watch · 18 min · with Crestwood leadership
A conversation with Crestwood Construction.
A long-form interview with the Crestwood leadership team about the searches we ran together, what the multi-touch process surfaced, and where the 90-day landing plan made the difference.
Audited outcomes.
We are assembling a reviewed, year-over-year record of how our matches hold: retention at one year, time to a signed offer, and how often a search ends in advice not to hire. It is not published yet because it is not audited yet. When it is, it will appear here with its method written next to it. Until then, we would rather show you two searches in full than a number you have no way to check.