Company load
Live wiring diagram
This is the view a manager opens before a hard conversation, a crew pairing, or a promotion. It reads the gap between two people and tells you what to expect when they collide.
Three buckets, each modelled from the map and stated with its assumptions. These are projections built on published replacement-cost multipliers, not measured outcomes. The point is not the decimal, it is the order of magnitude and where it sits.
Risk-weighted cost of the departures the map says are most likely in the next twelve months.
Annual rework above benchmark, traced to crews not raising problems early.
Contribution margin lost if the fourth location slips a year for want of a leader.
| Person | Role | Comp | Multiplier | Gross | Risk | Weighted |
|---|---|---|---|---|---|---|
| Risk-weighted twelve-month exposure | ||||||
Multipliers: leadership and management roles at 200% of annual compensation, per Gallup's finding that replacing a leader runs roughly twice salary against about 40% for frontline roles. Scarce-skill senior trades at 150%. Apprentices at 40%. The Center for American Progress puts executive replacement as high as 213% of salary across 30 case studies. Risk percentages are the map's own read, not a published figure.
Four people independently named unmet commitments as their main trust issue. The behavioural consequence is not sabotage, it is silence: crews stop raising small problems early because they do not believe follow-through is coming. Small problems raised late become callbacks.
Install revenue $7.8M · current rework 4.2% = $328K
Good-practice rework benchmark 2.5% = $195K
Attributable delta $133K per year. Assumes the full gap above benchmark is escalation-related. Treat as an upper bound on this one mechanism.
Dale wants Red Deer open inside eighteen months. That needs one person who can run a site without him. The map says exactly one person currently qualifies, and she is the same person already carrying eleven of twenty-four escalation paths at near-empty strain. Tyrell is the real candidate and is twenty-four to thirty months from ready with deliberate development, longer without it.
Red Deer is modelled at $2.4M first-year revenue. A twelve-month slip defers roughly $430K of contribution margin at Ridgeline's blended 18%. Opening it on Rhonda's back instead does not avoid the cost, it converts a 65% flight risk into something considerably worse.
Roughly $1.27M of modelled twelve-month exposure on a $12.4M revenue base, against a first-year engagement at about $205K. That is a 6.2× ratio, which sits inside the 5–6× band boutique advisory work is normally priced to defend. The number an owner should argue with is the risk column, not the multipliers.
Candidates take the same assessment at application. The map already knows the crew. So the question stops being "is this a good tech" and becomes "what happens when this person meets this specific foreman on this specific site."
Decision support only. The map never returns a score, a ranking, or a hire recommendation, and it is not a selection instrument. It tells an interviewer what to probe and a foreman what to expect.
Nobody draws this by hand. Every box below is a job that runs on its own once a person is tagged to a company. Add a person, the map redraws. Add a department, the map redraws. Finish a one-on-one, the map gets deeper.
Every person takes the Deep Insight Profile. Owner, operators, office, apprentices. Same instrument, whole company.
Site walkthroughs and one-on-ones are recorded and transcribed. The transcript is evidence, not a note.
Scores and transcripts write to that person's row under their company. Tagged to a company, they exist on the map.
A worker recomputes strain, reporting lines, pairings, friction and load-bearing edges, then writes the map.
The coach reads the company through the methodology and answers out loud in Christopher's voice.
Every one-on-one a manager runs is captured against the person it was about. Strain is recomputed, not remembered. A map that is six months old is a slide; a map that updated on Tuesday is an instrument. The difference is whether the one-on-one goes into the same record as the assessment.
The pipeline is identical for a roofing company, a mechanical contractor, a fleet, or a law firm. Only the dialect changes: fatigue and callbacks for trades, matter load and client friction for a firm. Sell it to trades first because the pain is loudest and the buyer decides alone.
The assessment engine, the scored profiles, and a coach speaking in a cloned voice are already running with paying customers. The South Peak map exists with people, edges and reads in the database. What this demo represents ahead of the build is the automatic redraw on every assessment completion, the transcript retrieval layer, and the cost projection. This page is a showroom model of the finished instrument, and its coach answers from a script.
This is the part that makes the rest of it work. An owner will always push to know who said what. The answer is no, and the refusal is not squeamishness, it is what keeps the data worth having.
If an owner can unmask a person from the map, people stop telling the truth in assessments and walkthroughs. One cycle of that and the map degrades into an expensive org chart. The confidentiality is not a constraint on the product, it is the thing that keeps the product accurate.
How a person is built. What pressure does to them. How to give them bad news. What their silence means. This is behavioural, it is consented to at assessment, and sharing it is the entire point of the exercise.
Shown named · consented at intakeResentment, broken trust, thinking about leaving. Surfaced as a count and a pattern, never as a name, unless that person has explicitly agreed to be named. The owner learns four people flagged unmet commitments and that two of them are hard to replace. The owner does not learn which four.
Aggregate only · names require opt-inHealth, family crisis, anything disclosed inside a coaching conversation, anything the crisis check flags. This never reaches an owner dashboard in any form. It routes to a named human path, and the person is told where it went.
Never surfaced · human escalation pathThe map does not say "Jesse told us he is leaving." It says the Calgary lead shows a high flight-risk pattern, names the drivers it can defend, gives the revenue that sits behind that person, and tells the owner to have a specific conversation this month. The owner gets the risk and the next action. Nobody gets handed a confession.
Protected material is excluded at retrieval, so it never enters the coach's context at all. Telling a model to keep a secret is not a control. Not giving it the secret is. That distinction is what makes this defensible under Alberta privacy law and survivable in a dispute.