LoS measures one thing: whether the people in a company see the same strategy their leaders see — and what the gap costs. The method comes from Line of Sight (Courser & Aries, 2023), built on research into why strategies fail in execution rather than design.
Twenty-nine questions, about nine minutes. Each answer is one of four — Aligned, Somewhat aligned, Misaligned, Not sure — which captures two things at once: what a person believes, and how sure they are. “Not sure” is a real answer, not a failure; confident wrong answers cost more than honest uncertainty, and the scoring treats them that way.
Every question rolls up to one of five conditions that execution research keeps finding behind strategies that stick:
Answers score on a 0–100 scale per key. Organizational health is a weighted score, not an average: each key counts in proportion to its measured contribution to overall line of sight in the validation study — Strategic Understanding .384, Activities & Structure .229, Balanced Metrics .201, Leadership .173, Human Capital .144 (Bullock, 2016, p. 18). Scores are computed separately for executives, managers, and everyone else — the gaps between those groups are usually the finding. Small groups are read with their size named, so a thin sample is always visible for what it is.
The instrument was independently validated in 2016 by Robert B. Bullock, Ph.D., against 338 survey responses across two samples. Overall line of sight explained 42% of the variation in how employees rate the quality of their company's products and services, 37% for innovation, 22% for job satisfaction, and 11% for intent to leave. The keys predict different things: Leadership and Human Capital relate most to turnover, Strategic Understanding to satisfaction, Activities & Structure to quality. The limits: the samples are modest, and the outcomes are employee perceptions, not financials. Read the validity report (PDF).
Peer numbers come in three grades, and each is labeled as what it is: government data first (for example, turnover from the Bureau of Labor Statistics JOLTS series), then published research with the source cited, then a model estimate marked low confidence when nothing better exists. We name a benchmark's population exactly — a figure for US managers is never dressed up as your industry.
Anonymity is structural, not promised. The survey stores a role group and a timestamp — no names, emails, IP addresses, or devices. There is nothing to join a response back to a person with, by anyone, including us. Free-text comments are summarized in themes and never quoted identifiably.
No “scientifically proven.” A survey measures what people believe at a point in time; the diagnosis is an argument from that evidence, and the pages always separate what was measured from what was inferred. Dollar figures come from arithmetic on your own numbers against the labeled benchmarks — shown with the work, never asserted in prose.