HDI Methodology

How does the HDI methodology work?

HDI measures human dependency the way it actually shows up — in behaviour, not in self-image. Instead of asking "how automated are you?" (which invites optimism), the methodology asks what happens in concrete situations: when the key person is away, when a problem appears, when a document needs approval, when knowledge has to transfer. Each answer maps to a behavioural anchor with a known dependency level, and the anchors aggregate into three layer scores and one overall 0–100 index. The exact weights and formulas are proprietary; this page describes what is measured and why it can be trusted.

What exactly is measured?

Three layers, each through its own block of behavioural questions:

• Coordination — how work travels: the number of manual approvals in a typical task, what a coordinator's absence does to throughput, how cross-department coordination happens, how much of the day is spent moving work rather than doing it.
• Control — how problems surface: whether detection is systemic or human, how incidents are discovered and escalated, how long detection-to-closure takes, how much review is manual re-checking.
• Cognitive — where knowledge lives: single-expert steps, undocumented procedures, how long onboarding to independence takes, how much reporting is assembled by hand from memory and scattered sources.

On top of the questionnaire, a small set of industry KPIs anchors the answers in operational reality — for cell & gene therapy, for example: deviation closure time, batch-record error rate, chain-of-identity documentation lag. Each industry pack replaces these with its own indicators and thresholds.

Why industry weighting?

The same behaviour carries different risk in different operations. A single person doing manual batch-record review is a moderate issue in a stable small-molecule plant and a critical one in cell & gene therapy, where every batch is a patient. The methodology therefore weights the three layers differently per industry — the relative importance of coordination, control and cognitive dependency is calibrated to where each sector's operational risk actually concentrates.

What is the confidence score?

Every screening result carries a confidence percentage. Self-assessments have known failure modes — questions skipped, contradictory answers, uniformly optimistic patterns — and the confidence score reflects how internally consistent and complete the responses are. A result of "HDI 93, confidence 83%" tells you both the finding and how firmly to hold it. Low-confidence results are exactly what the Phase 2 workshop is designed to resolve.

Why two phases?

Because diagnosis and interpretation are different jobs. Phase 1 (free, ~10 minutes, self-service) locates the dependency: score, layers, process heatmap, hours-leak. It deliberately contains no recommendations — a self-assessment can say what and where, but not reliably why. Phase 2 (the paid audit) is built around a structured workshop with the client's key people, scored by an auditor against behavioural anchors. It triangulates what people say in the room against what the Phase 1 self-assessment claimed — the gap between the two is itself a finding (the perception gap) — and turns the diagnosis into decoded findings, prioritised automation entry points, quantified losses on agreed assumptions, and an implementation plan.

Why isn't money in the free result?

The free screening reports the operational leak in hours and FTE, never currency. Converting hours into money requires cost assumptions — loaded rates, overhead treatment — that are meaningless unless agreed with the client. Publishing a headline dollar figure from a 10-minute self-assessment would be theatre; agreeing assumptions and quantifying properly is Phase 2 work. This is a deliberate credibility decision, not a teaser.

Where does the methodology come from?

HDI was developed from operational practice in regulated life sciences — training, quality and manufacturing operations in cell & gene therapy and conventional pharma — where human dependency has regulatory consequences and is therefore unusually visible. The behavioural-anchor approach borrows from structured interviewing and competency assessment; the industry calibration comes from field use. Benchmarks from anonymised, aggregated audits are published on the benchmarks page as the dataset grows — each published figure requires at least three independent audits behind it.