Root-cause analysis with Gilbert’s Behavior Engineering Model
Thomas Gilbert’s Behavior Engineering Model sorts the forces that shape performance into six cells: three about the environment (information, resources, incentives) and three about the individual (knowledge, capacity, motives). Its central claim — that environmental factors usually outweigh individual ones — is exactly the test to run on a “we need feedback training” request.
Method note: each cell below states a Brightline cause with its evidence source from discovery (interviews, survey analysis, document review). All evidence is simulated, constructed for portfolio purposes.
The BEM table
The six cells use Gilbert’s own labels. The case carries four primary causes — E-2, P-1/P-2, S-1, M-4 (the same IDs the is-it-training verdict uses); supporting evidence is listed inside each cell rather than given its own ID.
Environment — the three environmental cells
| Information (what people need to know/see) | Resources (time, tools, processes) | Incentives (payoffs, consequences) |
|---|---|---|
| E-2, definition gap. No definition or expectation of regular one-on-ones; policy says only “encouraged”. Evidence: HR-P-014 document review; M3 never saw it; HR1 confirms it is untracked. P-1/P-2, the form. The annual review form collects ratings only, with an optional comments box — no place to record behavior, impact, or agreed actions. Evidence: form review; S1, M2; HR1. | E-2, primary. One-on-ones are not scheduled; no template; no rostered time; spans of 12–18 direct reports (median 13) make informal contact impossible. Evidence: M1, M3; org chart extract; document review; survey Findings 2 & 4. P-1/P-2, the cadence. No qualitative feedback point anywhere in the year; the annual review is the only formal one. Evidence: S1, M2; document review. | M-4. Scorecards measure throughput only (0% weight on team development); feedback work is invisible and unrewarded. Evidence: scorecard document review; M1, M2, M3; HR1. |
Individual — the three individual cells
| Knowledge (knowing how) | Capacity (being able) | Motives (what the person wants) |
|---|---|---|
| Cause S-1 — the one training can address. Some managers (especially the newest) don’t know how to structure a feedback conversation or open a hard one. Evidence: M1 (public, vague feedback; has avoided 1:1s); M2 — an 11-year veteran who equates task direction with feedback while his team rates useful feedback 2.1/5 (survey Finding 1: the 1.5-pt structure gap dwarfs the 0.3-pt tenure gap, Finding 3); HR1 (new managers get no people-management support). The evidence base is the thinnest of the four causes — one interview directly plus two corroborating sources — which is why the training slice it justifies is deliberately narrow. Also in this cell: new managers want support (M1 asked for a management course; waitlisted to next year) — a missing-support problem, really an environment gap, not a knowledge deficit. | No deficit (checked and ruled out). Managers can describe good feedback and want to do it. Spans create workload pressure (workload 3.25/5), which is logged as an environment issue above, not an individual-capacity one. Evidence: M2, M3, M1. | Motives face of M-4. Giving hard feedback carries social risk with no upside (no recognition; can conflict with throughput pressure). Evidence: M1 (“hard, they used to be my colleagues”); HR1; scorecard. |
Cause-ID mapping (for readers cross-referencing earlier drafts or the discovery artifacts): E/P-1 → E-2, definition gap; P-2 and P-1b → P-1/P-2 (form and cadence); S-2 → the support note inside the S-1 cell; C-0 → “no deficit (ruled out)”; M-4b → the motives face of M-4.
What the model says
- All three environmental cells carry a cause; the individual side carries exactly one narrow knowledge gap (S-1) plus a support need that is really a missing environment resource. That is the shape the model predicts for a “we need training” request: the environment outweighs the individual.
- In Gilbert’s terms, Brightline’s largest deficits are information (no expectations, no definitions), resources (no time, template, or process) and incentives (scorecard misalignment). The knowledge cell is real but small and specific: structuring and opening difficult feedback conversations — the new-manager cohort feels it most (35% of managers, promoted in the last 18 months).
- That pattern is why the recommendation in section 4 is a blend: fix the environment and process first, support the conversation with a job aid, and use a short scenario to build the one skill training can build. A 2-hour course would touch only the S-1 slice — roughly the difference between the 2.3 and 2.6 tenure means, not the 1.9-to-3.4 structure gap.
Alternative explanations considered and set aside
| Hypothesis | Evidence for | Verdict |
|---|---|---|
| “It’s the wave of new managers” (HR’s intake theory) | 35% promoted in 18 months; new managers slightly less confident | Partly. Tenure gap is 0.3 points vs 1.5 for structure; new-manager cadence rate ≈ veteran rate. Training only new managers would miss most teams. |
| “Staff don’t want feedback” / engagement is generational | No supporting evidence in any source | Set aside. Staff describe exactly what useful feedback looks like (S1’s clinical-educator story). |
| “It’s a workload problem — clinics are understaffed” | Workload 3.25/5, mild team-size correlation | Contributing, not causal. Feedback happens inside work, not after it; the constraint is structure and time allocation, not headcount. |
| “Cedar Point proves meetings don’t help” (nihilistic read) | Cedar: 4.3/5 meeting frequency, 1.5/5 usefulness | Inverted. Cedar is the strongest evidence that content (structure) matters more than frequency. |
Next: the Mager & Pipe flow runs the same evidence through a decision procedure, then is-it-training gives the verdict per cause.