daddia Realisation Review
When to use
End-of-window planned-vs-actual portfolio review — optimism-bias calibration across initiatives.
What this skill does not do
- Does not re-litigate LAPSED reasons — carry forward from
benefits-recovery. - Does not blend benefit types in one headline cash figure.
Preconditions
| Input | If missing |
|---|---|
| Full register + tracking history | Ask |
| Prior realisation reviews (optional) | Skip optimism-bias cross-review |
Provisional mode
Single-initiative review without portfolio history — note optimism-bias check limited.
Trust spine
Narrative-synthesis BLUF; root-cause cluster diagnosis; calibration discount for future cases.
Workflow
Read the practice profile —
~/.claude/plugins/config/claude-for-strategy/org-profile.mdand~/.claude/plugins/config/claude-for-strategy/value-realisation/CLAUDE.md— the full benefits register, the complete tracking history, and anybenefits-recoverydecisions for benefits in the window under review.Score each benefit: REALISED IN FULL / PARTIALLY REALISED (state the percentage) / NOT REALISED / LAPSED (carry forward the specific reason recorded in
benefits-recovery, don't re-litigate it here).Roll up to portfolio level — total planned value vs. total realised value, split by type. Report cash-releasing, cash-releasable, and non-cash/qualitative separately; don't blend them into one headline figure. If any cash-releasable benefit was treated as guaranteed cash anywhere in the original case or subsequent reporting, call that out explicitly here — it's a common and material distortion in how a portfolio's "savings" get represented upward.
Run the optimism-bias check. If prior
realisation-reviewruns exist for other initiatives, compare this initiative's realisation rate to the pattern across those reviews. A single underperforming initiative may just be that initiative's story; a consistent realisation rate well below 100% across multiple independent reviews is a finding about how this organization's business cases get sized, and should be named as such rather than absorbed case-by-case. If a pattern emerges, state a suggested calibration discount for future business case estimates and flag it as a direct input totransformation:business-caseorcorporate-strategy:evaluate-strategic-option.Diagnose what should change about how benefits get planned, specifically — not a generic "do better" conclusion. If
benefits-recovery's root-cause classifications across this portfolio cluster on NOT EMBEDDED, the gap is in adoption planning, not estimation, and the fix belongs in how business changes get resourced, not in how numbers get forecast. If they cluster on THEORY PROBLEM, the gap is in how causal assumptions get tested before a case is approved — say which one it actually was, since the two point to entirely different fixes.Output a BLUF-structured PIR suitable as direct input to
consulting:exec-memoordeck-outlinefor a steering-committee version, if that plugin is installed.
Output format
RECOMMENDATION: [one to two sentences — overall verdict on this initiative's value delivery]
PORTFOLIO ROLL-UP:
| Benefit type | Planned | Realised | % |
|---|---|---|---|
| Cash-releasing | ... | ... | ... |
| Cash-releasable | ... | ... | ... |
| Non-cash/qualitative | ... | ... | ... |
CASH-RELEASABLE-AS-CASH FLAGS: [list, or "none"]
PER-BENEFIT SCORES:
[name]: [REALISED IN FULL | PARTIALLY REALISED — X% | NOT REALISED | LAPSED — reason: ...]
[repeat]
OPTIMISM-BIAS CHECK: [no prior reviews to compare] or
[pattern across N reviews: average realisation rate X%; suggested calibration
discount for future cases: ...]
ROOT-CAUSE CLUSTER (from benefits-recovery, this portfolio): [dominant cause(s) — what
this implies should change in HOW benefits get planned next time]
Worked example
Input: 60% realisation; three prior reviews averaged 55%.
Expected output: OPTIMISM-BIAS CHECK: pattern across 4 reviews; suggested calibration discount for future cases [review].
Quality checks before delivering
- Types reported separately in roll-up
- Optimism-bias check run when history exists
- Root-cause cluster named with planning implication
- BLUF recommendation first
Propose profile update
When a stable convention surfaces during this run (thresholds, naming, tone, output format, or recurring corrections), propose a profile update: show the exact diff against ~/.claude/plugins/config/claude-for-strategy/value-realisation/CLAUDE.md (org-wide facts go to org-profile.md), ask for confirmation, and write only on yes. Only /value-realisation:practice-setup auto-applies a full profile write.
Outputs
Follows plugin CLAUDE.md § Outputs. Next: transformation:business-case calibration, consulting:exec-memo, or steering PIR.