A transformation's north-star metric is the single number the program exists to move (a cost-per-unit figure, a customer-satisfaction score, a cycle time) stated with a target and a date. Left at that level alone, it is useless to everyone below the sponsor, because no individual workstream lead or frontline employee can act directly on a network-wide number. The KPI cascade is the discipline of decomposing that north star, level by level, until it lands on a metric a specific team can move this week, and the discipline only works if two rules are enforced without exception. First, the cascade must arithmetically reconcile: the sum of every workstream's stated contribution must equal the north-star target, exactly the way CN6.02's synergy tracker required every initiative's contribution to sum to the deal's total synergy case, that module applied this reconciliation discipline to one specific kind of target (M&A synergies); this module teaches it as the general requirement for any north star, financial or not. Second, every level needs a stated leading indicator, not just a lagging one. A lagging indicator (cost per case, this quarter's margin, the annual customer-satisfaction survey) confirms what already happened, often weeks or months after the fact, useful for accountability, useless for course-correction, because by the time it moves, the underlying behavior that caused it is long over. A leading indicator (this week's average handling time, this week's on-time-delivery rate, this week's process-adherence rate) moves before the lagging metric it predicts and can be seen, and acted on, while there is still time to change the outcome. Line of sight is the property that a leading indicator's movement causally, not just coincidentally, predicts the lagging metric several weeks out; a cascade without line of sight is a set of numbers on a dashboard that happen to be tracked, not a management tool. Worked Example 2 builds a full cascade, a network-wide cost target decomposed into three workstream contributions that reconcile exactly, with one workstream's leading indicator made explicit, and shows the arithmetic that connects a weekly operational number to an annual dollar figure a CFO cares about.
Advertisement
What is a vanity metric?
The failure mode worth naming here, because it is common and looks like diligence: the vanity metric, an activity count ("we ran twelve workshops," "we sent forty emails") reported as if it were progress, because it is easy to measure and always goes up, while carrying no demonstrated line of sight to the north star at all. A cascade with a vanity metric at any level is not a cascade with a weak link; it is a cascade that is silently disconnected from the number the program actually exists to move, and nobody will notice until the north star itself fails to move on schedule.
{
"skill": "Drafting a first-pass roadmap grid and KPI cascade skeleton from a transformation kickoff workshop's raw notes",
"use": "Have an AI tool turn a messy list of proposed initiatives and a stated north-star target into a structured two-axis roadmap draft (horizon x lane) and a matching KPI-cascade skeleton (workstream contributions, one candidate leading indicator each) that a PMO lead then corrects, re-sequences, and owns -- never a plan that ships as drafted.",
"tools": ["Chat assistants -- Claude, ChatGPT, Gemini", "Spreadsheet or project-planning tools for the actual Gantt/cascade build-out"],
"prompt": "Here are raw notes from a transformation kickoff workshop for <PROGRAM/INDUSTRY>, with a stated north-star target of <TARGET>: <NOTES>. Draft a two-axis roadmap (three time horizons x quick-wins/structural-moves/people-capability lanes) placing each proposed initiative, with an illustrative value contribution for each (do not invent real figures -- flag every number as a placeholder). Then draft a KPI-cascade table from the north star down through each workstream to one candidate leading indicator per workstream, and flag any workstream where you cannot suggest a plausible leading indicator.",
"verify": "Every placeholder value must be replaced with your own diligence-sourced number before this plan is used, and every dependency the tool omits -- especially a shared-resource/capacity dependency invisible on a simple list -- must be checked by asking each named specialist directly what else they are committed to. Recompute the cascade reconciliation yourself; do not trust an AI's claim that the workstream contributions already sum to the target (Guardrail 1-5).",
"diy": "You must already know how to build the two-axis roadmap, identify a capacity-driven dependency, and test a cascade's line of sight (this section) before you can tell a genuinely well-sequenced AI-drafted scaffold from one that merely looks organized. AI proposes a first cut; you dispose of whatever does not survive contact with the real team's actual constraints.",
"market": "IN"
}