KR Tree

1) Definition

A KR Tree is a hierarchical map that links your North Star Metric (NSM) to a small set of Objectives and their Key Results (KRs), showing how each KR (with owner, target, and data source) contributes to higher-level outcomes.

2) Why it matters (practical relevance)

• Causality, not chaos: Makes clear how today’s work moves the NSM.
• Focus: Trades feature debates for measurable deltas.
• Accountability: Each KR has an owner, a number, and a date.
• Decisions: Enables Scale/Iterate/Kill calls based on evidence, not opinion.

3) Core components / principles

• Levels: NSM → Objectives → KRs → Inputs/Levers → Bets/Initiatives, with Guardrails (SLOs, cost, security) alongside.
• Attributes per KR: Owner, baseline, target, time window, segment, data source, review cadence, confidence.
• Leading vs. lagging: KRs should include both direct outcome KRs and leading inputs you can influence.
• Quality bars: Count only successful value events (e.g., on-time, error-free).
• Small and sharp: 3–5 Objectives, each with 2–4 KRs. If it doesn’t fit on one page, it won’t be used.

4) How to build and use it (step-by-step)

1. Anchor the NSM: Confirm it reflects delivered customer value (and has a clear quality bar).
2. Set 3–5 Objectives: Plain-language statements of where you must improve to grow the NSM.
3. Define 2–4 KRs per Objective: Numeric, time-bound, with baseline → target and segment (e.g., new vs. existing).
4. Wire inputs/levers: For each KR, name the controllable drivers (activation rate, invite flow completion, p95 latency, etc.).
5. Map bets to KRs: Every initiative states the expected KR delta and a kill line.
6. Instrument & publish: One trusted dashboard with KR trends, error-budget burn, and cost.
7. Operate the cadence: Review weekly KR deltas; monthly SIK decisions; update targets quarterly.

5) Examples & analogies

Example A — Collaboration SaaS

• NSM: Weekly Active Teams sending ≥50 messages.
• Objective 1 (Activation): Get new teams to first value.
  • KR1: W1 teams reaching ≥50 msgs — baseline 22% → 35% this quarter.
  • KR2: % invited members who post within 24h — 48% → 60%.
• Objective 2 (Reliability): Remove friction in core flow.
  • KR3: p95 message send latency — 420ms → 250ms (guardrail with SLO).
• Inputs/levers: Setup wizard completion, invite CTA CTR, presence uptime.
• Bets: New invite UX (+6 pp to KR2), message pipeline tuning (−120ms to KR3).

Example B — Data/AI Automation Platform

• NSM: Weekly automated workflows executed on-SLA and accepted by customers.
• Objective 1 (Time-to-Value):
  • KR1: Median time to first accepted workflow — 3.2 days → 1.0 day.
• Objective 2 (Expansion):
  • KR2: Accounts with ≥3 active workflows — 18% → 35%.
• Objective 3 (Trust & Cost):
  • KR3: Error-budget burn ≤ 20%; unit cost per run − 25%.
• Bets: Templates gallery, data-source auto-mapping, model-latency caching, cost-aware routing.

Analogy:

• Tree: Trunk = NSM, big branches = Objectives, leaves = KRs, roots = Inputs/Levers; soil health = Guardrails.
• Scoreboard: NSM is the final score; KRs are the quarters and key stats that win the game.

6) Common mistakes to avoid

• Outputs disguised as KRs: (“Ship feature X”) instead of measurable outcome changes.
• Too many KRs: Dilutes focus; no one knows what matters.
• No baseline or data source: Targets become guesses; reviews become arguments.
• Ignoring quality/guardrails: Hitting numbers while burning SLOs or unit economics.
• Wrong level of aggregation: Measuring “users” when value is at the team/account level.
• No ownership or cadence: KR without a named owner and weekly update is theater.

7) Quick KR Tree checklist

• NSM defined with quality bar and clear data source
• 3–5 Objectives, each plain-language and NSM-linked
• 2–4 KRs per Objective, numeric, time-bound, with baseline → target and segment
• Inputs/Levers named for each KR (controllable drivers)
• Bets mapped with expected KR delta + kill line
• Guardrails visible (SLOs, cost, security) on the same dashboard
• Owners + cadence (weekly review, monthly SIK) assigned
• Single dashboard: NSM, KRs, inputs, guardrails—one source of truth

8) Actionable takeaways

1. Write KRs as outcomes, not deliverables—use baseline → target within a time window.
2. Limit scope: If your KR Tree doesn’t fit on one page, cut it until it does.
3. Attach every bet to a KR delta and declare a kill line before starting.
4. Review weekly with numbers only (NSM, KRs, guardrails)—then decide: Scale / Iterate / Kill.
5. Defend guardrails: Reliability and unit economics are first-class citizens, not footnotes.

Use this KR Tree to turn strategy into a weekly scoreboard that steers funding, focus, and speed.