okr-cascade ยท diff
git:20260718.7c1285c to git:20260728.fac4544
17 added, 15 removed. Audit A to A.
---
name: okr-cascade
description: Cascade OKRs so team and individual goals ladder up to company strategy without sandbagged targets or scoring that rewards easy wins. Use when setting quarterly or annual objectives across more than one team and you want alignment, not a spreadsheet of restated tasks.
---
# OKR cascade
OKRs (objectives and key results) fail less because the format is hard than
because the incentives around them are: people set targets they know they will
hit, dress activity up as outcomes, and score generously so the numbers look
good. A cascade done well aligns effort to strategy. Done as theater, it
produces a quarterly ritual everyone quietly games.
## Method
1. **Write objectives as outcomes, key results as measures.** The objective is a
- short qualitative statement of where you want to be ("new users reach value in
- one session"). Each key result is a metric with a number ("cut median
- time-to-first-action from 9 minutes to 3"). A key result you either did or did
- not do is a checklist item, not a result.
+ short qualitative statement of where you want to be ("new users reach value
+ in one session"). Each key result is a metric with a number ("cut median
+ time-to-first-action from 9 minutes to 3"). A key result you either did or
+ did not do is a checklist item, not a result.
2. **Align, do not literally cascade.** A team's OKRs should name which company
objective they serve, but they are not the parent's key results copied down a
level. Let teams express their contribution in their own terms, and combine
top-down direction with bottom-up proposals so the people doing the work own
the target.
- 3. **Separate committed from aspirational.** Mark each OKR as committed (you will
- deliver it, a score near 1.0 expected) or aspirational (a stretch where 0.7 is
- a strong result, per Google's convention). Mixing the two silently is how a
- missed stretch reads as failure and a sandbagged commitment reads as triumph.
+ 3. **Separate committed from aspirational.** Mark each OKR as committed (you
+ will deliver it, a score near 1.0 expected) or aspirational (a stretch where
+ 0.7 is a strong result, per Google's convention). Mixing the two silently is
+ how a missed stretch reads as failure and a sandbagged commitment reads as
+ triumph.
4. **Kill sandbagging at review.** When every proposed key result looks certain
- to score 1.0, push back: aspirational targets that always hit full are set too
- low. Ask what would have to be true to reach 30% further, and calibrate
+ to score 1.0, push back: aspirational targets that always hit full are set
+ too low. Ask what would have to be true to reach 30% further, and calibrate
targets across peer teams so no one wins by aiming low.
5. **Cap the count.** Three to five objectives per level, each with about three
key results. A team carrying fifteen OKRs has a to-do list, not a set of
priorities, and nothing has been chosen over anything else.
6. **Score honestly and decouple from ratings.** Grade key results 0.0 to 1.0 at
- cycle end on the evidence, and read a 0.6 on a genuine stretch as healthy, not
- a miss. Tie OKR scores straight to compensation and you guarantee sandbagging
- next cycle, so keep scoring a planning signal, not a performance verdict.
+ cycle end on the evidence, and read a 0.6 on a genuine stretch as healthy,
+ not a miss. Tie OKR scores straight to compensation and you guarantee
+ sandbagging next cycle, so keep scoring a planning signal, not a performance
+ verdict.
## Litmus tests
- Pick any team OKR: can you name the company objective it serves and how it
moves that needle?
- If every key result is tracking toward 1.0, were the targets real or safe?
- Does a low score trigger a learning conversation, or a penalty that teaches
people to aim lower next time?
## Boundaries
OKRs are a goal-setting and alignment tool, not a project plan or a
performance-review instrument: the moment scores feed directly into ratings,
honesty collapses, so hold that line. Conventions vary (Google's 0.7 aspiration,
- Intel's origins, committed-only shops), so follow your organization's cadence and
- scale, and route individual promotion evidence to the promo-packet skill.
+ Intel's origins, committed-only shops), so follow your organization's cadence
+ and scale, and route individual promotion evidence to the promo-packet skill.