The People Analyst Guide to First, Break All the Rules. Format: what the book argues → what the research actually says → how you run it → the analysis you can run → the AI-era turn → what to do Monday. No reproduction of the book's text. The substance is ours; research anchors verified on read.
What the book argues
Key 2 is define the right outcomes — not the steps. The book's claim is that great managers tell people what to achieve and then get out of the way on how; standardizing the steps produces compliance and mediocrity, because each person's best route runs through their own pattern. Define the result, make it clear, and trust people to find their own path to it.
What the research actually says
Two robust literatures meet here and both support it, with boundaries. Goal-setting (Locke & Latham) is one of the most replicated findings in the field: specific, challenging outcome goals lift performance. And self-determination theory (Deci & Ryan) shows that autonomy over the how drives intrinsic motivation and sustained performance — scripting the steps removes exactly the autonomy that fuels the work. So "outcomes, not steps" is well grounded.
The boundaries a people analyst must keep. First, some work genuinely needs the steps — safety- critical procedures, regulated processes, and novices who don't yet have a route all benefit from standardization; "outcomes only" assumes competence and low interdependence. Second, and bigger: an outcome you can't measure cleanly becomes a number people game — Goodhart's law ("when a measure becomes a target, it stops being a good measure"), the same failure behind bad OKRs and the pay-for-performance caution. So the honest version: define clear, measurable outcomes; grant route autonomy; and watch the metric for gaming — outcomes manage for performance only if the outcome is real and the route is free.
How you run it
Specify the outcome and its measure, hand over the how, and instrument for two things: is the outcome clear and aligned (do people agree on what "good" is — the Nine Lies cascade point), and is it being gamed (is the proxy diverging from the real goal). Manage the outcome, not the keystrokes.
The analysis you can run
An outcome / alignment analysis — leadership-quality (alignment: do priorities and definitions of
"good" actually agree) with survey-orchestrator (autonomy / CAMS) — that checks whether outcomes are
clear and shared and whether people have the route-autonomy that makes outcome-management work. (Braids
Nine Lies Lie 3 "cascade meaning" and Work Rules Ch 7 on OKRs/calibration.)
The AI-era turn
AI makes continuous outcome-tracking cheap, which strengthens this key — less reason to manage the steps when you can see the result in near-real-time. But it sharpens the Goodhart risk: an AI optimizing a proxy metric will game it faster and more thoroughly than any human, and an AI that scripts "next best action" can quietly re-impose the steps. Measure the real outcome, keep the route human and free, and watch the proxy.
What to do Monday
- Define outcomes and their measures; hand over the how — script steps only where safety/regulation/novice-status demands it.
- Check the outcome is clear and aligned (do people agree what "good" is) before trusting it.
- Watch for gaming — Goodhart's law; if the proxy diverges from the goal, fix the measure.
- Use AI to track outcomes, not to re-script the steps; guard the proxy.