The dilemma
By mid-cycle, Quinn saw three versions of truth about the same people: dashboards, manager narratives, and final ratings. When the stakes rose, alignment broke. Stories outweighed proof. Decisions lost the plot.
What was fractured
- Signal
Evidence came in uneven slices across quarters and projects. - Sense-making
Recency, halo, and likability steered judgments. - Scale
Calibration rewarded the loudest voice, not the clearest case. - Stakes
Ratings drove pay and promotions. Bias compounded into outcomes.
Quinn’s playbook: Repair the review system
- Evidence packets, not vibes
One page per person. Three outcomes with links. Two growth deltas. One forward bet. - Behaviorally anchored rubrics
Define “meets” and “exceeds” with level-specific examples. No free-form labels. - Two-pass calibration
Pass 1 silent pre-reads. Pass 2 live debate with a neutral moderator and time limits. - Decision logs
Each rating gets an owner, rationale, and evidence ID. Logged within 24 hours. - Bias interrupters
Rotate first-speak. Cap stories at 120 seconds. Require one counterexample before a rating locks. - Opportunity check
Map project scope and visibility against ratings to surface allocation bias. - Comp alignment guardrails
Tie ratings to band movements with defined ranges to prevent drift. - Growth contracts
Close with a 90-day skill goal, named mentor, and one resource. - Quality audit
Sample files after the cycle. If evidence is thin, reopen the rating.
The shift at SoulCode
After one cycle, calibrations ran faster, appeals dropped, and comp decisions traced cleanly back to work. Managers spent less time defending and more time developing.
Leadership takeaways
Clarity beats charisma.
Proof beats memory.
Decisions need receipts.
Teaser for Episode 25: The Autonomy Audit
Quinn tests where freedom fuels ownership and where it quietly erodes it.
.png)
