Hell is other people: performance management at Big Tech
Performance management systems in large tech companies are portrayed as opaque, politicized processes that often reward self-promotion and project luck more than actual contribution. Commenters describe stack ranking, recency bias, legal threats, non-solicit and NDA clauses, and weak feedback as creating perverse incentives: engineers hide future plans, optimize around review cycles, and focus on visible wins over unglamorous but critical work. Some argue these systems function less as objective performance measurement and more as tools of managerial control, with calls for behaviorally informed alternatives and simpler, low-drama processes for steady, competent employees.
Job transitions, legal threats, and NDAs
- Multiple anecdotes of employees moving between large tech firms and being threatened with lawsuits or subjected to invasive audits, especially around alleged “solicitation” or “bleeding talent.”
- Settlement agreements often include non-disparagement and confidentiality, which posters say are de facto enforced by fear of costly litigation, even if some clauses may be legally shaky.
- Some note recent US/California changes limiting NDAs around workplace mistreatment and NLRB scrutiny of separation-agreement NDAs, but practical impact is unclear.
Non-solicit, non-compete, and disclosure norms
- Strong advice: never disclose your next employer before you’ve left and the new offer is fully signed; avoid detailed exit interviews or give only bland positives.
- Reports that HR and managers aggressively push for destination details; others counter you can simply refuse, especially for optional exit interviews.
- Claims that non-solicits are generally unenforceable in California/Illinois but enforced elsewhere; others emphasize they’re “no joke” regardless.
Stack ranking and performance systems
- Historical roots traced to GE and Microsoft; distinctions made between stack ranking and peer review.
- Many argue forced curves distort teamwork, misalign incentives, and create perverse internal competition.
- Defenders say without forced differentiation, managers systematically avoid firing or downgrading anyone, leading to mediocrity at scale.
Biases, self‑promotion, and introversion
- Strong sentiment that evaluations reward perception, talkativeness, and self-promotion over actual output.
- Reference to cognitive biases and studies where talkative participants were rated more intelligent, independent of true ability.
- Concerns that classic “quiet, technically strong” engineers are now penalized, while “impact inflation” and credit-stealing thrive.
Limits of performance reviews and measurement
- Many describe performance processes as theater: low inter-rater reliability, strong recency bias, and heavy dependence on project luck.
- Some see them primarily as tools of control/discipline and legal cover for firing, not genuine performance measurement.
- Difficulty tying individual engineering work to revenue or clear impact is repeatedly highlighted; comparisons with professions like finance, law, and medicine underscore this gap.
Manager quality, incentives, and corporate goals
- Posters distinguish “good managers” from “real leaders,” claiming most optimize for survival and pleasing upper management/shareholders.
- Acknowledgement that managers usually know their best and worst performers in small teams, but can’t transparently justify rankings, hence complex formal systems.
- Broader critique that Big Tech’s goals center on shareholder value, monopoly power, and lucrative (including military/AI) contracts, undercutting earlier “don’t be evil” narratives.
Alternative approaches and ideas
- Suggestions include:
- Minimal reviews for “steady-state” employees (“keep up the good work” plus inflation adjustment).
- Anonymous peer feedback alongside manager reviews, leveraging I/O psychology findings.
- Systems where high performers gain more autonomy and internal “startup” authority rather than people-management roles.
- Panel-based review combining self-appraisal, manager evaluation, and cross-team managers to mitigate direct bias.
- Some small-company experiments (e.g., peer-allocated bonus pools) are mentioned as brutally honest but informative.