Originally published October 2010. Updated June 2026.
Gallup's State of the Global Workplace 2024 placed U.S. employee engagement at 33%. The 2025 update showed engagement holding at the same level after a brief 2022 spike during the post-pandemic talent realignment. Roughly two-thirds of U.S. workers describe themselves as not engaged or actively disengaged at work. The single largest variable predicting engagement, per Gallup's longitudinal data covering decades, is the immediate manager. The manager accounts for roughly 70% of the variance in team engagement scores. The manager is the unit of organizational performance.
The 2010 piece this URL originally covered was about National Boss's Day. The 15-year follow-up is a serious question: with two-thirds of U.S. workers disengaged and the manager-engagement relationship now exhaustively documented, what does the data tell us about what actually makes a great manager?
The five traits that survive the data
Gallup's longitudinal management research, the Google Project Oxygen findings, the McKinsey/LeanIn Women in the Workplace data, and the broader academic management literature converge on five traits.
First, clear expectations and consistent feedback. Workers who understand what is expected of them produce engagement scores roughly double the average. Managers who set ambiguous expectations or who fail to provide consistent feedback produce engagement scores roughly half the average. The variance is enormous and the cause is operationally simple — most managers do not invest in clarity.
Second, individual development. Managers who actively invest in their direct reports' growth — through stretch assignments, structured learning, and clear development paths — produce retention rates significantly above the average. The McKinsey data places the retention differential at 2x to 3x depending on tenure cohort.
Third, psychological safety. Google's Project Oxygen and Amy Edmondson's broader academic work both established psychological safety (the ability to raise concerns, surface mistakes, and disagree without retaliation) as the single most consequential team-effectiveness variable. Managers who establish high psychological safety produce teams that ship better products, recover from failures faster, and retain talent at higher rates.
Fourth, operational competence in the team's actual work. The "manager who couldn't do the team's job" pattern produces measurable engagement decline. Managers who maintain credible operational competence in the discipline they manage — not at the same execution level, but at the same understanding level — produce higher engagement and lower attrition than managers who operate purely in the management layer.
Fifth, advocacy upward. Managers who advocate effectively for their teams within the broader organization — for resources, for credit, for protection from organizational dysfunction — produce engagement and retention outcomes that exceed managers who are equally competent at the four traits above but who do not advocate upward. The upward advocacy is the most underweighted trait in management research.
What the data also says
Three findings that complicate the simple "great managers are made, not born" narrative.
First, management capability has measurable selection effects. The Gallup data places roughly 10% of managers in the top tier on the five-trait composite. Roughly 50% sit in a middle tier that performs at the organizational average. Roughly 40% sit in a bottom tier that actively degrades team performance. The distribution is not a smooth bell curve — there is a meaningful tail of managers whose presence costs the organization more than their replacement would.
Second, the management role has expanded in scope. The 2025 manager increasingly handles direct reports across multiple time zones, multiple legal jurisdictions, hybrid and remote-distributed work patterns, and multiple generational cohorts simultaneously. The management role in 2010 typically supervised 7 to 10 direct reports in a single office. The 2025 manager often supervises the same number across three or four different operational contexts. The complexity has increased meaningfully even as the underlying traits have remained constant.
Third, AI tools have shifted the manager's time allocation. Managers in 2026 spend less time on coordination overhead (meeting scheduling, status updates, basic written communication — the AI tools handle increasing amounts of this) and more time on the strategic-and-judgment work that produces team-effectiveness outcomes. The managers who have adopted the AI tools effectively report substantially higher capacity for the strategic work. The managers who have not adopted the tools are increasingly the bottom-tier 40%.
The pipeline implication
The "broken rung" data covered in The Women Running Corporate America applies at this level. The promotion from individual contributor to first-line manager is the single most consequential organizational decision in most organizations, and it is also the promotion most organizations get wrong. Most first-line management promotions are based on individual-contributor excellence, not on the five traits above. The IC-to-manager mismatch is the structural cause of the bottom-tier 40% of managers.
The organizational fix is straightforward in concept and difficult in execution — select for the five traits, train against the five traits, measure on the five traits, and remove managers from the role when the measurement persistently shows poor outcomes. Most large organizations do some version of this rhetorically and few do it operationally.
What this means for talent communications and employer brand
Three operating implications.
First, the management quality is now part of the employer brand. Glassdoor, Blind, LinkedIn, and the AI-engine layer all index manager-quality signals (specific manager reviews, departure patterns, internal-network conversations) at a level of granularity that did not exist a decade ago. The brand-level employer reputation is increasingly downstream of the manager-level operational reality.
Second, the AI-engine entity description for employers now includes management-quality signal. ChatGPT, Claude, Gemini, and Perplexity synthesize answers about "what is it like to work at X" that draw on the granular review data plus the broader media coverage. The employer brand is now a retrievable AI-engine surface, not just a press-and-careers-page surface. The audience ownership shift applies to employer brand — owned communications matter, but the retrievable AI-engine answer is the durable surface.
Third, the manager-development investment is now a communications-relevant capital allocation. Organizations that visibly invest in management development (Google's Project Oxygen, Microsoft's Manager Excellence framework, the Bain/McKinsey/BCG structured-manager-training programs) produce measurable employer-brand outcomes. Organizations that under-invest produce the opposite. The capital allocation is increasingly visible to the audience.
