Three pieces of news from the last four months frame the state of work in 2024.

In June, Gallup's State of the Global Workplace report found that only 23% of employees worldwide are engaged at work. Another 62% are "not engaged", doing what is required and little more, and 15% are actively disengaged. Gallup estimates the cost of low engagement at 8.9 trillion dollars, about 9% of global GDP. Forty-one percent of employees said they experienced a lot of stress the previous day, and one in five reported daily loneliness.

In May, Microsoft and LinkedIn's Work Trend Index reported that 75% of knowledge workers already use generative AI at work, and that 78% of those users bring their own tools rather than waiting for their employer to provide them.

And last week, on September 16, Amazon CEO Andy Jassy told corporate employees they will be expected in the office five days a week from January 2, 2025, alongside a plan to increase the ratio of individual contributors to managers and remove layers.

Underneath all three is the oldest question in management: what makes people want to do good work, and what makes a group of them perform better than the sum of its parts? The research answers are more stable than the headlines. This post collects them and then looks at how this year's shifts test them.

1. Engagement, motivation, performance: getting the words right

Term Meaning Measured by
Motivation The reasons and energy behind behaviour: why someone starts, persists and puts in effort Self-report scales, behaviour (persistence, voluntary effort)
Intrinsic motivation Doing something because it is interesting, satisfying or meaningful in itself Free-choice behaviour, interest ratings
Extrinsic motivation Doing something for a separable outcome: pay, approval, avoiding punishment Same scales, different reasons
Engagement Gallup's definition: involvement in and enthusiasm for work and workplace Gallup's Q12 survey items, such as "I know what is expected of me at work"
Team performance Output quality and quantity, plus the team's capacity to keep performing and the growth of its members Outcomes, delivery metrics, peer and stakeholder ratings

Engagement surveys measure the conditions for motivation, not motivation directly. That is a strength: Gallup's 12 items (clear expectations, the right materials, the chance to do your best every day, recognition, someone who cares about you, someone who encourages your development, your opinions count, a mission that makes your job feel important, colleagues committed to quality, a best friend at work, recent talk about progress, opportunities to learn and grow) read like a checklist of things a manager can actually change.

2. Self-determination theory: three needs

The best-supported general theory of motivation is self-determination theory (SDT), developed by Edward Deci and Richard Ryan since the 1970s and tested in hundreds of studies across work, education, health and sport. It proposes that people do their best and most sustained work when three basic psychological needs are met:

Need Meaning At work it looks like Common ways managers frustrate it
Autonomy Feeling that you endorse what you are doing, not that you are being controlled Choice over how to do the work, a voice in what to do, understanding the why Micromanagement, unexplained orders, surveillance
Competence Feeling effective and growing in skill Challenging but achievable work, useful feedback, learning Constant firefighting, no feedback, impossible goals, never-ending trivial work
Relatedness Feeling connected to and cared about by others Trust, belonging, colleagues you like Isolation, internal competition, treating people as interchangeable

Daniel Pink popularised a version of this in Drive (2009) as autonomy, mastery and purpose. Purpose is not one of SDT's three needs, but SDT research shows that people who understand and endorse the value of a task can be highly motivated even when the task is not interesting in itself. That is the manager's real job with boring work: explain why it matters, honestly, and let people own the how.

The reward debate, settled enough to use

Do rewards undermine intrinsic motivation? Deci, Richard Koestner and Ryan's 1999 meta-analysis of 128 experiments found that expected, tangible rewards contingent on doing an interesting task reduce intrinsic motivation on average. Verbal feedback that conveys competence tends to increase it.

That does not mean pay does not matter. It means:

  • Pay fairly and transparently, then take money off the table. Unfair pay is a powerful demotivator. Beyond fairness, bonuses tied to narrow metrics tend to buy the metric, not the behaviour you wanted (Goodhart's law in practice).
  • Use rewards that feel informational, not controlling. Recognition for specific good work says "you are competent". A points system for pull requests says "we are watching your numbers".
  • Rewards work better for routine, well-defined tasks. For creative and complex work, controlling incentives are riskier.

3. Herzberg: hygiene factors are not optional

In 1959, Frederick Herzberg and colleagues proposed that the things that make people dissatisfied at work are different from the things that make them motivated.

  • Hygiene factors (pay, policies, working conditions, job security, relationships with supervisors): when bad, they cause dissatisfaction. When good, they remove it, but they do not create motivation.
  • Motivators (achievement, recognition, the work itself, responsibility, growth): these create motivation.

The two-factor theory's methodology has been criticised, but its practical message survives: you cannot inspire people out of a broken environment. If the build takes 40 minutes, on-call is brutal and salaries lag the market, talking about mission will sound hollow. Fix the hygiene first.

4. The progress principle: the small wins that drive inner work life

The most practical research on day-to-day motivation comes from Teresa Amabile and Steven Kramer, who analysed nearly 12,000 daily diary entries from 238 professionals in 26 project teams across seven companies. Their book, The Progress Principle (2011), concluded:

Of all the things that can boost emotions, motivation, and perceptions during a workday, the single most important is making progress in meaningful work.

The flip side was equally strong: setbacks had a larger negative effect on people's inner work life than progress had a positive one. And the things that most blocked progress were often managerial: shifting priorities, unclear goals, being pulled off work, ideas being dismissed.

Amabile and Kramer also surveyed managers on what they thought motivated employees most. "Supporting progress" came last of five options. Most managers ranked recognition first.

What this means for managers: your highest-leverage activity is removing obstacles from meaningful work and making progress visible. Break work into pieces small enough to finish. Protect people from churn. End weekly meetings with what moved, not only what is late.

5. What makes teams perform

Individual motivation is necessary but not sufficient. Teams have their own dynamics.

Project Aristotle

From 2012, Google's People Analytics group studied 180 internal teams to find what distinguished effective ones. They expected the answer to be the right mix of skills and personalities. It was not. Who was on the team mattered less than how the team worked together. Their 2015 findings ranked five dynamics:

  1. Psychological safety: team members feel safe to take risks and be vulnerable in front of each other.
  2. Dependability: people get things done on time and to a high standard.
  3. Structure and clarity: clear roles, plans and goals.
  4. Meaning: the work is personally important to team members.
  5. Impact: people believe their work matters and creates change.

Psychological safety was by far the most important, and the one that enabled the others.

Psychological safety, precisely

Amy Edmondson introduced team psychological safety in a 1999 study of work teams, defining it as "a shared belief that the team is safe for interpersonal risk taking". Her earlier work on hospital teams had found, surprisingly, that better teams reported more errors. They were not making more mistakes; they were more willing to talk about them.

Psychological safety is often misunderstood as niceness or comfort. Edmondson is explicit that it is compatible with high standards. The combination of high safety and high standards produces learning and performance. High safety with low standards produces comfort. Low safety with high standards produces anxiety, and silence about problems.

Collective intelligence

Anita Woolley and colleagues (Science, 2010) found that groups have a measurable "collective intelligence" that predicts performance across varied tasks. It was not strongly correlated with the average or maximum intelligence of members. It was correlated with:

  • the average social sensitivity of members (measured by reading emotions from photos of eyes),
  • equality of conversational turn-taking: groups dominated by a few voices did worse,
  • the proportion of women in the group, largely explained by social sensitivity.

Hackman's conditions

J. Richard Hackman, after decades studying teams from orchestras to airline crews, argued that leaders can rarely cause great performance directly, but they can create the conditions for it: a real team (clear membership, stable over time), a compelling direction, an enabling structure (the right size, the right tasks, clear norms), a supportive context (resources, information, rewards for team performance) and expert coaching at the right moments. His recurring warning: teams are usually too big. Beyond about 6 to 8 people, coordination costs grow faster than capacity.

6. How 2024's shifts test the evidence

The return to office

The debate about where people work has hardened into mandates. Amazon's move to five days is the most prominent this year; others have tightened hybrid rules. Jassy's stated reasons are cultural: it is easier to "learn, model, practice, and strengthen our culture", collaborate and invent together in person.

The evidence is more mixed than either side admits.

  • Hybrid can retain people without hurting performance. In June, Nicholas Bloom, Ruobing Han and James Liang published a randomised trial in Nature: 1,612 Trip.com employees were assigned to either five office days or two days at home. Over six months, the hybrid group's quit rate fell by a third, with no difference in performance reviews or promotions over the following two years.
  • Fully remote has real costs for some work. Studies of mentoring and code review suggest junior employees get less feedback when no one is co-located. Relatedness and learning are harder at a distance.
  • Mandates are an autonomy question too. Viewed through SDT, a top-down rule that removes a valued choice without a convincing explanation frustrates autonomy, whatever its merits. The likely result is not a revolt but quiet disengagement or departures among those with the most options.

For managers who do not set the policy, the practical move is to make in-person time worth the commute: concentrate it on collaboration, mentoring and relationships, and do not fill it with video calls people could have taken at home.

AI at work

The Work Trend Index numbers say that AI adoption is happening bottom-up, often without guidance. That has motivational consequences.

  • Competence: AI can remove drudgery and make people feel more capable, or it can make them feel deskilled and replaceable. Which one depends on how it is introduced.
  • Autonomy: employees bringing their own tools is a sign of autonomy, and also of a vacuum. Teams need clear rules on what data can go where, framed as enabling rather than forbidding.
  • Relatedness and meaning: if AI absorbs the tasks people found meaningful and leaves them the checking and approving, motivation will fall even if output rises.

A useful default: let teams experiment openly, share what works, protect learning time, and be honest about what AI means for roles. Uncertainty about job security is a hygiene factor; leaving it unaddressed corrodes everything else.

Flatter organisations, busier managers

Amazon is not alone in removing management layers. Gallup has long reported that managers account for about 70% of the variance in team engagement. Flattening may remove bureaucracy, but it also raises span of control. A manager with 15 reports cannot hold weekly one-on-ones, notice early burnout or coach each person. If you are going flat, invest in the managers who remain, and consider senior individual contributors as coaches.

7. A manager's playbook

Translated into weekly habits, the research points to a short list.

Autonomy

  • [ ] Explain the why behind every significant request. If you cannot, question the request.
  • [ ] Agree on outcomes and constraints; let people choose the approach.
  • [ ] Ask for input before decisions that affect the team, and say what you did with it.

Competence

  • [ ] Give specific feedback weekly, positive and corrective. Specific beats frequent beats annual.
  • [ ] Match people to stretch work at the edge of their ability, with support.
  • [ ] Protect focus time. Constant interruption is a competence killer.

Relatedness and safety

  • [ ] Hold regular one-on-ones that are about the person, not just the status.
  • [ ] Respond to bad news with curiosity, not blame. The first reaction you give to a mistake sets the team's reporting rate.
  • [ ] Make sure everyone speaks in meetings, not just the confident few.
  • [ ] Admit your own mistakes out loud.

Progress

  • [ ] Break work into pieces that can be finished in days, not months.
  • [ ] Remove one obstacle every week, and say which one.
  • [ ] Minimise priority churn. If you must change direction, explain why and acknowledge the lost work.

Hygiene and context

  • [ ] Fight for fair pay, reasonable on-call, working tools and clear career paths.
  • [ ] Keep teams small and stable.
  • [ ] Be honest about uncertainty: return-to-office rules, reorganisations, AI's effect on roles.

8. Myths worth retiring

Myth What the evidence says
"People are motivated mainly by money." Pay fairness is essential, but beyond fairness, autonomy, progress and relationships drive sustained effort
"Great teams are made of great individuals." How a team interacts predicts performance better than individual talent alone
"Psychological safety means being nice." It means people can speak up. It works best with high standards
"Engagement is the employee's responsibility." Most engagement drivers are under the manager's control
"Gamification and leaderboards boost motivation." They can boost short-term activity on routine tasks; they often undermine intrinsic motivation and cooperation on complex work
"Remote work kills productivity" / "Office work is pointless." Effects depend on the kind of work, the person's career stage and how the time is used

Why this matters

  • Engagement is low and stuck. Gallup's 23% means most people are doing their jobs without much enthusiasm, and the cost is measured in trillions.
  • The fundamentals are stable. Autonomy, competence, relatedness and progress explain most of what managers can influence.
  • Teams succeed on how they work together. Psychological safety, clarity, equal voice and meaningful work beat star line-ups.
  • This year's big changes are motivational experiments. Office mandates, AI adoption and flatter structures each touch autonomy, competence or relatedness. Leaders who manage them with that lens will lose fewer good people.

Final Thoughts

Every few years, work changes shape: open offices, agile, remote, hybrid, and now AI and return-to-office mandates. Each change is sold as the thing that will unlock performance. The evidence suggests a duller truth. People do their best work when they understand why it matters, have room to decide how, are getting better at it, make visible progress, and belong to a group where they can speak honestly.

None of that is new, and none of it is expensive. It is just hard to do consistently while the organisation shifts around you. That is the job. If you manage people, pick one item from the playbook this week and do it every week until the end of the year.

Sources