A group of workers take hold of a rope. One by one, they have already shown how hard they can pull alone. Now they pull together. The gauge rises, but not as much as the arithmetic says it should.
At first, nothing looks wrong. The group is stronger. Work is getting done. The loss appears only when Max Ringelmann looks at the contribution per person: the average has fallen.
Modern teams rarely pull the same rope, which is probably for the best. Team dependability fails more quietly: the handoff everyone thought was covered, the risk mentioned only after the deadline, the dependable person quietly doing a second job called "making sure."
A century later, Google ranked dependability second among the 5 dynamics in Project Aristotle, behind psychological safety and ahead of structure and clarity. Psychological safety got the keynote. Dependability got the follow-up email nobody wanted to send.
When follow-through breaks, managers tend to reach for character judgments. Someone is lazy. Someone does not care. Someone needs more accountability. Sometimes they are right. Often they are looking at poor coordination, fuzzy ownership, more work than the team can carry, invisible effort, or a team that has made bad news expensive to share.
So the useful question is not whether dependable people are good and unreliable people are bad. It is what makes a team more likely to keep its promises, how to tell a motivation problem from a badly designed commitment, and why the person who always steps in can keep the real problem hidden.
What is team dependability?
Team dependability is the shared confidence that commitments will be completed to the agreed standard and on time. When that becomes unlikely, the responsible person says so early, explains what changed, and helps reset the plan. Dependability concerns the reliability of the team system, not just the virtue of individual people.
Google's re:Work account of Project Aristotle used a shorter definition: dependable teammates "reliably complete quality work on time." Its survey item was even plainer: when teammates say they will do something, they follow through.
The scale behind that finding was unusual. Google examined 180 teams, ranging from 3 to 50 people with a median of 9, reviewed more than 250 survey items, and ran more than 35 statistical models. Dependability finished second among the 5 team dynamics.
That does not make it the second law of team physics. As we noted in our account of what Project Aristotle found, Google published an internal research program, not a peer-reviewed study. Its definition is useful because it gives managers something observable to discuss. Its ranking should be held more lightly.
There is another reason to be careful with the word. Dependability sounds like a character trait, which makes every miss feel like evidence about the person. Teams are systems of promises. A sensible diagnosis starts with how the promise was made, whether the work was possible, and what happened between commitment and delivery.
Otherwise, "accountability problem" becomes management's version of "miscellaneous." It explains very little and makes the person carry all of it.
Why do people sometimes work less in groups?
People sometimes work less in groups because their individual effort is pooled, harder to evaluate, or less obviously necessary to the result. Social psychologists call that motivation loss social loafing. It is only one kind of group-performance loss, and that distinction matters.
Ringelmann's rope matters because it contains 2 different problems that modern managers routinely bundle together. Groups produced more total force, but less force per person as they grew. That gap became known as the Ringelmann effect.
Ringelmann was a French agricultural engineer who was more interested in human and mechanical work than office culture. His 1913 report examined how workers pushed and pulled loads alone and together. The paper later became famous through secondary citations, while the original was difficult to locate. David Kravitz and Barbara Martin finally reconstructed the history in "Ringelmann Rediscovered" in 1986.
The gap can come from 2 places. Coordination loss happens when people are trying but their efforts interfere, arrive in the wrong order, or require more synchronization. Motivation loss happens when a person contributes less because the output is shared with others.
Alan Ingham and his colleagues separated the 2 in a 1974 rope-pulling study. One experiment used actual groups of 1 to 6. A second used pseudo-groups where each person pulled alone but believed up to 5 others were pulling too.
Individual performance fell when the first 2 perceived coworkers were added, even without a coordination problem. It then leveled off rather than declining in a neat straight line.
In 1979, Bibb Latané, Kipling Williams, and Stephen Harkins moved from ropes to noise. Across 2 experiments with 84 undergraduates, participants clapped and shouted alone, in real groups, and while believing their individual sound was pooled with a group. They exerted less effort when they thought they were part of a combined output. The researchers gave that motivation loss its name: social loafing.
This is good science and a terrible excuse to call a coworker lazy. Clapping in a laboratory is not the same job as coordinating a product launch, running a hospital unit, or teaching a class. The studies established that pooled effort can change motivation. They did not give managers a personality test disguised as a deadline.
Is social loafing at work just laziness?
No. Social loafing at work can reflect several features of the task and group, including whether individual effort can be evaluated, whether a contribution feels necessary, what effort coworkers are expected to make, and whether the result matters to the person doing the work.
The strongest broad evidence is Steven Karau and Kipling Williams's 1993 meta-analysis of 78 studies. It found social loafing across many tasks and populations. It also found substantial variation. Evaluation potential, expectations of coworker performance, task meaningfulness, and culture were especially influential moderators.
Managers like a single-cause story because it fits neatly on a slide. Teams decline the invitation.
Visibility changes effort, but visibility is not enough
In 1981, Williams, Harkins, and Latané ran 2 cheering experiments on identifiability. Making each person's output identifiable eliminated social loafing in the group condition. Making output unidentifiable could produce loafing even when a participant performed alone.
That result is often translated into "make all work visible." The next study made the advice less tidy. In a 1985 brainstorming experiment with 4-person groups, Harkins and Jackson found that identifiability raised output only when individual contributions could also be evaluated through comparison. A visible contribution that nobody can interpret does not create useful accountability. It creates a dashboard.
Complex work makes comparison especially difficult. One engineer may close fewer tickets because the work is harder. One manager may resolve a problem before it becomes a project. Counting visible units without understanding them rewards legibility, not necessarily value.
The practical point is narrower: make ownership and progress visible enough for the team to coordinate. Do not confuse surveillance with clarity.
People coast when their effort feels dispensable
Norbert Kerr and Steven Bruun called this the free-rider effect. Across 3 experiments published in 1983, they tested whether effort fell when people believed their contribution was unnecessary to group success. It did, under the task conditions where their effort appeared dispensable.
This explains why "everyone owns this" so often means nobody knows what their part changes. Shared purpose is useful. Shared ownership can become a fog bank.
A dependable team can answer 2 questions without scheduling a workshop: Who is directly responsible for this result? What becomes harder or impossible if their contribution does not arrive?
People notice what coworkers put in
Social loafing is partly social. In a 1985 experiment, Jeffrey Jackson and Stephen Harkins changed what participants believed a coworker intended to contribute. Participants adjusted their own effort toward that expectation, supporting an equity explanation sometimes called the sucker effect.
Nobody wants to be the only person taking a collective promise literally. Once low effort becomes expected, reducing your own contribution can feel less like shirking and more like restoring fairness.
That response is understandable. It is also how dependability erodes without anyone announcing that it has.
Meaning changes the calculation
The meta-analysis found task meaning mattered, and later experiments help explain why. When people believe the result is valuable and that their effort can affect it, collective work has a reason to earn their attention. Remove either link, and "the team needs this" becomes background noise.
That is why a list of assigned tasks is not yet a dependable system. People need to see the result, their part in it, and a credible path from one to the other.
Why does the dependable person end up carrying the team?
Sometimes the most dependable person lowers their effort to avoid being exploited. Sometimes they do the opposite and work harder to protect the group result. The research supports both responses.
Williams and Karau tested the second pattern in 3 experiments on social compensation. Participants generated ideas either collectively or side by side. When they expected a coworker to perform poorly on a meaningful task, they worked harder collectively than individually. When the task carried little meaning, they did not compensate.
That finding complicates the familiar story about free riders. The reliable person may not stop carrying the slack. They may become better at carrying it.
From above, the team still appears dependable. The launch happens. The client gets the answer. The report arrives 4 minutes before the meeting, apparently by magic.
The same 2 people keep performing the magic, which is how an allocation problem acquires excellent quarterly metrics.
Social compensation is not proof of burnout, and the experiments did not follow office teams for years. The managerial inference is still worth testing: successful delivery can conceal an unhealthy distribution of rescue work. If the same people repeatedly save the promise, the system is borrowing dependability from them.
This is also why individual praise can miss the team problem. Thank the person who rescued the work. Then ask why a rescue was required, who was supposed to see the risk, and whether the plan is likely to produce the same hero next week.
A team that depends on heroics is not dependable. It is lucky in a repeatable-looking way.
How do psychological safety and dependability work together?
Psychological safety makes it possible to say a promise is in danger. Dependability determines what the team does with that information.
That connection is easy to miss because safety and accountability are often presented as opposing management styles. Our piece on why psychological safety is not about being nice makes the opposite case. Safety serves the work by letting important information travel before it becomes expensive.
Suppose a Thursday deadline is slipping on Tuesday. On a safe team, the owner can say so without preparing a legal defense. On a dependable team, that admission starts a concrete conversation: reduce scope, move the date, add help, or accept the risk.
The original promise does not vanish. It becomes a better-informed promise.
Now remove safety. The owner waits until Thursday because every earlier warning has been treated as a lack of commitment. The manager receives the news at the deadline, calls it an accountability failure, and tightens the reporting process. Everyone learns to make the next warning look greener for longer.
Early bad news is not a breach of dependability. Silence is.
That does not mean repeated warnings excuse repeated misses. A person who regularly cannot meet a clear, realistic commitment may need a smaller workload, more skill, better support, or direct performance management. Psychological safety should make that diagnosis more accurate, not make the standard disappear.
The reliable team standard is simple to state and harder to practice: tell the truth about the commitment early, then do what the new plan requires.
How can managers build a more dependable team?
Start by auditing the promise before judging the person. Most commitment failures become easier to diagnose when the original agreement is written in plain language.
Ask 7 questions:
- Who owns the result? One accountable owner can coordinate many contributors. "We" is not an owner.
- What exactly counts as done? A date without a quality bar is an argument scheduled for later.
- When is it due, and why then? Deadlines with no consequence tend to become preferences. Fake urgency teaches people that dates are negotiable.
- Does the owner have the capacity and skill? A promise made against impossible workload is not motivation. It is arithmetic with optimistic branding.
- Which dependencies can stop the work? Name the approvals, inputs, and decisions that sit outside the owner's control.
- When should risk be raised? Do not wait for the due date to inspect the due date. Pick an earlier point when changing the plan is still cheap.
- What happens when someone raises a risk? If the answer is public blame, expect private delay.
The pattern behind a miss usually points to a different response:
| What you see | Likely problem to test | Manager response |
|---|---|---|
| "We all own it" | Ownership ambiguity | Name 1 owner |
| Work is late at a handoff | Coordination loss | Simplify the handoff |
| Progress is visible, quality is disputed | Undefined standard | Define "done" together |
| The same people rescue every deadline | Hidden social compensation | Rebalance work and authority |
| Risk appears only on the due date | Unsafe or absent escalation | Add an early risk check |
| Clear, realistic promises still slip | Skill, capacity, or motivation | Manage the actual constraint |
Visibility matters, but the goal is coordination. A lightweight commitment log, project plan, or weekly review can make ownership and risk legible. It should answer what changed and where help is needed. It does not need to document every 11 minutes of a person's day.
Close loops visibly. When a commitment changes, record the new owner, scope, and date. When work finishes, say so. Many teams are reasonably good at starting work and surprisingly mystical about whether it ended.
Finally, notice early warnings and rescues. If managers celebrate only the dramatic save, they teach the team to produce work that requires saving. Quiet, on-time delivery is less cinematic. That is part of its appeal.
What can an org chart actually help with?
An org chart can clarify formal reporting lines, team boundaries, and where an escalation should go. Those are real inputs to dependability. People cannot coordinate cleanly when they do not know who owns a decision or who has the authority to unblock it.
That is why the first step in creating an organizational chart is getting the relationships right, not decorating the boxes. A clear chart removes one category of avoidable ambiguity.
It still cannot show whether people keep their promises. It cannot render an overloaded manager, a hidden dependency, a vague quality standard, or the teammate who quietly saves every deadline. The next layer is making ownership explicit for the work between reporting lines. Team behavior tells you whether it can live there.
The research carries its own limits too. Project Aristotle was internal Google research. The classic social-loafing experiments often used undergraduates performing short, artificial tasks such as rope pulling, clapping, shouting, and brainstorming. A 2021 review of field research on effort gains found that people often work harder in groups when their effort feels indispensable and group success feels possible.
So there is no Ringelmann percentage you can apply to a product team, and no missed deadline that proves someone loafed. The evidence gives managers a better set of questions. It does not grant permission to replace curiosity with a label.
Draw the structure clearly. Make the promises equally clear. Then pay attention to what happens between them.
Frequently asked questions
What is dependability in a team?
Dependability is the shared confidence that team members will complete agreed work to the required standard and on time. It also includes raising risks early, renegotiating responsibly when conditions change, and closing the loop on the revised commitment. A dependable team is predictable, not perfect.
What is social loafing at work?
Social loafing is a reduction in individual effort when work is pooled into a group result compared with working individually or alongside others whose output remains separate. It is a group-motivation effect, not a synonym for laziness, ordinary downtime, or every missed deadline.
What is the difference between the Ringelmann effect and social loafing?
The Ringelmann effect is the loss in average individual output that can occur as a group grows. It may include coordination loss and motivation loss. Social loafing refers specifically to reduced individual motivation or effort when contributions are combined into a group product.
How can a manager reduce social loafing without micromanaging?
Make ownership, expected output, dependencies, and progress visible enough for coordination. Explain why the work matters and how each contribution affects the result. Add an early point for raising risk. Visibility should help the team make decisions, not create a surveillance record of activity.
Can a dependable team still miss a deadline?
Yes. Dependencies change, estimates fail, and new information arrives. A dependable team surfaces the risk early, explains what changed, agrees on a new plan, and follows through. Repeated misses still require management, but surprise is not a necessary part of accountability.
Dependability is not the exciting part of a team. It lets the exciting people trust one another enough to do the work.