Skip to content

Accountability vs responsibility

Responsibility is being involved in doing the work. Accountability is being individually and ultimately answerable for whether it happened. Several people can be responsible for one outcome; only one can usefully be accountable, and conflating the two is how outcomes end up owned by everyone and nobody.

Accountable parties are held individually and ultimately responsible for the success or failure of a task, which is why the number of them should be one.

Why it matters when the plan changes

Organisations distribute responsibility easily and accountability reluctantly, because naming one person answerable for an outcome is an uncomfortable conversation. The result is outcomes with several contributors and no owner, which look well staffed and stall at the first trade-off nobody can settle. This shows up hardest at the dependencies a plan depends on, the handoffs and cross-team commitments that sit beneath any strategy, where research on strategy execution has found that only 9% of managers say they can rely on colleagues in other functions and units all the time, a gap that unowned accountability widens rather than closes.

The tension is that single accountability feels like it removes collaboration. It does the opposite: it lets several people contribute without the work stalling when they disagree, because someone can decide. Shared accountability preserves the appearance of partnership while removing the one mechanism that makes partnership work, someone able to settle the question when contributors cannot agree.

In practice

An integration plan lists data migration as a shared responsibility between two engineering leads. Both are working; neither holds the sequencing call. The dependency slips twice before anyone says out loud that shared responsibility was never assigned to a person, and the fix takes an afternoon once someone does.

Evidence

What it cannot tell you

Accountability versus responsibility tells you who should decide, not whether that person has the standing, information or authority to decide well. Naming one accountable owner does not by itself fix broken dependencies or missing information; it only creates the single point where a disagreement can be settled, or where a bad decision now has a clear author.

Questions

Contribution can be shared; ultimate answerability cannot, in any way that helps. Two people jointly accountable means that when they disagree there is no mechanism left to settle it, which is precisely the situation that single accountability exists to resolve in the first place.

It enables it. Several people can contribute fully when someone can decide if they disagree. Research reported in Harvard Business Review in 2015 found only 9% of managers could rely on colleagues in other functions all the time, the exact gap that a single accountable owner is meant to close.

Because it is an uncomfortable conversation that assigns risk to a named person, and because shared ownership sounds collaborative. Both reasons are real. Neither reason survives the first slipped dependency, at which point the uncomfortable conversation happens anyway, under noticeably worse conditions than it would have.

Take the outcomes the plan depends on and ask separately of each who is answerable by name and by date. The gaps show as either two names or none, and the exercise takes an hour. It is usually done after the first milestone has already slipped.

A RACI, short for the Responsibility assignment matrix documented on Wikipedia in 2026, is one convention for writing this down, separating responsible, accountable, consulted and informed. The convention matters less than the discipline behind it: exactly one accountable name per outcome, not two.