Arrangements that involve another person consistently improve follow-through on intentions. Several different mechanisms are at work, and they are not equally present in every arrangement.
Observation changes behaviour by itself
Knowing that an action will be reported alters how it is approached, even where the observer has no authority and imposes no consequence.
The effect is strongest when reporting is specific and scheduled. A vague understanding that someone is interested produces much less than a fixed weekly account of what was done.
This is why written logs shared with another person tend to outperform private ones, despite containing exactly the same information.
Commitment made aloud is harder to revise quietly
An intention held privately can be adjusted at any point without anyone noticing, and it usually is, often without the adjustment being registered as a change.
Stating it to someone creates a record. Abandoning it then requires an explanation, and the anticipation of that conversation carries weight at the moment of decision.
The mechanism depends on the commitment being concrete. A stated aim to get fitter creates no such friction, since nothing specific has been promised.
Scheduled contact supplies the cue
Much of what accountability partners provide is timing. A standing appointment means the behaviour has a fixed slot rather than competing with everything else for attention.
Group formats work similarly, and part of their effect comes simply from the meeting existing on a calendar that others also hold.
When the schedule lapses, follow-through usually declines quickly, which suggests the cue is doing more work than the encouragement.
Encouragement and accountability are different functions
A partner who reassures without asking questions provides support but removes the observation that produced the effect.
Arrangements that work generally include a straightforward account of what was and was not done, delivered without either criticism or excuse-making.
People often choose partners for warmth and then find the arrangement ineffective, because the two roles pull in opposite directions and are hard to combine in one person.
The structure can be built without another person
Public logs, automated tracking and pre-committed arrangements reproduce parts of the effect by making behaviour visible or by fixing consequences in advance.
These substitutes capture the observation and commitment elements while missing the relational one, and they suit people who find social arrangements uncomfortable.
What none of them replaces is a partner's judgement about when the goal itself is wrong, which is the part of the arrangement that most often matters over longer periods.