IX Coach

Accountability Coaching, and the Statistic You Should Not Trust

Does accountability coaching work, and what does the research actually show?

Accountability coaching sells one mechanism: someone is expecting a report from you. The figure everyone quotes for it, a 65% or 95% success rate attributed to the ASTD, has no study behind it that could be found. There is real research, and it is smaller and more specific: in Gail Matthews' study, the group that wrote goals, made action commitments and sent weekly progress reports to a friend achieved significantly more than every other group, with achievement rated by the participants themselves.

Accountability coaching is a standing arrangement in which you report to someone: what you said you would do, and whether you did. What is for sale is the audience. Insight and method are mostly absent by design, because the people who buy it already know what to do and keep not doing it, which is a real and common problem and a fair thing to pay for.

That makes the useful question who the audience should be: a friend, a group, a paid coach, or a conversation with software. IX Coach, an AI coaching product, publishes this page and is one of the options, so it is placed against that question near the end rather than argued for. First, though, the evidence, because this corner of coaching has an unusually poor relationship with it.

A number with no study behind it

If you have read that committing to someone makes you 65% more likely to reach a goal, or that a specific accountability appointment lifts that to 95%, you have met a figure credited to the American Society for Training and Development. A search for the study turns up the figure repeated on many business and self-help sites, and none of them links to a paper, a methodology or a dataset. That does not prove it was invented. It does mean a number nobody can trace is not evidence, and a coach who offers it as proof has not checked their own central claim.

A related piece of folklore has a clearer history. Matthews' own research summary opens by calling the supposed Yale or Harvard study, in which 3% of graduates with written goals went on to out-earn the rest, an urban myth, reporting that reviews by her and by the social psychologist Steven Kraus, and investigative reporting by Fast Company, found that no such study had ever been done.

What Matthews' five groups actually show

Her study is the one people are usually reaching for. A total of 267 participants were recruited from businesses, organisations and networking groups, and 149 completed it. They were randomly assigned to one of five conditions: an unwritten goal; a written goal; a written goal with action commitments; those plus commitments sent to a supportive friend; and those plus weekly progress reports to that friend. Group 5 members were also sent weekly reminders to email the report. The goals were ordinary four-week ones, such as completing a project, increasing income or getting organised.

Matthews' own summary states the result this way: "Group 5 achieved significantly more than all the other groups; Group 4 achieved significantly more than Groups 3 and 1; Group 2 achieved significantly more than Group 1." Writing the goal beat only thinking about it; sending your commitments to a friend beat writing them privately and beat leaving the goal unwritten; and sending that friend weekly reports beat every other arrangement.

Three limits belong next to those results. The outcome was not independently measured: at the end of four weeks participants were asked to rate their own progress and the degree to which they had accomplished their goals. Of the 267 recruited, 118 did not finish. And the design cannot separate the friend from the rhythm, because Group 5 had both, along with the researchers' reminders. The percentage figures that circulate alongside this study do not appear in the summary and are not used here.

Explore: accountability partners · commitment contracts

What binding yourself looks like when it is tested

Commitment contracts, where you stake something you will lose if you fail, are the other mechanism that accountability products borrow. The closest tested version is a randomised trial by Giné, Karlan and Zinman of a contract for quitting smoking, in which participants deposited money that was returned if they passed a nicotine test and forfeited to charity if they did not. Of the smokers offered the contract, 11 percent took it up, and those randomly offered it were 3 percentage points more likely than the control group to pass the six-month test. The effect persisted in surprise tests at twelve months.

Real, small, and chosen by few is roughly what the evidence for binding yourself looks like. It is a result about smoking, so it should not be stretched to a work goal, but it is a useful corrective to the idea that a strong enough commitment device makes follow-through automatic.

Explore: commitment contracts · precommitment devices

Choosing who the audience is

If the weekly report to a person is the studied mechanism, then a friend and a shared spreadsheet reproduce most of it for nothing, and for some readers that is the right answer. A group can do the same with more people watching. A paid coach adds two things a friend often cannot: a judgement about whether the plan is worth following, and social weight from someone you have agreed to answer to.

The better practitioners work on why the commitment failed, not only that it did. After two or three cycles patterns show: promises made in an optimistic mood on Sunday, work scheduled into hours that were never free, tasks that are really projects. Without that, accountability coaching is an expensive streak counter. And at some point a good coach asks whether you should still be doing the thing at all, since an accountability relationship will make you follow through on a poor plan as efficiently as a good one.

There is also a failure mode particular to this format. Some people answer outside expectation by performing it, producing reports instead of progress. If you recognise that from school, say so in the first session; a good coach will change the design.

Explore: crucial accountability · accountability partners · smart goals · process vs outcome goals

When being watched is the wrong tool

If starting has been the problem in every job you have held, not just this one, take that history to a clinician for an assessment before paying anyone to watch you not start. If the block is that the work frightens you rather than bores you, closer watching tends to make it heavier; that is coaching material, but of a different kind. And if the hours genuinely are not there, no reporting rhythm creates them.

Groups deserve a word here as well. They can supply an audience cheaply and a sense that others face the same problem, but the more people there are, the less likely any one of them is to notice a single missed report. If you pick a group, find out who actually reads the reports and what happens when yours is blank.

What a report worth sending contains

The studied report was brief, so a useful one can be too. It names the action commitment in the words it was made, says plainly whether it happened, and gives one sentence on what got in the way if it did not. It does not narrate the week or apologise. A coach or friend who receives that every Friday has what they need to ask the one question that matters, which is what you will change.

Who it suits is fairly clear from this. The arrangement works best for a person who has a decided goal, honest reasons for missing it, and no difficulty telling someone the truth. It suits people poorly when the goal is not yet chosen, or when reporting itself feels like the threat.

Report on actions, not outcomes

A report is easier to keep honest when it covers what you did rather than what resulted. Sent three runs this week is something you either did or did not; got fitter is not. This is the logic of process goals as against outcome goals, and it matters here because an outcome you only partly control makes a weekly report a weekly excuse.

It also protects the relationship. A friend or coach who is told about outcomes ends up commenting on luck, whereas one who is told about actions can ask about the plan.

Explore: process vs outcome goals

IX Coach measured against the five groups

Set against Matthews' conditions, IX Coach sits nearest Group 3 with a conversation attached. It is an AI coaching product, not a person, so it supplies none of what Groups 4 and 5 had: nobody receives the report, and nobody notices if it stops coming. What it does offer is that it answers without a booked slot, including at the hour you skip a commitment, and that it uses relevant detail from earlier conversations, so the fourth missed Tuesday can be looked at with the first three in view. The ideas it works from, accountability partners, commitment contracts and process versus outcome goals, are published pages you can read before deciding anything.

The sensible arrangement for most people is therefore a compound one. If the expectation of a specific person is what moves you, arrange that with a friend, a group or a coach, and use whatever is available in the hours between reports for thinking about why you skipped.

Judging an accountability coach before the first session

Ask what exactly gets reported, in what form and how often. Ask what happens when the report is that you did not do it; a coach whose answer is encouragement is teaching you to report comfortably rather than accurately. Ask what the smallest unit of progress worth reporting is, since a weekly report on a goal with no weekly evidence is theatre.

On credentials, the ICF's ACC records training, logged coaching hours and mentor coaching; nothing in it records whether a coach's clients follow through, so ask how they know theirs do. Settle the fee and what happens to a missed session before you start.

Frequently asked questions

Does accountability coaching work, and what does the research actually show?

Accountability coaching sells one mechanism: someone is expecting a report from you. The figure everyone quotes for it, a 65% or 95% success rate attributed to the ASTD, has no study behind it that could be found. There is real research, and it is smaller and more specific: in Gail Matthews' study, the group that wrote goals, made action commitments and sent weekly progress reports to a friend achieved significantly more than every other group, with achievement rated by the participants themselves.

Does accountability coaching actually work?

The mechanism has some support and the marketing figure has none that could be found. In Matthews' study, people who sent weekly progress reports to a friend rated their goal achievement higher than every other group. The 65% and 95% figures credited to the ASTD have no traceable study behind them.

What should an accountability coach actually ask me?

What gets reported, in what form and how often, and what happens when the report is that you did not do it. After a few cycles a good one asks why commitments failed and whether the goal is still worth pursuing.

Can a friend do this instead of a paid coach?

For the mechanism Matthews studied, yes: weekly reports to a supportive friend is what her strongest group did. A coach adds judgement about whether the plan is worth following and a stronger sense of obligation than a friend may want to apply.

Where do the 65% and 95% accountability figures come from?

They are credited to the American Society for Training and Development, but no paper, method or dataset could be found. They are repeated across many websites, which is not the same as being studied.

Is a commitment contract better than a coach?

The one randomised trial cited here, on quitting smoking, found a small effect and few people took up the contract. Whether a contract or a coach suits you depends on whether a stake or a person is what you respond to.

Research

Practice this with IX Coach

Try this practice

Keep reading