Why Working With a Coach Works (When Going It Alone Doesn't)

You've probably heard some version of it: a study found that the 3% of Harvard (or Yale, depending on who's telling it) graduates who wrote down their goals ended up earning ten times more than the other 97% combined. It's one of the most repeated "facts" in the entire self-improvement world. It's also, as far as anyone has ever been able to verify, completely made up. There's no such study, no such class, and no researcher who can produce it. So what does the real research say about goals, accountability, and whether working with another person — a coach — actually changes your odds? As it turns out, quite a lot, and it's more interesting than the myth.

The study you've heard was fake. Here's the one that's real.

Dr. Gail Matthews, a psychology professor at Dominican University of California, has spent part of her career tracking down and debunking the Harvard/Yale goals myth. But she also ran her own actual study on the subject in 2015, and it's the one worth paying attention to. She recruited 149 participants — entrepreneurs, professionals, and others working toward personal or professional goals — and split them into five groups with progressively more structure:

• Group 1 — unwritten goals. Just think about your goal. No writing, no plan.
• Group 2 — written goals. Write the goal down, nothing else.
• Group 3 — written goals plus action commitments. Write the goal and a list of actions you'll take.
• Group 4 — commitment sent to a friend. Same as Group 3, plus you send your goal and action commitments to a supportive friend.
• Group 5 — weekly progress reports. Same as Group 4, plus a weekly written progress update sent to that same friend.

On a 10-point goal-achievement scale, Group 1 (unwritten goals) averaged 4.28. Simply writing the goal down (Group 2) jumped that to 6.08 — a real effect on its own. But the highest score by a clear margin was Group 5, at 7.6, the group that combined a written goal, action steps, and a standing weekly check-in with another person. Matthews reported that this group achieved significantly more than every other condition, including the ones with written goals alone.

What the research says about coaching specifically

Accountability from any person is one piece of it. But there's also a growing, more specific body of research on health and wellness coaching as its own intervention, separate from general goal-setting research. A 2023 systematic review and meta-analysis pooled 30 randomized controlled trials and more than 8,600 participants, published between 2005 and 2023, looking at coaching's effect on outcomes that actually matter to the people receiving it — not just whether they hit a number on a chart.

The review found statistically significant improvement in three areas: quality of life (measured across 14 of the included studies), self-efficacy, meaning someone's belief in their own ability to make a change (13 studies), and depression symptoms (15 studies). We want to be equally clear about the fourth: anxiety, measured in 7 of the studies, was the one outcome that did not reliably improve with coaching. That's an important, honest limit, and it's exactly why coaching is positioned as support for building a life back up — not as a replacement for licensed mental health treatment when that's what's actually needed.

Coach Nadira

The Resilience Rebuild — 90-Day 1:1 Coaching

Weekly 45-minute sessions, support between calls, and a plan built around your life — exactly the kind of structured, accountable relationship the research above points to.

See How It Works

Why a person still beats an app

Apps, trackers, and AI-driven coaching tools have gotten genuinely useful, and we've written honestly about the good evidence behind simple tracking elsewhere on this blog. A 2025 systematic review comparing human, AI, and hybrid coaching in digital health found that both human coaches and AI-driven tools showed a positive impact on engagement and lifestyle outcomes. But it also flagged something worth knowing before you assume an app will do the same job as a person: only about 1 in 5 of the digital-coaching studies reviewed were full randomized controlled trials — most were small pilot or feasibility studies, a much thinner evidence base than the research behind human coaching. The reviewers also noted that engagement metrics varied so much between studies that direct comparisons were hard to make. A human coach can also do something a script can't yet: notice that this week is different from last week, and adjust the plan in real time instead of running the same sequence regardless of what's actually happening in your life.

Three signs coaching is worth trying right now

• 1. You already know what to do, but you're not doing it. If the gap is information, a book or a guide might close it. If the gap is follow-through, the research above says a person you report to closes it better than information does.
• 2. You've tried the generic plan and it didn't stick. A plan that isn't built around your actual schedule, energy, and setbacks is easy to abandon the first hard week. A coach adjusts the plan to you instead of asking you to fit the plan.
• 3. You're rebuilding after something, not starting from a blank slate. Grief, burnout, and a health scare all change what "doable" looks like week to week. Coaching that flexes with that is different from a fixed 30-day program that assumes a steady baseline.

And one honest sign it's not the right tool yet: if you're not ready to make any changes at all, no amount of accountability fixes that — a gentler starting point, like a free mini-guide, makes more sense until you are. And if what you need is clinical or emergency mental health support, that's a licensed therapist's job, not a coach's, and a good coach will say so and help you find the right referral rather than pretend otherwise.