Self-Awareness: The Predictor of Whether Coaching Works

There’s a line people say under their breath in calibration meetings, almost always near the end, when nobody has the energy to argue anymore: “that person isn’t coachable.”

It’s an elegant verdict. It closes the conversation, absolves everyone in the room, and sounds like a diagnosis. It almost never is. In the vast majority of cases we see, what’s being described isn’t an attitude problem. It’s a map problem.

That person does want to improve. What’s happening is that they’re working with a map of themselves that doesn’t match the territory, and every piece of feedback you give them lands in a place that, from where they’re standing, doesn’t exist.

And unlike attitude, that can be measured in advance.

The problem: you invest six months without knowing if you’ll move anything

Think about how the decision to invest in someone’s development gets made today.

Someone is underperforming, or performing well but choking on their relationships with the team. The protocol kicks in: conversation with the manager, development plan, maybe an external coach, maybe a twelve-week leadership program. Real cost: between USD 3,000 and 15,000 per person, plus the manager’s time, plus six months on the calendar.

And in the meeting where that spend gets approved, nobody asks the one question that matters before signing it:

“Is this person in a position to receive this feedback, or will they reinterpret it until it confirms what they already believe about themselves?”

We don’t ask because we assume feedback is a button: you press it and something changes. The research says something considerably more uncomfortable.

The meta-analysis by Smither, London and Reilly (2005), which reviewed 24 longitudinal studies on multisource feedback, found that the average improvement after receiving 360-degree feedback is small. Mean effect sizes were 0.24 for direct-report ratings, 0.12 for peer ratings and 0.14 for supervisor ratings. For self-ratings: 0.00.

Translated: on average, feedback barely moves the needle. But “on average” is the treacherous phrase here. The central conclusion of that work isn’t that feedback doesn’t work. It’s that it works enormously for some people and not at all for others, and that the difference between the two is predictable.

The right question, then, isn’t “do we give them coaching?” It’s “which group is this person in?”

The science: self-other agreement

The concept is called self-other agreement, and it sits on one of the most solid and least applied research lines in organizational psychology.

The idea is simple. You take a person’s self-assessment across a set of behaviors and compare it with how their direct reports, their peers and their boss assess them. The gap between the two isn’t noise. It’s information.

Leanne Atwater and Francis Yammarino formalized the model in the early nineties, and their classification is still the standard. Four groups:

1. Over-estimators. They rate themselves higher than others rate them. They think they listen; their team says they interrupt. They think they delegate; their team says they micromanage.

2. Under-estimators. They rate themselves lower than others rate them. They’re better than they think, and usually don’t know it.

3. In agreement, high performers. They see themselves well, and so does everyone else. The map matches the territory.

4. In agreement, low performers. They see themselves poorly, and so does everyone else. It stings, but they know where they stand.

So far it looks like a harmless taxonomy. What’s interesting is what comes next.

Over-estimators are consistently the worst rated

This is the finding that keeps repeating. In Atwater and Yammarino’s work, superiors rated in-agreement good performers as more effective and also rated under-estimators above over-estimators.

Read that again. Someone who underrates themselves — who believes they’re worse than they are — turns out to be a better leader in their boss’s eyes than someone who overrates themselves. Modesty isn’t a moral virtue here: it’s a symptom that the person is calibrated, or even over-calibrated, and therefore still listening.

The field study by Atwater, Roush and Fischthal (1995), published in Personnel Psychology, confirms it from another angle: managers who overrate themselves relative to how others see them tend to perform worse. And it adds a key data point about feedback: when those over-estimators received upward feedback from their teams, their self-ratings came down.

In other words: feedback can recalibrate the map. But first you have to know the map was broken.

And almost nobody knows theirs is

Tasha Eurich, in a research program spanning nearly five years with around 5,000 participants, published in Harvard Business Review (2018), found the number that sums up the whole problem: 95% of people believe they know themselves well. Between 10% and 15% actually do.

That gap — between confidence in self-knowledge and actual self-knowledge — is exactly the ground where coaching goes to die.

Because the person who overrates themselves isn’t lying. They aren’t being arrogant or difficult or “not very coachable.” They’re reporting, with complete honesty, what they see on their map. When you tell them “your team feels you don’t listen to them,” they don’t hear an observation. They hear an injustice. And their reaction isn’t deliberate resistance. It’s the logical response of someone being described a place that doesn’t appear in their coordinates.

One more finding from Eurich’s research, worth including for how counterintuitive it is: people who spend enormous amounts of time on introspection are not more self-aware. They tend to be less so. Staring into the mirror longer doesn’t fix a crooked mirror. Only outside data fixes that.

The practical consequence: the order of interventions

This is where it stops being a curious fact and becomes a budget decision.

If self-awareness is the entry condition for feedback to operate at all, then giving skills coaching to someone with low agreement is doing things in the wrong order. You’re teaching them to read the map before you fix it.

Look at how the three most expensive decisions in your function change when you have this data up front.

Development. Instead of assigning this person a six-month leadership program, you assign them four weeks of calibration work first: structured feedback from multiple sources, concrete evidence, conversations about specific incidents. Only then does skills development start. Same budget, different order, different results.

Promotion. Multi-evaluator assessment stops being a year-end formality and becomes a real filter. A manager candidate with high potential but low agreement is a silent risk: they’ll reach the role with no read on their own impact, and the first hard piece of feedback will bounce off the same wall. Potential Mapping tells you whether they have the profile for the next level. Self-awareness tells you whether they’ll be able to learn in it.

Hard conversations. You know what you’re walking into before you enter the room. If the gap is in conscientiousness and not in extraversion, the conversation is about concrete deliverables, not about style. The discussion stops being a fight over perceptions and becomes a joint reading of data.

It’s the difference between investing with evidence and investing with hope.

How we measure it at Talen.to

The instrument Talen.to uses to put a number on this is called the Convergence Score: it compares a person’s self-assessment with the assessments from their manager and peers across the same six dimensions of OCEAN+, and returns the distance between the two readings.

I won’t explain it here because it’s already explained in detail — how it’s calculated, what the ranges are, how to read each level — inside the Talento Index guide. What matters for this post is what that number enables: it lets you know, before you sign off on the coaching budget, whether you’re facing a skill problem or a map problem.

Two more things, because these are the objections that always come up.

It isn’t a punishment score. Low convergence doesn’t mean “difficult person.” It means “person who doesn’t yet have the information about themselves.” It’s a development starting point, not a label. The person sees their own result; that’s the design, and it’s the same reason we keep insisting on algorithmic transparency: if you can’t show a score to the person it describes, it isn’t a development tool.

It isn’t measured once. Self-awareness moves. That’s the entire point of running the talent lifecycle on a single engine: today’s convergence is a snapshot, and the one six months from now tells you whether the calibration work paid off. It is, in fact, one of the very few development metrics you can evaluate honestly.

What you can do this week

Without buying anything:

One. Take the last person someone described as “not coachable.” Write down, on a sheet of paper, the two or three specific behaviors behind that sentence. Now ask yourself: has that person ever received concrete evidence of those behaviors, with examples, dates and consequences? Or did they receive adjectives? If it was adjectives, you don’t have a low-coachability case. You have a badly delivered feedback case.

Two. Before your next development plan, ask the person to rate themselves on the same dimensions you’re about to rate them on. Five minutes, same scale. Compare the two columns before the meeting. The rows with more than 20 points of difference are the real agenda for that conversation, and they’re probably not the ones you had written down.

Three. Do the exercise on yourself. Ask two people on your team to rate you on the same dimensions you rated yourself on. It’s uncomfortable, which is why almost nobody does it. If 95% believe they know themselves and only 10% to 15% actually do, the statistics aren’t on your side.

---

Coaching doesn’t fail for lack of willingness. It fails when you talk to someone about a territory that isn’t on their map.

Measuring self-awareness before you invest isn’t a methodological luxury. It’s the difference between spending six months fixing the map and spending them arguing with it.

Meet Talento Index, the assessment of your current team with the same engine you use to hire. The Convergence Score is included in every plan.

---

Sources cited:

• Atwater, L. E. & Yammarino, F. J. (1992). Does self-other agreement on leadership perceptions moderate the validity of leadership and performance predictions? Personnel Psychology, 45(1).
• Atwater, L. E., Roush, P. & Fischthal, A. (1995). The influence of upward feedback on self- and follower ratings of leadership. Personnel Psychology, 48(1), 35-59.
• Smither, J. W., London, M. & Reilly, R. R. (2005). Does performance improve following multisource feedback? A theoretical model, meta-analysis, and review of empirical findings. Personnel Psychology, 58(1), 33-66.
• Eurich, T. (2018). What Self-Awareness Really Is (and How to Cultivate It). Harvard Business Review.