There is a particular kind of bewilderment that belongs exclusively to experienced coaches.
You have logged thousands of hours in the session chair. You have read the research, earned the credentials, and built an intuitive feel for when a client is genuinely shifting versus performing the language of change. And then, without warning, a client who just two sessions ago articulated a sophisticated, multi-stakeholder perspective on a thorny organizational conflict walks into Thursday's session and reacts to a minor scheduling disruption as though the very foundations of his career are crumbling.
You did nothing wrong. The client did nothing wrong. And yet something clearly happened.
What happened was development — real, scientifically predictable, messy human development. The problem is not the regression. The problem is that the field of coaching has not yet given practitioners the conceptual tools or the measurement infrastructure to recognize it for what it is, let alone harness it as evidence of coaching's value.
This article makes the case that the variability in a leader's performance — the distance between their ceiling and their floor, the width of their developmental range — is not a coaching failure. It is a coaching signal. And when coaches learn to measure it with the same rigor that industrial engineers apply to process capability, a new and more defensible case for coaching's return on investment becomes possible.
The popular imagination of adult development is architectural: a staircase, each step earned through insight and practice, leading steadily upward. Progress graphs point to the upper right. Pre- and post-assessments show the client at Level 3 before coaching and Level 4 after. Case closed, ROI established.
This model is compelling. It is also wrong.
Kurt Fischer's Dynamic Skill Theory, developed across more than three decades of empirical research, establishes something more unsettling and more useful: human cognitive performance is not a fixed property of a person. It is a dynamic output of a person in a context (Fischer, 1980; Fischer & Bidell, 2006). Remove the environmental support — the trusted colleague in the room, the familiar setting, the absence of time pressure — and what looked like a consolidated capability can contract to an earlier, more automatic mode of functioning within minutes.
Fischer called this the distinction between optimal level and functional level. Optimal level is what a person can do with maximal support: the right cues, low threat, high familiarity, sufficient cognitive resources. Functional level is what actually shows up under normal or adverse conditions. The gap between them is not small. Across hundreds of studies, Fischer and his colleagues found that this gap routinely spans one to two full developmental levels — the equivalent of years of supposed growth (Fischer & Pipp, 1984).
What this means for executive coaching is difficult to overstate. A leader who can articulate elegant second-order thinking in a coaching session — where she is at ease, psychologically safe, and operating with her coach's scaffolding — may well revert to first-order reactive decision-making the moment she steps into a board meeting with an adversarial chair and inadequate sleep. She has not forgotten how to think at the higher level. The skill is real. It is simply not yet robust enough to deploy without support.
This is the invisible architecture of leadership development. And it is almost entirely hidden from coaches who rely on periodic, snapshot-based assessments.
Fischer's framework maps closely onto the Model of Hierarchical Complexity (MHC) developed by Michael Lamport Commons and colleagues (Commons et al., 1998; Commons & Richards, 2002). Where Fischer illuminates the dynamic, context-sensitive nature of skill performance, Commons provides a mathematically grounded taxonomy of the cognitive orders themselves — from concrete single operations, through formal, systematic, and metasystematic reasoning, up to paradigmatic and cross-paradigmatic thinking at the frontier of adult development.
The MHC is not a personality theory or a soft typology. It is a formal measurement framework in which each order of hierarchical complexity is defined by the logical structure of the cognitive operations it coordinates. A systematic thinker is not "more strategic" in a vague sense — she coordinates multiple formal systems simultaneously in a way that cannot be reduced to the operation of those systems individually. A metasystematic thinker goes further, comparing and critiquing entire systems as objects of analysis.
What makes the MHC indispensable for coaching practitioners is precisely what makes it uncomfortable: it reveals that the developmental range most leaders actually operate in — the distance between their best thinking and their worst — is wider than their coaches, their organizations, or they themselves typically recognize.
The leader who presents as a metasystematic thinker in a strategic offsite and then micromanages a direct report's slide formatting the following Monday has not lost his capability. He has moved from his optimal level to his functional level. In Fischer's terms, his skill has "collapsed." In MHC terms, he has temporarily regressed to a lower order of hierarchical complexity because the conditions supporting higher-order operation were withdrawn.
The coaching question is not why did he regress. The coaching question is: what is the shape of his developmental range, and what would it take to narrow it?
The dominant methodology for evaluating coaching effectiveness remains the pre-post design: administer a 360-degree instrument or a validated leadership assessment before the engagement begins, coach for six to twelve months, administer the same instrument again, and report the difference. If the score went up, coaching worked. If it did not, the conversation becomes uncomfortable.
This design is not useless. It provides a rough map of directional movement. But it systematically obscures the most important feature of human development: its within-person variability over time.
Consider what a pre-post design actually captures. It takes two snapshots — one before coaching, one after — and computes the difference. It cannot tell you anything about what happened between those two moments: the peaks, the valleys, the sessions where the client operated at the cutting edge of her capability, and the sessions where she regressed to defensive, reactive, early-stage patterns. It produces a line between two points and asks you to believe that line describes the trajectory.
But development is not a line. It is a trajectory with variance. And that variance is, scientifically and practically, where the action is.
Here is the claim that industrial and systems engineers have understood for decades and that coaching has only begun to articulate: reducing variation is itself a form of improvement, independent of any change in the mean.
This principle sits at the heart of Six Sigma methodology and Taguchi's robust design framework (Pande, Neuman, & Cavanagh, 2000; Taguchi, Chowdhury, & Wu, 2005). A manufacturing process that produces parts averaging exactly the right dimension but with high variance around that mean is not a good process — it is an unreliable process. The individual parts that hit the target are not evidence of quality; they are evidence of luck. Quality is defined by the distribution, not the occasional peak.
Process capability indices such as Cpk and Cpm were designed precisely to capture this insight. They express the relationship between the natural variation in a process and the tolerance limits within which that process must operate. A high Cpk means that variation is tight and the process is reliably hitting its target. A low Cpk means that even if the average is on target, individual outputs are scattering outside acceptable limits.
Now translate this directly to leadership development. A leader whose cognitive complexity varies widely across contexts — who thinks at the formal operational level in low-stakes settings but collapses to concrete operations under pressure — has low process capability. The occasional brilliant strategic insight is real, but it is not a reliable organizational resource. The organization cannot plan around it, delegate based on it, or use it as the foundation for decisions that depend on consistent execution.
A coaching engagement that succeeds in raising the floor — that tightens the developmental range so that the leader's worst-day thinking is closer to his best-day thinking — has delivered substantial organizational value, even if the mean score on a post-assessment looks statistically similar to the pre-assessment score. The leader is more reliable. His decisions are more consistent. His team can anticipate him with greater confidence. He is, in the language of industrial engineering, a higher-capability process.
This is not a metaphor borrowed from manufacturing for rhetorical effect. It is a direct parallel with rigorous empirical grounding. And until now, coaching had no way to measure it.
When coaching session data is captured and analyzed at each session — rather than once at the beginning and once at the end — a fundamentally different picture emerges. Instead of a line between two points, you have a time series: a sequence of measurements that reveals the actual trajectory of development with all its nonlinearity, regression, consolidation, and acceleration intact.
A longitudinal time series of complexity-weighted coaching data shows you things that a pre-post design cannot:
Consolidation patterns. After a genuine developmental advance, there is often a period of apparent regression as the new capability integrates. The leader seems to perform worse for a session or two before the new order stabilizes. Without longitudinal measurement, this consolidation dip looks like evidence that coaching is not working. With longitudinal data, it is recognizable as the signature of genuine structural change — the developmental equivalent of muscle soreness after effective training.
Context sensitivity signatures. Repeated measurement reveals which contexts systematically pull a given leader down to lower-order functioning. For some leaders it is perceived status threats; for others, ambiguity without clear authority. For still others, it is the presence of a particular person or role. This information is invisible in a 360 — it requires session-by-session observation over time.
Variance trajectories. As coaching progresses, the within-person variance in complexity level across sessions should, in a well-functioning engagement, decrease. The range between the leader's best sessions and worst sessions should narrow. This narrowing — the tightening of the developmental distribution — is a measurable coaching outcome with real organizational consequences. It can be quantified, reported, and compared across clients and engagements.
Leading indicators of regression. Longitudinal data eventually reveals the precursors to performance collapse: the early-session signals — particular linguistic patterns, a shift in reasoning architecture, a narrowing of the scope of consideration — that reliably precede a regression event. Coaches who can see these patterns in real time have a fundamentally different tool for intervention than coaches who can only observe after the fact.
For coaches credentialed at the ICF Master Certified Coach level — practitioners with demonstrated expertise in co-creating the conditions for deep, sustainable client change — the measurement gap creates a particular kind of professional frustration.
MCC coaches know, in their bones, that the development they facilitate is real and complex. They have witnessed the moments when a client's meaning-making structure genuinely shifts — when the frame expands, when a previously unconsidered perspective becomes not just cognitively accessible but emotionally metabolized. They have also witnessed the regression, the consolidation, the nonlinear journey that no pre-post design can adequately represent.
What they have lacked is a way to make this visible to the organizations that write the checks — a way to translate the lived phenomenology of deep developmental work into the language of reliability, consistency, and quantifiable capability improvement that business decision-makers require.
The ICF's own competency framework points toward this need. Competency 6.06 calls on coaches to notice "trends in the client's behaviors and emotions across sessions to discern themes and patterns." Competency 7.08 invites coaches to help clients identify "factors that influence current and future patterns of behavior, thinking or emotion." Competency 8.07 asks coaches to "partner with the client to integrate learning and sustain progress throughout the coaching engagement."
Each of these competencies, taken seriously, requires longitudinal data. They require not a snapshot but a trajectory. They require measurement infrastructure that can distinguish signal from noise across time — that can tell you whether the pattern you observed in session seven is consolidating into a stable developmental advance or retreating into familiar territory.
TruMind.ai's measurement architecture was designed to address exactly this gap. By analyzing session transcripts through the lens of both the MHC and Fischer's dynamic skill framework, TruMind produces not a single score but a developmental profile that evolves across time — tracking the nine dimensions of leadership capability (Adaptability, Coachability, Resilience, Boundary-Bridging, Charisma, Persuasion, Environmental Scanning, Strategy, and Digital Orchestration) alongside all eight ICF coaching competencies at each session.
This session-by-session architecture makes variance visible as a formal measurement object. TruMind tracks not only where a leader is performing on the MHC stage continuum but also how stable that performance is across contexts and time — the developmental equivalent of a process capability index.
For the coach, this means something that has not previously existed: objective, longitudinal evidence that their work is narrowing the gap between a client's optimal and functional levels. Evidence that the leader who once varied by two or more MHC stages across sessions is now varying by less than one. Evidence that the floor has risen, even when the ceiling — the peak performance that the client can achieve under ideal conditions — has remained at the same order.
This is coaching impact that was always real. It has simply, until now, been invisible.
The developmental science community has a language problem. The concepts that most precisely describe what coaches do — hierarchical complexity, optimal-functional level gaps, within-person variance, developmental consolidation — are not the concepts that HR directors, chief human resources officers, or CFOs reach for when evaluating program ROI.
Those decision-makers reach for concepts from domains they already trust: finance, engineering, operations. They ask whether the investment produced reliable returns. They want to know whether the capability improvement is real and repeatable, or whether they are paying for occasional flashes of brilliance that the organizational system immediately extinguishes.
This is where the parallel with Six Sigma and process capability indices becomes not a metaphor but a genuine bridge. When coaches can report that an engagement produced measurable variance reduction in a leader's developmental complexity — that the Cpk equivalent of his reasoning performance under stress increased meaningfully across the engagement — they are speaking a language that maps onto existing organizational quality frameworks.
The claim is no longer "our coaching produced growth." The claim is "our coaching produced reliability." And reliability is something organizations know how to value.
Practically, this suggests a reframing of how coaching engagements are scoped, conducted, and reported — particularly for coaches who work in organizational contexts with formal HR sponsors:
At contracting: Rather than committing only to directional movement (the client will develop), commit also to variance reduction as a measured outcome. Establish a baseline range of developmental complexity in the intake sessions. Define success partly in terms of how much the within-person variance has narrowed by the end of the engagement.
During the engagement: Use longitudinal data to have an honest, evidence-based conversation with the client about their developmental range. Most leaders are entirely unaware of how dramatically their cognitive complexity shifts across contexts. Making this visible — sharing the actual trajectory data with the client — is itself a powerful developmental intervention. It externalizes a pattern that the client could not previously observe, turning implicit variability into an object of reflective attention.
At reporting: Present the variance reduction data alongside the directional data. Show not only that the client's mean complexity level increased but also that the standard deviation of that measure decreased. Translate this into organizational language: the client's decision-making is now more predictable under pressure; her team can rely on consistent reasoning even in high-stress conditions; the risk of costly regression events has measurably declined.
For coach trainers and mentor coaches: Use the longitudinal profile as a teaching tool. The variance trajectory of a coaching engagement reveals something about the coach as well as the client — the sessions where complexity collapsed often correspond to moments where the coach's own presence or skillfulness wavered, where the scaffolding was withdrawn prematurely, or where the coaching approach did not match the client's developmental zone. This is professional development data of a kind that no self-report or supervisor observation has previously been able to provide.
Fischer's concept of the optimal-functional gap draws directly from Vygotsky's zone of proximal development — the space between what a learner can do independently and what she can do with skilled support (Vygotsky, 1978). The coach's role, in this framework, is not simply to provide insight or accountability. It is to provide the scaffolding that allows a client to operate temporarily at a complexity level above her current functional floor — and, through repeated supported practice, to gradually raise that floor.
This is a precise and demanding specification of what coaching actually does developmentally. It suggests that the most powerful coaching interventions are those that provide exactly the right level of challenge — operating in the Goldilocks zone between too familiar (which produces no growth) and too challenging (which produces collapse and regression). Too much scaffolding, and the client never builds independent capability at the higher level. Too little, and the client cannot reach the higher level at all.
The implication for coaching practice is significant. The coach's judgment about when to press and when to support, when to stay with a generative edge and when to back off and consolidate, is not merely an art. It is a developmental intervention with measurable consequences for the shape of the client's complexity trajectory over time. Getting it right — consistently placing the client in productive challenge rather than threatening overwhelm — is precisely what narrows the variance. It builds robustness at the higher order, not just occasional access to it.
Traditional coaching measurement requires a pause in the coaching process: stop coaching, administer the instrument, wait for results, return to coaching. This interruption introduces reactivity — the mere act of measurement changes the thing being measured. It also introduces a temporal gap between the data and the coaching session in which the data might be applied.
Unobtrusive measurement — the analysis of the coaching session itself, from the transcript that the session naturally produces — eliminates both problems. The coach does not stop to measure. The session proceeds as it always would. The measurement happens retrospectively, from the artifact that the coaching naturally generates, without altering the session's dynamics or imposing any additional burden on the client.
This matters for variance measurement in particular. Because unobtrusive measurement can occur at every session rather than only at prescribed intervals, it captures the full developmental trajectory including the consolidation dips, the regression events, and the plateaus that snapshot measurement misses entirely. The resulting time series is genuinely longitudinal — not two points connected by an imaginary line, but a continuous record of where the client actually was, session by session, across the arc of the engagement.
The coaching field has, for understandable reasons, measured itself by the standard that organizational buyers most readily grasp: did the average go up? The pre-post design, the snapshot assessment, the directional score — these are the metrics the industry has built its legitimacy around.
They are not wrong. But they are incomplete. They tell you whether the client's ceiling moved. They say nothing about whether the floor moved too, nothing about whether the variance narrowed, nothing about whether the developmental gains are robust to the stressors and context shifts that real organizational life continuously delivers.
The regression that keeps experienced coaches awake — the Thursday collapse of the leader who was genuinely brilliant on Tuesday — is not a mystery. It is the optimal-functional level gap made visible. It is Fischer's dynamic skill architecture operating exactly as science predicts. It is the clearest possible signal that there is still developmental work to do, not because the growth was illusory, but because the growth has not yet become robust.
When coaches can see this gap — when they have session-by-session longitudinal data that shows them the shape of the developmental distribution rather than a single summary statistic — they gain something that changes the nature of the work. They gain the ability to notice when variance is narrowing, to recognize consolidation events for what they are, to identify the specific contexts that trigger regression and address them directly.
And they gain, at last, the ability to stand in front of an organizational sponsor and say: your investment in this leader produced something that your previous assessments could not see. It produced reliability. It produced a leader whose worst-day reasoning is now measurably closer to her best-day reasoning. It produced a reduction in the variance of the most important cognitive process in your organization: the one that runs through her.
That is a claim that industrial engineers understand. That is a claim that finance officers understand. That is, finally, a claim that does justice to what skilled coaching has always actually accomplished.
The victory was always there. We simply lacked the measurement to see it.
Commons, M. L., & Richards, F. A. (2002). Organizing components into combinations: How stage transition works. Journal of Adult Development, 9(3), 159–177.
Commons, M. L., Trudeau, E. J., Stein, S. A., Richards, F. A., & Krause, S. R. (1998). Hierarchical complexity of tasks shows the existence of developmental stages. Developmental Review, 18(3), 237–278.
Fischer, K. W. (1980). A theory of cognitive development: The control and construction of hierarchies of skills. Psychological Review, 87(6), 477–531.
Fischer, K. W., & Bidell, T. R. (2006). Dynamic development of action and thought. In W. Damon & R. M. Lerner (Eds.), Handbook of child psychology: Theoretical models of human development (6th ed., Vol. 1, pp. 313–399). Wiley.
Fischer, K. W., & Pipp, S. L. (1984). Development of the structures of unconscious thought. In K. Bowers & D. Meichenbaum (Eds.), The unconscious reconsidered (pp. 88–148). Wiley.
Pande, P. S., Neuman, R. P., & Cavanagh, R. R. (2000). The Six Sigma way: How GE, Motorola, and other top companies are honing their performance. McGraw-Hill.
Taguchi, G., Chowdhury, S., & Wu, Y. (2005). Taguchi's quality engineering handbook. Wiley.Vygotsky, L. S. (1978). Mind in society: The development of higher psychological processes. Harvard University Press.