Three chapters in which the framework stops describing and starts calculating — and accepts the exposure that comes with it.
In which developmental positioning becomes calculable, and the framework acquires the one thing that can destroy it: a number that could be wrong.
The mathematics do not make the ocean blue. They tell you where it already is.
It repairs the one flaw capable of collapsing everything in Parts I and II — and it repairs it before deriving a single instrument, because an instrument built on a broken scale is worse than no instrument.
Then it builds the position coordinates, the throughput and capacity measures, the blue-ocean detection instruments, the readiness score, and the arrival-time inversion. It closes with the result that makes the whole apparatus commercially operative.
Every developmental strategy framework that assigns numbers to stages — Beige = 1, Purple = 2, and onward — commits the same silent error. It treats an ordinal scale as an interval scale, and then performs arithmetic on it. Means, standard deviations and differences all require equal spacing between units. Developmental colours do not supply that. Nothing in the colour model guarantees that the distance from Blue to Orange equals the distance from Green to Yellow. If the spacing is unequal, every mean and every variance computed on it is meaningless.
This is the single objection capable of collapsing the entire framework, and a competent quantitative reviewer needs about ninety seconds to raise it. §7.1 answers it. Everything after §7.1 depends on that answer holding.
None of them is hard. Several are one line. The difficulty in this chapter is never the algebra — it is the assessment that produces the inputs, which is judgement work dressed as data entry. A centre of gravity is only as good as the six scores behind it, and those scores are the part a practitioner should be nervous about. Treat every number here as carrying an error bar you did not write down.
Quantification confers false authority. The moment a developmental read carries two decimal places it will be quoted in a board paper as though it were a revenue figure, and the judgement underneath will become invisible. Report ranges. Report the evidence for each domain score. When someone asks for the number without the census behind it, decline — the number without the census is exactly the artifact this chapter exists to prevent.
The resolution is not to defend equal spacing in the colour model. The colour model cannot supply it and no amount of argument will make it do so. The resolution is to anchor the number line to something that supplies equal spacing by construction, and let the colours ride on top as qualitative texture.
That anchor is the Model of Hierarchical Complexity. Its orders are equal-interval not by assertion but by definition: each order n+1 is defined as an operation that coordinates two or more actions from order n, non-arbitrarily, on the outputs of order n. The ordering is therefore analytic rather than empirical. The spacing is a property of how the orders are constructed, not a finding that could have come out otherwise.
This gives a legitimate interval scale. The developmental colours become names for regions of that scale — which is, informally, what they were always doing. Nothing of the qualitative richness is lost. What is gained is the right to compute.
Two things follow, and the second is the reason this repair is cheap rather than catastrophic.
First, the correspondence between colour and order is a stated hypothesis, not a proof. It is falsifiable: if organisations independently scored at Green consistently produce reasoning artifacts at metasystematic rather than systematic order, the table shifts. It is published here so that it can be tested rather than assumed, and empirical recalibration is expected.
Second — and this is the part worth pausing on — because the conversion turns out to be linear with slope one, every instrument built on differences is completely unchanged by the anchoring. Bandwidth, velocity, variance, developmental distance, gap index, mismatch tension: all invariant. They were always measuring intervals, and intervals do not care where the zero sits.
Exactly one instrument breaks. Any measure that multiplies a raw level produces a quantity whose value depends on where you happened to start counting, which is not a real quantity at all. That measure is rebuilt in §7.7, and it needed rebuilding for independent reasons.
The anchoring therefore costs one equation and buys the entire numeric foundation. It is forced, and it is a bargain.
where s is the legacy colour index (the display variable) and o is the MHC order (the computational variable). Report in colours. Compute in orders. Never the reverse.
It does not claim that organisations pass through orders in a fixed sequence, that higher orders are morally better, or that an order can be inferred from a survey. It claims one thing only: that if you are going to do arithmetic on developmental position, the arithmetic must be done on a scale with defensible spacing. That is a claim about measurement, not about people.
Because they carry qualitative texture the orders do not. "Orange" evokes a recognisable organisational personality in a way that "order 10, formal" never will, and a framework nobody can feel is a framework nobody applies. The division of labour is deliberate: colours for the room, orders for the spreadsheet.
Table 7.1 is Calibration Hypothesis H1. It is the load-bearing empirical claim of this chapter, and it is stated as a hypothesis so that it can be attacked. Everything numeric in Parts III through VII rests on it.
Because the conversion is linear with slope 1, adding a constant to every position leaves every difference untouched. Formally, for any two positions o1 = s1 + 5 and o2 = s2 + 5:
Every instrument in the following pages that is built on a difference is therefore identical before and after the anchoring. This is not a convenience. It is the evidence that the repair was structural rather than cosmetic: a fix that changed all the answers would have been a new framework wearing the old name.
The order from which the organisation predominantly makes meaning under pressure. Not aspirational. Behavioural.
External communications is deliberately down-weighted to 0.5. It is the domain most easily performed above actual altitude, and weighting it heavily inflates every score an organisation will ever produce about itself.
The reading. Meridian is Blue–Orange with a genuine Blue floor, performing Green externally. Note what the table exposes that no single number could: external communications sits at 6.0 and conflict resolution at 4.0. The house markets at an altitude it cannot deliver under stress. That gap is the finding — not the centre of gravity, which merely locates it.
Conflict resolution serves as the floor proxy because it is the domain least amenable to performance — it reveals order under load. PG > 1.5 indicates a promise the organisation cannot consistently keep, with predictable costs: internal trust erosion, external relationship fragility, and high churn among clients who bought the performed altitude. Meridian's PG is 2.0.
Ceiling is the highest order sustainable under pressure, not in ideal conditions; floor is the order the organisation defaults to under significant stress. Wide bandwidth is not automatically an asset. An organisation with ceiling 12 and floor 8 has dB = 4 and is floor-fragile: under duress it becomes territorial and blame-driven, and in that moment the ceiling is irrelevant. Floor-raising precedes ceiling-raising wherever floor < 9.
Negative velocity is not automatically pathological. Regression toward far-from-equilibrium instability is frequently the precondition of a phase transition. What determines the outcome is not the regression but the organisation's interpretation of it: a system that pathologises its own disintegration and reimposes the prior order forecloses the transition it was in the middle of.
Managed variance (IDVm) is named, bridged and actively leveraged — an asset, because it constitutes bandwidth. Unmanaged variance (IDVu) produces unmetabolised friction and surfaces in the minutes as "personality conflict", "communication problems", or "culture issues". The strategic objective is never to minimise IDV. It is to convert IDVu into IDVm, and §7.10 shows what that conversion is worth.
Meridian, from Table 7.2: floor 4.0, ceiling 6.0, so dB = 2.0 — narrow, and the ceiling is performed rather than held. PG = 2.0 against a bandwidth of 2.0 means the entire span is being spent on the gap between what the house says and what it does under load. There is no bandwidth left over for actual range. This is a common and unlovely configuration, and it is invisible to any single one of the four measures.
Always floor before ceiling, and always IDVu before either. An organisation that raises its ceiling while its floor is unattended has widened the distance it falls, and the fall happens in precisely the moments — the crisis, the lawsuit, the funding gap — when the ceiling would have mattered most.
Earlier drafts of this framework carried two separate quantities: a mismatch tension defined between a role and a person, and a "complexity load" defined as environmental complexity minus organisational capacity. The second had no units — environmental complexity and organisational capacity were never placed on a shared metric, which makes the subtraction undefined. Once both are expressed in orders, the two equations collapse into one, and the merge yields something neither had alone: a holarchic ladder.
|T| > 1.5 produces chronic dysfunction that is routinely misattributed. Upward mismatch — the holder below the demand — reads to observers as not ready. Downward mismatch — the holder above the demand — reads as an attitude problem. Both diagnoses are wrong, and both are made daily.
T > 1.0 sustained and growing at the organisational level indicates accumulating developmental pressure and approaching phase-transition availability. This is the one reading in the chapter that is good news wearing bad clothing: the system is being asked for more than it can currently give, which is the only condition under which it reliably becomes more.
Organisation. The environment Meridian now operates in — capital entering the niche, attention fragmenting, distribution consolidating — demands roughly o = 11 (systematic) to read accurately. Meridian's CoG is 9.80. T = 11 − 9.80 = 1.20: sustained, growing, and below the dysfunction threshold. This is a house under genuine developmental pressure and still able to metabolise it. It is the best moment they will get.
Individual. Their managing director's role now demands o = 11; she reads at 12. T = −1.0. Downward mismatch. She is not disengaged, and the board's reading of her as "difficult about process" is the misattribution §7.5 predicts almost word for word. The role is too small for the holder, and the cost of that is measured in her eventual resignation rather than in any current metric.
Multiplicative, and the multiplication is the whole point rather than a modelling convenience.
Material and financial resource processing.
Signal generation, detection and response.
Coherent purpose sustained across uncertainty.
Rate of complexity increase in people and systems.
All four are normalised to [0, 1]. Any factor approaching zero collapses the product regardless of the others, and that is the assertion — not an artifact of choosing multiplication over addition, but the reason multiplication was chosen. A sum would let a strong balance sheet compensate for an absent purpose. Nothing in the observed behaviour of organisations suggests it can.
High M, high I, low Mn, low D. Take M = 0.9, I = 0.9, Mn = 0.1, D = 0.1:
Functionally dead, while every standard dashboard reads healthy. Revenue is fine. Reporting is fine. The failure, when it arrives, will be described as sudden. It was not sudden. It was legible the entire time to anyone measuring all four factors instead of the two that are easy to instrument.
M and I are the easy ones and every organisation already instruments them under other names — cash conversion, margin, cycle time, incident response. Mn and D are the ones nobody measures, which is exactly why the dangerous signature is so common. For Mn, ask what proportion of staff can state the organisation's purpose in a sentence that is not the marketing sentence. For D, ask how many people are doing work this year that they could not have done last year.
Because a sum permits substitution, and the empirical claim is that these four do not substitute. No quantity of cash buys purpose; no clarity of purpose survives insolvency. Each factor is a necessary condition, and the correct algebraic expression of a set of necessary conditions is a product. If you find a case where a genuinely near-zero factor is compensated, the model is wrong and should be reported as such.
This is the one equation the anchoring forced us to rebuild. Centre of gravity now enters as a difference from a declared floor rather than as a raw level, which makes the result a real quantity: orders of effective organising capacity.
The floor o0 = 8.0 is the MHC concrete order — the order below which sustained organisational coordination does not hold. Note a property worth stating explicitly: β·ν can exceed υ−1, meaning a wide-bandwidth, fast-moving organisation computes above its nominal capacity. This is intended. Such an organisation genuinely punches above its current centre of gravity, and a model that forbade it would be wrong.
Given CoG = 5.8 legacy → o = 10.8; dB = 1.8; dV = 0.4 orders/yr; IDVm = 1.1; IDVu = 0.6; horizon τ = 3 yr.
Effective capacity 2.29 against a nominal 2.80 — Aperture is operating at 82% of its own centre of gravity. Unmanaged variance is consuming nearly a fifth of the firm's capability before any of it reaches a client.
Note their relative force at typical values. β and ν are gentle — bandwidth and velocity each buy roughly a tenth to a fifth. υ is brutal: an IDVu of 0.6 costs 37.5% outright. The asymmetry is the finding, and it is why §7.10 reaches the conclusion it does. Unmanaged variance is the most expensive thing on this page and the cheapest to fix.
wB, wV and wU are engagement priors, not constants of nature. They are listed as Calibration in Table 7.7 for that reason. Before quoting a C(org) to a counterparty, state which weights you used — a capacity figure without its weight set is not a measurement, it is an opinion with a decimal point.
Leaders operate above emerging demand. There is no blue ocean on the developmental axis. Compete by translating downward — the market's problem is access, not altitude.
Emerging demand is developing beyond what the leaders can consistently deliver. A blue ocean is open at precisely the altitude the leaders cannot reach.
This is the central derivation of the framework. It converts blue ocean from a phenomenon recognised in retrospect into a condition detected in advance — which is the entire difference between a description and an instrument.
Sign alone is insufficient for positioning. Define provision density ρ(o) as the count of credible providers operating at order o, and demand density δ(o) as the count of buyers whose requirements sit at order o.
The competitor census by altitude — the empirical table of ρ(o) — is the primary evidentiary artifact of this method. It is what makes the claim auditable rather than merely asserted, and §7.11 shows why a dated copy of it is the difference between a defensible contingency contract and an unenforceable one. Produce it, and date it.
Aperture's census finds eleven credible leadership-development providers. Nine cluster at o = 10 (formal: competency models, 360s, behavioural frameworks). Two reach o = 11. None operates at 12. Weighted by revenue, CoGleader = 10.3.
Their buyer interviews find a cohort of chief people officers whose stated requirements — holding contradictory strategic logics simultaneously, developing others across altitude, designing for a system they cannot control — sit at o = 11.6. So DGI = 10.3 − 11.6 = −1.3. Negative. Void depth 1.3, with ρ(12) = 0 and δ(11.6) demonstrably greater than zero.
The ocean is real and the census proves it. Whether Aperture can enter it is a different question, answered in §7.10, and the answer at first computation is not yet.
Below +0.5 the offer registers as irrelevant — it tells people what they already know. Above +1.5 it registers as incomprehensible, or as threatening, which in practice are the same response wearing different words. This is the geometric basis of "meet people where they are", and unlike the slogan it has a computable range.
A 2.5-order span is spannable only if delivered in successive steps each ≤ 1.5 orders. The bridge range and the tolerance window are not in conflict, and the apparent contradiction between them dissolves once the relation is stated: the bridge is the integral of tolerance-window-sized steps. Earlier drafts gave both numbers without the relation, which read as incoherence and was in fact only incompleteness.
Blue ocean position is emerging demand + 0.5, not + 1.5. You are ahead by half an order, not by stages. Far enough to be the next place, close enough to be visible from where the market is standing, and able to receive it on arrival. Further ahead than that and you are invisible to precisely the people you built it for — which is the most common way a genuinely correct blue-ocean read still ends in failure.
Emerging demand sits at 11.6, so the blue-ocean position is 12.1 — not 13, which is where the temptation lies and where three of Aperture's competitors have already failed by publishing material nobody could use. Check it against the tolerance window: a buyer at 11.6 can receive an offer in [12.1, 13.1], so 12.1 sits exactly at the near edge. Recognisable as the next place, not the far one. Check the bridge: 12.1 − 11.6 = 0.5, below the bridge floor of 1.0, which tells you correctly that this is a product position rather than a developmental partnership position. Those are different businesses with different margins, and conflating them is how firms end up delivering bespoke transformation at the price of a workshop.
Capacity available against capacity required. Every term appears exactly once — earlier drafts placed velocity and unmanaged variance inside both the capacity index and the readiness expression, penalising variance quadratically and rewarding velocity twice.
Target CoG 7.25 (Yellow-integral) against a current 5.8, so Dtarget = 1.45. The move crosses Green → Yellow, the first-tier to second-tier boundary, so κ = 1.5.
The second figure is what happens when IDVu is converted from 0.6 to 0.2, which raises υ and lifts C(org) to 3.05. From marginal to underwritable on variance management alone — no new hires, no altitude move, no capital. Convert unmanaged variance before attempting the altitude move: it is the highest-leverage intervention available and it is cheaper than every alternative.
On gaming the score. Every term in BORS is supplied by the party being scored, which makes it trivially inflatable and therefore worthless as a self-report. It becomes an instrument only when the inputs are evidenced: the domain scores behind CoG, the census behind the target, the named and dated interventions behind IDVu. A counterparty accepting a BORS without those four artifacts is not underwriting anything. They are reading a brochure with a fraction in it.
Developmental ascent is sigmoid. Any framework that establishes logistic growth in one section and then computes arrival time linearly in another has contradicted itself, and the contradiction is not academic — it is worth about a year on a five-year contract.
Calibrate k from observed velocity at the current position: k = dV / [ (CoG − o0)(1 − (CoG − o0)/L) ]. The linear estimate t = D/dV is a first-order approximation valid only near the inflection point — outside that band it is materially wrong, and which way it is wrong depends on which side of inflection you sit.
The linear estimate gives 3.6 years. The logistic gives 4.6 — a full year longer, a 28% under-forecast. And the reason matters more than the number: Aperture's inflection sits at (CoG − o0) = L/2 = 2.5, and they are at 2.80. They are already past inflection and decelerating. Their fastest developmental year is behind them, and every subsequent order costs more than the last. Sign a gainshare on three and a half years against a true curve of four and a half and you have written a cheque against a model you did not invert.
If CoGtarget ≥ o0 + L, arrival time is infinite. The target sits at or above the ceiling of the current arc and no amount of velocity reaches it. Raising L is ceiling work, not positioning work — a different project with a different budget.
If dV ≤ 0, arrival is undefined. Positioning strategy is premature: the organisation is in regression, and the only question worth asking is whether that regression is transitional or terminal.
The general rule, worth memorising. Before inflection the curve is accelerating, so the linear estimate — anchored to today's slower velocity — is pessimistic; you will arrive early. After inflection the curve is decelerating and the linear estimate is optimistic; you will arrive late. Since almost every organisation commissioning this work is doing so precisely because it has already been developing for some years, the post-inflection case is the common one, and the common error therefore runs toward promising too much. Any contract priced on a linear estimate is, on the balance of cases, underpriced.
Resonance is not monotonic in the gap. It peaks at moderate distance — close enough to induct, far enough to pull — and decays at both ends. Let Δ = |CoGA − CoGB|, with optimal gap Δ* = 1.0 and Δmax = 6.0.
Higher-order party actively supports the other's development; the gap is acknowledged.
The gap is unacknowledged. Nothing malicious; nothing generative either.
Value flows one direction with no developmental return.
Bounded [0.5, 1.5] by derivation rather than by assertion. Behaviour check: at Δ = 1.0 generative, R = 1.50 — the induction sweet spot. At Δ = 0, R = 1.21: peers, some resonance, little pull. At Δ = 3.0 generative, R = 1.01 — good intentions do not close a three-order gap. At Δ = 3.0 extractive, R = 0.75: large power differentials make extraction more damaging, not less.
This is the arithmetic behind the most persistently puzzling finding in team research: brilliant individuals with poor relational quality operate at under half their composite capacity, while merely competent people with exceptional coherence operate at nearly double theirs. Raising Qrel is multiplicative; replacing individuals is additive. The multiplication is why the coherence intervention outperforms the talent intervention in nearly every case.
Positive margin compounds: the ecosystem becomes more capable, which raises the capacity of everything embedded in it, including you. Negative margin accrues systemic debt payable in talent attrition, reputational erosion, and ecosystem fragility. This is not an ethical overlay on the mathematics. It is a term in it — and §7.13 makes it a precondition of the contract rather than a preference of the contractor.
Earlier drafts made resonance rise monotonically with the gap under positive direction, which contradicted the prose sitting beside it and broke its own stated bounds. The error is instructive rather than embarrassing: it is what happens when a formula is written to express an intuition about direction without checking its behaviour at the extremes. A monotonic rule implies that a Turquoise consultancy is maximally generative to a Red street operation, which no practitioner believes and no engagement has ever shown. Test every kernel at Δ = 0 and Δ = Δmax before trusting it in between — most formulations that feel right in the middle fail badly at one end.
This is the result that makes the framework commercially operative rather than merely descriptive. The problem with every gainshare structure is attribution: if a provider claims to have created X in value, the auditor asks whether X was created or merely displaced — captured from a competitor, a supplier, or another line of the client's own business. On a flat plane that question is frequently unanswerable, because a flat plane is close to zero-sum by construction. Unanswerable attribution is what kills these deals. The developmental axis resolves it.
Let o* be the target altitude. If ρ(o*) = 0 at contract inception, then value created at o* has no prior claimant. It is net-new by construction, and the profit share is attributable to the provider.
Conditions 1–3 establish that the value is new; 4–5 that the provider can produce it; 6 that it is not being taken from somewhere unexamined.
BORS is not a self-assessment. It is an underwriting instrument — the answer to the only question a counterparty needs settled: can this firm deliver the altitude move it proposes, within the horizon, with margin? And DGI answers the question that kills these deals: is the value net-new or displaced? Together they constitute a computable condition under which value is provably created rather than transferred — the premise of abundance, in a form that survives a balance sheet.
Every value below is a declared calibration parameter, not a derived constant. The current values are priors from applied engagement, published here so that they can be tested and revised. Publishing them is what distinguishes applied mathematics from numbers that merely appeared. Nine are marked Calibration and require empirical refinement as engagement data accumulates; each is falsifiable against outcome data, and each is stated rather than hidden.
Return to Equation 1.1. A holon has π = 0.3, A = 2, cE = 10, cI = 4 and σ = 1.5. Compute dq/dt. Then find the value of cI at which flux is zero, and say in one sentence why that state is not the same as health.
Score your own organisation across the six domains of Table 7.2 using the default weights. Compute CoG in legacy index and convert to o. Compute PG. If your PG exceeds 1.5, name the specific promise your external communications is making that your conflict-resolution behaviour cannot keep — in one sentence, with no mitigating clause.
Build ρ(o) for your own market: list every credible provider and assign each an order, with one line of evidence per assignment. Then estimate δ(o). Compute DGI. If it is negative, state the void depth and the position you would take, and check it against the tolerance window. Date the table. This is the artifact §7.13 requires, and building it once teaches more than the rest of this chapter.
Take a firm at CoG − o0 = 1.2 with L = 5.0 and dV = 0.4, targeting 3.0. Compute arrival time both linearly and logistically. The error will run the opposite direction to Aperture's. Explain why in terms of inflection, and state the general rule for which side of inflection produces optimistic error.
Hypothesis H1 asserts a specific colour-to-order correspondence. Design an empirical study capable of falsifying it: state the reasoning artifacts you would collect, the independent scoring procedure, the sample, and the result pattern that would force the table to shift. Then state what would have to be true for the whole MHC anchoring — not merely the calibration — to be the wrong move. A framework whose author cannot describe its own disconfirmation is not a framework.
7B and 7C are the two that change how you see your own firm, and both take longer than they look — 7C is a week of real work, not an evening. Do them on your actual organisation rather than a hypothetical one; a hypothetical has no inconvenient facts in it.
You will be tempted to score competitors low. Resist it, and require a line of evidence for every assignment — a published artifact, an observed engagement, a client account. A census built from impressions produces the void you hoped for rather than the one that exists.
Worked solutions to 7A and 7D are in Appendix B, with the intermediate steps. 7B, 7C and 7E have no key — all three depend on facts only you can gather, which is the point of setting them.
Standard strategy frameworks map the terrain at a single altitude — usually formal, occasionally abstract — and are therefore structurally blind to the most decisive information available: whether the gap between an organisation and its market is widening or narrowing, and at what order.
The war metaphors that dominate strategic thinking are not wrong. Zero-sum reasoning is correct on a plane, and a plane is what those frameworks can see. They are dimensionally incomplete rather than mistaken, and the distinction matters, because you do not repair an incomplete model by arguing with its conclusions.
Adding the developmental axis does not overturn them. It embeds them as the special case that holds when every competitor is clustered at one order — which is, empirically, most of the time, and which is why the metaphors have worked well enough for a century.
But when demand moves and provision does not, the plane opens. That opening is detectable in advance, its depth is measurable, the readiness to enter it is computable, and the value created within it is attributable. Those four sentences are the chapter.
What happens to every instrument in this chapter when the holon is a joint venture, a platform, or a supply chain — and why aggregating altitude across subsidiaries is not an averaging problem.