Strategy for organisations that are alive, nested, and still becoming.
Strategy for organisations that are alive, nested, and still becoming.
This is a textbook, not a manifesto. Every chapter carries four things and carries them in the same order, so that you can enter at any point and know where you are standing.
The order matters. Argument first, because a formalism you cannot motivate is a formalism you will misapply. Formalism second, because an argument you cannot compute with will not survive contact with a budget cycle. Case third, because both are useless until they touch a real organisation with real people and a real problem. Exercises last, because you do not know a framework until you have used it on something that could embarrass you.
You may skip the formalism boxes on a first pass; the prose stands without them. You may not skip the exercises — several of the most important claims in these pages are established only by working a problem, and are merely asserted in the body text.
One organisation — Meridian Press, an independent publisher of forty people — is carried through all seventeen chapters. You will watch a single living system read at successively greater depth rather than meet a new anonymous company each time. Meridian is a composite of four real houses.
Formalism appears in boxes, always after the idea it formalises and never as decoration. Chapter 7 is the full primer; earlier chapters seed only the notation they need. Nothing here requires more than first-year calculus, and the equations are meant to be computed, not admired.
Developmental models are contested, and the strongest criticisms of them are good ones. They are answered in Chapter 3, at length, without hedging. If you find yourself resisting the altitude language in this chapter, that resistance is legitimate and it has an address.
Three chapters that replace the box with a membrane, the snapshot with an arc, and the analyst with a participant.
In which the most-used tool in strategy is taken seriously enough to be found insufficient.
A part is a whole, and a whole is a part, and the difference is only where you stand.
It recaps classic SWOT with genuine respect, locates the single assumption that limits it, replaces that assumption with the holon, and introduces the altitude lens that makes every quadrant readable from more than one height.
What it does not do is discard the grid. A tool used by every business school on earth for sixty years is not used because everyone is foolish; it is used because it solved something real, and Chapter 2 is devoted to naming exactly what. The claim here is narrower and harder: that the grid's limit is not in how it is filled but in what it assumes about the thing being filled in.
A word on the movements below. They are drawn as a wave rather than a ladder because the chapter does not climb: it goes out from the familiar tool, down into the assumption underneath it, across into a different ontology, and back up to something you can use on Monday. Readers who want the practical payoff first may read the fourth movement, then return to the first.
SWOT analysis is the most widely used strategic planning tool in human history. It was born from research at the Stanford Research Institute in the 1960s and 70s, is attributed to Albert Humphrey, and was refined through decades of consulting practice. It maps strengths and weaknesses — held to be internal — against opportunities and threats, held to be external, in a clean two-by-two grid.
Its beauty is its accessibility. A first-year business student learns it in five minutes. A seasoned chief executive still finds it useful. It fits on a whiteboard, a napkin, or a single slide. It works for start-ups and for the Fortune 500, for product launches and career decisions, for non-profit strategy and military planning.
Every MBA programme teaches it. Every consulting engagement begins with it. Every team offsite, at some point, has a whiteboard with four quadrants drawn on it. SWOT is so ubiquitous that it functions as the default language of strategic thinking — the way any conversation about where do we stand and where should we go begins.
The problem with SWOT is also its accessibility: almost everyone does it badly. People confuse aspirations with strengths, listing "innovative culture" when what they have is a culture that talks about innovation and ships the same product year after year. They catalogue weaknesses without recognising which are actually features of their developmental stage. They list opportunities they have no capacity to pursue and threats they have no power to influence. The grid gets filled. Strategy rarely emerges.
These are real failures of practice, and better facilitation improves them. Give a room a skilled facilitator, two days instead of two hours, and a discipline about evidence, and the quality of a SWOT rises considerably. This is worth doing and most organisations should do it.
But it is not the interesting failure. The interesting failure is the one no amount of facilitation reaches, because it is installed before the first quadrant is filled — before the marker is uncapped, before the room is booked, in the shape of the grid itself.
Notice what the two axis labels commit you to. Internal and external are not observations about your organisation; they are a claim that your organisation has an interior and an exterior, that the line between them is locatable, and that every fact worth recording sits cleanly on one side of it. That claim is made silently, by the furniture, before anyone speaks. And it is the claim this chapter is about.
A framework's assumptions are hardest to see precisely where they are most load-bearing, because they present themselves as the shape of the paper rather than as a position anyone took. Nobody argues for the axes. They are simply drawn, and then everyone argues inside them.
Forty people. Twenty-two years old. Eleven million in revenue, flat for three years. A backlist that is genuinely loved and a front list that is genuinely late. Last spring they ran a conventional SWOT. It produced fourteen strengths, nine weaknesses, six opportunities and seven threats, and not one decision. We will return to Meridian in every chapter of this book.
Note what the two axis labels are doing. They are not describing the organisation. They are asserting that it has an inside and an outside, and that every fact belongs to exactly one of them.
Attributes of the organisation that assist in achieving the objective.
Attributes of the organisation that obstruct achieving the objective.
Conditions of the environment that assist in achieving the objective.
Conditions of the environment that obstruct achieving the objective.
Try to place three of Meridian's real facts. Their best editor is married to their largest distributor's acquisitions lead. Their backlist is beloved by a readership that is ageing out. A rival house has begun paying advances Meridian cannot match, and three of their authors have stayed anyway, out of loyalty.
The marriage is a strength on Tuesday and a governance exposure on Wednesday, and it is a single fact. The ageing readership is the same fact as the beloved backlist, read at a different time constant. The three loyal authors are not an asset Meridian holds; they are a decision three people renew annually, for reasons Meridian has never asked about.
Each of these is a relationship, and a relationship has no interior. The grid does not fail to hold them because Meridian filled it in carelessly. It fails because the grid was drawn for objects, and none of these three facts is an object. Move the line anywhere you like; the facts will still straddle it, because they are made of the line.
The deepest flaw in standard SWOT is not methodological. It is ontological. It lives in what the framework assumes about the nature of the thing being analysed, before a single quadrant is filled.
SWOT assumes the organisation is a bounded object — a thing with clear insides and outsides, like a box or a machine. Strengths and weaknesses live inside the box. Opportunities and threats live outside it. The boundary between them is hard, clear and unambiguous. You look inward to assess yourself, outward to assess your environment, and then devise strategies to leverage the good and mitigate the bad.
This is a mechanistic view of what is, in fact, a living system. And when you misidentify the nature of the thing you are analysing, every conclusion drawn from the analysis inherits the misidentification. A SWOT built on a mechanistic ontology produces mechanistic strategies — strategies that optimise a machine rather than develop a living system.
Notice that this is not a complaint about rigour. The bounded-object model is perfectly rigorous. It is rigorous about the wrong object. And rigour applied to a misdescription is worse than no rigour at all, because it produces confident answers.
The tell is always the same: the categories blur exactly where the organisation is most alive. Is a twenty-year supplier relationship a strength or an opportunity? Is a cultural shift toward slower reading a threat to Meridian's release cadence, an opportunity for its backlist, or a signal that Meridian's own sense of what it is for is ready to change? The grid demands you choose. The living system refuses.
This blurring is not noise to be tidied up by a stricter facilitator. It is data. It is the model announcing its own boundary. Every place the quadrants blur is a place where the relationship, not the object, is doing the work.
There is a reason the bounded-object model felt natural for so long. It is inherited. SWOT was formalised inside a mid-century managerial worldview that took the firm to be a machine with inputs, outputs, and an engineering problem in between — a worldview that predates the living-systems turn in biology, complexity science, and organisational development by a generation. The framework did not choose its ontology. It absorbed the one available in the room where it was built.
The worldview has since changed. Cell biology stopped describing membranes as containers and began describing them as organs of exchange. Ecology stopped modelling populations in isolation. Systems theory stopped asking what a thing is made of and started asking what it is coupled to. The framework did not change with them.
What follows is not an attack on SWOT. It is an attempt to give it the ontology the rest of the sciences moved to sixty years ago.
Arthur Koestler coined holon in 1967 for an entity that is at once a whole and a part. Ken Wilber extended it into an account of reality as a holarchy — nested wholes, each transcending and including the one below.
Your editorial team is a whole, and a part of Meridian. Meridian is a whole, and a part of independent publishing. Publishing is a whole, and a part of the attention economy. At no level is there a hard boundary. There are only membranes.
Read this way, every quadrant changes character. A strength stops being a possession and becomes an expression of the quality of participation between holon and field. Meridian's backlist is not stored inside Meridian; it lives in the relation between an editorial sensibility, a readership, and twenty-two years of trust. Damage any term of that relation and the strength evaporates, though nothing "inside" the company has changed.
A weakness stops being an isolated deficit and becomes a place where that relationship has not yet found full expression — a growing edge rather than a fault. This is not consolation — a growing edge can still kill you. It is a claim about where to look: not inward for a broken part, but at the coupling that has not yet formed.
An opportunity stops being a prize waiting to be seized and becomes a signal from the larger holon about where it needs contribution. A threat stops being an external danger and becomes evolutionary pressure — the field reporting that some part of your configuration no longer serves the whole.
None of this is optimism. Immune systems are built by exposure, muscle by resistance, ecosystems stabilised by disturbance: living systems grow through pressure rather than in spite of it. The only open question is whether they grow consciously or reactively.
Everything below is developed properly in Chapter 7. It is stated here only so that the words membrane and health have referents you can compute with, rather than referents you must feel.
A holon is a triple H = (I, M, E) in which I is the interior state, E the field the holon is embedded in, and M the membrane: a selective map governing exchange between them. The bounded-object model is the special case in which M is a constant, impermeable, and therefore uninteresting.
A holarchy is a sequence H0 ≺ H1 ≺ … ≺ Hn where Hk ≺ Hk+1 means Hk+1 transcends and includes Hk — that is, Ek ⊇ Ik+1. Each holon's exterior is the next holon's interior. This is the whole content of "there is no outside".
Here q is the quantity exchanged, π ∈ [0,1] the permeability of the membrane, A its area, and cE, cI the concentrations either side. The term σ is autopoietic production — what the holon makes itself. A dead system has σ = 0 and runs to equilibrium, cE = cI. A living system holds a gradient against equilibrium, indefinitely, and this costs energy. Health is not equilibrium. Equilibrium is the technical description of a corpse.
Viability is not monotonic in π. At π → 0 the holon is sealed and starves: no information, no nutrient, no correction. At π → 1 there is no membrane at all and the holon dissolves into its field — it stops being a whole. Health lives in a window, and strategy is largely the art of finding and holding it. Meridian's editorial membrane is closer to 0.2 than they think: they read their market almost entirely through their own backlist.
Note the asymmetry of the two failures. Sealing is slow, comfortable, and produces excellent quarterly reports until it doesn't; dissolution is fast and obvious. Almost every organisation that dies of its membrane dies on the left of this curve, having spent years being told it was disciplined.
Standard SWOT assumes a single perspective — whoever holds the marker. But every quadrant looks different depending on the developmental altitude of the observer: the stage of consciousness from which meaning is made. Altitude determines what you can see, what you value, what you fear, and which strategies you are able to conceive at all. Below, the six altitudes named by their traditional colours, each with what strength means from that height and where the stage characteristically lives in the body. The second column carries the number that matters: colours name regions of a scale, and the order o — from the Model of Hierarchical Complexity — is the scale itself. Chapter 7 explains why arithmetic must be done on o and never on the colours.
Most SWOT conflicts inside organisations are altitude conflicts wearing the costume of strategic disagreement. What Orange calls a weakness — we are slow to decide because we include too many voices — Green calls a strength. What Blue calls a weakness — there is no clear chain of command — Yellow calls a feature. The marketing team and the people team are rarely disagreeing about facts. They are disagreeing about the altitude from which facts should be read.
Three cautions before using this table on anyone. Altitudes are not people: a person operates at several and moves between them under load — everyone goes Red at three in the morning. Later is not better: a room that cannot go Blue when Blue is required is not enlightened, it is unsafe. And the table describes what a stage can see, never what a person is worth. Used as a ranking it becomes a weapon, which is the subject of Chapter 3.
This is the single most useful formal statement in the chapter, and it is almost embarrassingly simple. Classical SWOT treats a quadrant assignment as a function of the organisation alone:
Luminous SWOT Ecology holds that the assignment is a function of the organisation, the altitude of the observer, and the moment:
Write Aa for the reading operator at altitude a. Then the hidden assumption of classical SWOT is exactly this: that Aa = id for every a — that altitude does nothing, that any competent observer reads the same event the same way. That assumption is false, and it is falsifiable in any room with two departments in it. Below, one event, four altitudes, four sound and incompatible strategies.
A rival house two streets away has raised a fund and begun paying advances Meridian cannot match.
A threat. Someone stronger is nearby. Crush them, or submit before you are humiliated.
A competitive signal. We cannot win on price, so differentiate. What can we offer that money cannot?
A relational question. What does this mean for our authors? Three stayed. Why? Ask them, and believe the answer.
An ecosystem dynamic. Capital has entered the niche. That reprices every house in it, including the rival. What does the field want?
Same external event. Four coherent responses. Which is correct depends on which altitude the situation actually requires — and, separately, on which altitude the organisation is currently able to operate from. Naming the gap between those two is the diagnostic work of Chapter 5.
The practical consequence is worth stating plainly, because it is the whole reason this formalism earns its place. If Aa ≠ id, then a SWOT with no altitude recorded is an unlabelled measurement. It reports a reading without reporting the instrument. You would not accept a temperature with no scale attached, and the four quadrants deserve the same courtesy: every item on a Luminous SWOT carries the altitude it was read from, and items read from different altitudes are never averaged.
Traditional SWOT spends half its energy on what is wrong. That deficit orientation creates a gravitational pull toward problem-solving rather than possibility-making, and it is why the exercise so often feels heavy. Appreciative Inquiry, developed by David Cooperrider, supplies a different operating question. Applied to each quadrant, it produces the four names used for the rest of this book. Chapter 6 builds them into the mandala; here they are simply introduced, with the question each one replaces.
The evidence for the renaming is not merely that it feels better. Cooperrider's original claim, and the body of work that followed it, is that the question a group is asked determines the range of answers it can generate: a deficit question recruits defensiveness and produces remedial actions, while a generative question recruits imagination. Same room, same facts, different yield. The four names above put that finding into the furniture, so the facilitator need not be unusually skilled for the room to work.
Renaming is not softening. "Emerging capacity" does not mean the weakness is fine. Meridian's front list is late, and late is late. The reframe changes what the fact is evidence of — not whether it is true, and not whether it is urgent.
Altitude language can be weaponised. Told that a colleague is "operating from Orange", most people hear a ranking, and rankings end inquiry. Chapter 3 takes this objection seriously rather than answering it; it is the strongest criticism the framework faces.
A room that has learned the new vocabulary but not the new ontology will produce a beautifully worded grid and the same four lists. If your Life-Giving Forces are the old strengths with better adjectives, nothing has happened yet. The test is whether any item now names a relationship rather than a possession.
Using only the Meridian Press facts given in §1.1 and §1.2, complete a conventional four-quadrant SWOT. Fill every quadrant. Do not skip anything because it is awkward. Keep the sheet; you will need it for 1B and again in Chapter 6.
Mark every item on your 1A sheet that could defensibly have gone in a different quadrant. For each, write one sentence naming the relationship it actually describes and the two holons that relationship runs between. Expect between three and six. If you found none, you have described a machine, not a publisher — go back.
Over one quarter, Meridian's editorial holon received 60 manuscript submissions from outside its existing author network and 240 from inside it. Treat π as the outside share of total inflow.
(i) Compute π. (ii) Locate it on Figure 1.3 and say which failure mode Meridian is nearer. (iii) The house wants a healthier membrane by next year. Give two interventions that raise qout and one that would raise π without improving anything — and explain why the third is the one a quarterly target will produce.
Take one live fact from your own organisation. Write the Red, Orange, Green and Yellow readings of it, in that order, each in no more than thirty words. Then mark which reading your last all-hands actually produced. That mark is your organisation's current centre of gravity, and it is more reliable than any assessment instrument.
Construct the strongest available argument that the bounded-object model is correct for a specific decision Meridian must make this quarter — that treating the house as a box with an inside and an outside will produce a better decision, faster, than treating it as a holon. Do not strawman it. Then say precisely what the holarchic reading buys that is worth its additional cost, or concede that for this decision it buys nothing. A framework that is never the wrong tool is not a tool; it is an identity. No answer is provided for this exercise.
Difficulty measures how much of the answer the chapter supplies, not how much thinking is required. 1A is easy because the method is handed to you; 1E is hard because it asks you to argue against the book while holding its terms accurately.
1B and 1D are better done with three or four people than alone: both depend on noticing a disagreement you would not have generated by yourself. Do 1A silently first, then compare sheets before discussing.
Worked solutions to 1A–1C are in Appendix B. 1D and 1E have none by design; both depend on facts only you have. If a colleague's 1D differs sharply from yours, that difference is the most useful finding either of you will get here.
How a tool built at Stanford Research Institute to explain why corporate planning kept failing became the thing corporate planning does instead of thinking — and what has to be honoured in it before anything can be added.