Agile: Bought, Not Held.
The earlier essay on Agile-as-stance named three meanings the word now carries: a disposition the Manifesto authors articulated, a family of practices that emerged around it, and an organisational programme that grew up to deliver the first two. The essays since then have worked the first two meanings, with the second now substantially covered by the pair on Scrum and Kanban. What remains is the third.
The third meaning is the one most people encounter first, and frequently the only one they encounter at all. The certifications. The transformation initiatives. The frameworks sold as products. The consultancy engagements that install them. The aggregate of these is what most organisations mean when they say they have adopted Agile, even when the people doing the saying would also acknowledge, if asked carefully, that the stance and the practices are separate things from what was bought.
The claim this essay examines is the one the buying side makes when it commissions the programme. We can adopt the stance by buying its delivery. The claim is not usually phrased that bluntly, but the commissioning act assumes it. The practitioner the essay eventually returns to is the one who lives with that assumption’s consequences. They enter later. The system that produces those consequences is worked first.
The system being examined
The programme form is not one thing. It is a small system of actors, each operating inside their own incentive structures, each making locally rational decisions, each producing a piece of what the buying side eventually receives. Naming them is a precondition for the analysis, even though the analysis will treat them more abstractly than the naming suggests.
There are four. The first is the buying side itself: the organisation, usually represented by an executive sponsor, that commissions the programme and signs the cheque. The second is the certification industry, which credentials individuals as proficient in the practices the programme installs. The third is the framework producers, who design and maintain the products that the programme delivers. The fourth is the consultancies, who deliver those products into the buying organisation and stay long enough to declare the engagement complete.
A note on the supply side. Frameworks and consultancies are not parallel actors in the sense that buyers and certifiers are. They stand in a producer-deliverer relationship. The framework is a product, designed once and offered repeatedly. The consultancy is the mechanism by which the product reaches the buyer’s organisation. Many consultancies deliver multiple frameworks. Many frameworks are delivered by multiple consultancies. The market is fuzzy at the seam: some consultancies have produced their own frameworks, some framework bodies have commercial arms that look consultancy-shaped. Certified individuals, working as contractors or as employees inside the buying organisation, also operate as delivery agents in the same structural role, and for the purposes of this analysis are folded into the consultancy category that the essay uses throughout. But the producer-deliverer logic is real underneath the fuzz, and treating the two roles as one would lose something the analysis needs.
The analysis treats the four actors structurally. No specific institutions or frameworks are named. The reader will recognise the shapes regardless. The point of the analysis is not to single anyone out; it is to show what the system constituted by these four actors structurally cannot deliver, and what that costs the people who eventually work inside what it does deliver.
What would have to be true for the stance to be commissionable
The buying-side claim, in its strongest form, is that the stance can be acquired through commission. That a disposition characterised by provisionality, empiricism, and orientation toward uncovering better ways of working can be installed in an organisation by an external party operating to a defined engagement scope, on a defined timeline, against a defined deliverable.
This is an existence claim about the stance: that it is the kind of thing a commission can transact in. It assumes the stance has the shape of a deliverable.
The shape of a deliverable is specific. A deliverable can be defined in advance, against criteria the buyer and the supplier agree to. It can be produced through identifiable activities. It can be measured for completion. It can be handed over and signed off. These four properties are what commissioning consists of. Without them, the transaction has no structure to occur within.
The stance has none of these properties. It is the disposition to not commit prematurely to definitions, to let what works be discovered through doing, to treat current understanding as provisional. The criteria for its presence are not agreeable in advance because the stance is precisely the willingness to find that the criteria one would have agreed in advance were the wrong criteria. Activities that produce it are not identifiable because the stance is constituted in part by remaining open to what activities the work actually requires. Completion cannot be measured because the stance has no completion state; an organisation that holds the stance continues to hold it through ongoing practice, and an organisation that no longer holds it has not failed to deliver but ceased to embody.
The four actors each require the stance to have a shape it does not have, for reasons specific to their role. The buyer needs it to be a deliverable they can procure against. The certifier needs it to be a property of individuals that can be verified through examination. The framework producer needs it to be encodable as a structure of roles, events, and artefacts that survive being written down. The consultancy needs it to be installable through an engagement that can begin, conclude, and be invoiced. None of these are bad-faith requirements. Each is the natural shape the role requires the stance to take in order for the actor to participate in delivering it.
The shape the actors require is the shape the stance does not have. Not in some elusive or romanticised sense, but in the precise sense that the properties commissioning depends on are the properties the stance is constituted by the absence of. The category error is structural to what the stance is and what the actors can transact in.
This is where the essay on change first explored a closely related point, though at a different level. That essay described how change-management approaches treat the conditions for change as themselves deliverable, when the conditions are produced by the system the change is meant to alter. The wrong level was the level at which the intervention was aimed. The wrong shape, in this essay’s territory, is the shape the intervention assumes its target has. The error has a similar structure. It mistakes the form of the system that does the delivering for the form of what is being delivered.
What this means for the practitioner, who is the figure the essay returns to later, is that they enter an organisation where the work is expected to conform to a structure that was specified before the work began. The structure was specified that way because that is the form the programme could take. The stance the structure was supposed to embody is the disposition to not specify in that form. The practitioner does not encounter this as a philosophical contradiction. They encounter it as the daily fact that the work, which is irreducibly uncertain, must be reported, planned, and reviewed as though it were a deliverable being executed against a specification. That fact is structural to the system the practitioner has been hired into. It is the existence assumption’s failure showing up at the level where the work happens.
What would have to be true for the programme to produce the stance
Granting, for the sake of the next step, that the stance could be commissioned at all, a second assumption is doing work the first one needs help carrying. The buying side assumes that the programme, the integrated activity of certification, framework adoption, and consultancy engagement, actually produces the stance it claims to deliver. This is the mechanism assumption. It asks whether the actors’ selection effects, in combination, produce practitioners and organisations that hold the stance.
The selection effects do not run in that direction.
Certification selects for fluency in a vocabulary. The credential is awarded to individuals who can demonstrate, in an examination setting, that they have absorbed the framework’s definitions, can identify its roles, can describe its events, can recognise its artefacts when presented with them. These are not bad things to know. The question is whether knowing them produces, or even predicts, the disposition the framework is supposed to express. Fluency in describing the stance and operating from the stance are different capabilities. The first is what the examination can test; the second is what the examination cannot. The certification mechanism reliably produces practitioners who can describe the framework, because that is what its selection effect actually selects for. Whether those practitioners hold the stance is a question the credential does not answer and structurally cannot.
The framework producers’ selection effect operates differently. A framework that survives in the market is one that can be taught, demonstrated, audited, and adopted at scale by organisations under pressure to demonstrate that they have adopted something. The market selects for frameworks that have clean surface signatures the buying organisation can show to its stakeholders: roles assigned, events scheduled, artefacts produced, training completed, certifications obtained. These are not incidental features of frameworks that succeed. They are why those frameworks succeed in the market. A framework that delivered the stance but had no clean surface signature would not propagate, because the buying side, the consultancies, and the certifiers all depend on visibility to participate in delivering it. The frameworks the market produces are the frameworks the market’s selection mechanism permits, and that mechanism permits surface-visible frameworks over stance-delivering ones whenever the two come apart. They come apart often.
The consultancies’ selection effect operates at the engagement level. A transformation engagement is sold against a roadmap and judged against milestones: assessment completed, structures stood up, training rolled out, practices observed in use. These are the artefacts of a programme that can be reported on. Whether the organisation, after the consultancy has left, holds the stance is a longitudinal question that the engagement structure was never designed to answer. The consultancy’s professional incentives reward the visible execution of the roadmap. Whether the visible execution corresponds to the stance is, again, a question the mechanism does not answer. By the time the answer becomes available, the engagement is concluded, and the cost of finding out that the answer is unfavourable is borne by the buying organisation, not by the supplier whose engagement produced it.
Get Bruno Baketarić’s stories in your inbox
Join Medium for free to get updates from this writer.
Two earlier essays do most of the structural work here. The chain-of-pressure dynamic described in one of them is operating throughout this mechanism. The buying organisation commissions the programme under pressure, often substantial, to demonstrate Agile capability to its board, its market, or its own employees. That pressure converts a question the stance treats as open, what would actually serve our work, here, in our conditions, into a procurement decision: which programme do we buy. The conversion is not a failure of intent on the executive’s part. It is the chain doing what it does, with the cognitive space for the open question compressed at every link above the decision-maker, and frequently at the decision-maker’s own link too. The four actors, in turn, supply the form the procurement decision needs. They are not producing the conversion. They are producing the supply that becomes available because the conversion has happened.
The earlier essay on management formation does the rest. The buying-side practitioner is unlikely to receive the stance even if the programme delivered it, because the disposition the stance requires is the disposition the manager’s formation has selected against. The mechanism assumption fails not just because the supply side does not deliver the stance, but because the demand side could not receive it if it did. The two failures are not redundant. Each one is sufficient on its own. They compound.
Here a small distinction inside the supply side is worth surfacing, because it changes how the mechanism fails. The certifier’s mechanism failure is downstream of an individual examination. The framework producer’s mechanism failure is downstream of market selection over the framework as a product. The consultancy’s mechanism failure is downstream of an engagement’s structural design. These are different failures at the operational level, but they share something at the structural level: each one selects for a surface property (vocabulary fluency, market viability, milestone completion) that the actor’s role can transact in. None of them selects for the disposition the stance is constituted by. The actors are not aligned by intent; they are aligned by the structure of what each can transact in, which is the same structure: visible, defined, deliverable. Across four actors, the same selection logic runs four times, and each time it filters out the same thing.
What would have to be true for the programme to be the appropriate route
The third assumption is the one most rarely examined, in the same way the exclusivity question is most rarely examined for any popular claim. Granting, for the sake of this step, that something could be commissioned and that the programme could produce something useful, the buying side still assumes the programme route is the appropriate adoption mechanism, against the alternatives.
The alternatives are not other practices. The pair of essays on Scrum and Kanban examined that question. The alternatives at the programme level are other routes to adoption: slow internal evolution, locally grown practice, the stance acquired through years of doing the work and reflecting on it with sustained attention. These routes exist. They are how the stance was acquired, demonstrably, by the practitioners who first articulated it. They are how it continues to be acquired by individuals and small teams who pursue it through accumulated practice. They are not theoretical alternatives. They are the routes that produced the thing the programme claims to deliver.
The buying side does not, in most cases, weigh these alternatives against the programme route. The reason is the same reason that runs through the earlier essay on Agile-as-stance, operating one level up. The alternative routes cannot be commissioned. They have no supply side. They do not present as procurable. An executive who asks the procurement department to commission slow internal evolution will not receive a quote. The alternative is not absent from the option space because it is bad. It is absent because the option space is constituted by what can be procured, and the routes that cannot be procured are not in the space the executive is choosing from.
This is local optimisation operating at the system level. Each actor on the supply side, locally rationally, produced what they could sell. The certifiers produced examinations and credentials, because that is what could be assessed and priced. The framework producers produced structured products, because that is what could be packaged and propagated. The consultancies produced engagement models, because that is what could be sold and delivered. None of them produced slow internal evolution, because slow internal evolution is not a product. The aggregate of these locally rational supply decisions is an option space in which the alternatives that would have delivered the stance are systematically absent. The buying side then makes a selection from the available options, locally rationally, against criteria the chain of pressure has shaped. The selection is real. The space it was made from is the space the supply side’s local optimisation produced.
The exclusivity assumption fails for a reason the section has been building toward. The alternatives were not in the space within which the selection happened. The reader is asked to notice the structure of that absence. The option space the executive chooses from is not the option space that contains all the routes to the stance. It is the option space that contains all the routes that can be procured to the stance. These are not the same set. They have only loose overlap, and the overlap is shaped by what could be made procurable, not by what would actually deliver.
A note on tone here, because the analysis has now arrived at the part of the essay where the four actors look most like the targets of a critique, and they are not. The actors did not design the option space. They populated the part of it they could supply, under their own incentive structures, often with genuine belief in what they were producing. Many of the people working inside these institutions are doing serious, careful work within the form they have been given. The form they have been given was selected through a process they did not control, the same chain-of-pressure dynamic that shapes most of what organisations end up adopting. The actors are inside the system the essay is describing, not outside it. The verdict the essay declines to issue is the verdict it declines to issue on them as well.
The practitioner
The system the four actors constitute functions, at the level of those actors, on its own terms. Programmes are commissioned. Certifications are issued. Frameworks are deployed. Engagements are concluded and invoiced. The economics, internal to the supply side, are coherent. The reporting, internal to the buying side, demonstrates what was bought. From inside any of the four actors’ positions, the system is not visibly incoherent. It is operating.
The incoherence shows up somewhere else. It shows up where the work happens, in the daily experience of the practitioner who lives inside the organisation after the programme has been installed and the consultancy has left. By practitioner here, the essay means the engineer, the team lead, the internal coach, the technical specialist, the product person inside the buying organisation: the people whose actual work is the building of software, or the supporting of those who build it from inside the same organisation. Not the consultancy’s coaches, who are part of the delivery mechanism and have been trained in the framework the consultancy sold. The buyer-side practitioner: several links down the chain from the executive who commissioned the programme, with no direct relationship to the supply side, navigating the gap between what the work actually requires and what the programme requires the work to look like.
Some of these practitioners arrived hoping for the stance the Manifesto named, and discovered the programme instead. Others did not seek the role at all; the programme arrived around them and they were enrolled into it without consultation. The two groups have different relationships to the disappointment. What unites them is that the system’s incoherence shows up at the level where they operate, in ways that are not visible at the level where the system is constituted.
The shape of that experience tracks the three assumption failures already worked through. The existence failure shows up as the gap between the work, which is irreducibly uncertain, and the structures the work is required to be reported and reviewed against, which assume the work is a deliverable being executed against a specification. The mechanism failure shows up as the recurrent experience that the disposition that would actually serve the work, the disposition to hold understanding provisionally, to revise commitments as evidence accumulates, to refuse premature definition, is the disposition the surrounding system reads as insufficient seriousness, insufficient alignment, insufficient commitment. The chain of pressure compresses the cognitive space the disposition requires, link by link, every day. The exclusivity failure shows up as the slow realisation that the route the organisation is on is not, and could not be, the route that would have delivered the thing it claims to be delivering, and that the routes that might have delivered it were not available to the organisation when the choice was made.
None of this is visible from the level of the four actors. The buying-side executive sees the programme operating; the consultancy sees the engagement completed; the framework producer sees the framework deployed; the certifier sees the credentials issued. The practitioner sees the gap, every day, and bears the cost of it in their work. The cost is asymmetrically distributed, in the same sense the chain-of-pressure essay described. The decisions that produced the system were made at the top; the consequences of those decisions are borne several links down. The practitioner is the locus where the system’s structural failure shows up, not because they are the cause of it, and not because the essay is arrayed in their defence against the actors, but because the structure of the system places the cost of its incoherence where they happen to stand.
This is not a complaint on their behalf. It is an observation about where the analysis lands when it is followed all the way through. The four actors’ assumptions, taken together, jointly require the stance to be what it cannot be. The system the assumptions produce is internally coherent at the actors’ level and structurally incoherent at the practitioner’s. The reader, having followed the analysis, has the structure visible. What they do with that visibility is theirs.
The recursion
The analysis rests on its own assumptions. The four-actor model is a position another framing could redraw. The decision to spine the essay on the buying side, with the supply side as the supply the demand produced, could legitimately be reversed. The treatment of alternative adoption routes as effectively absent from the option space is an empirical claim that admits exceptions worth a longer examination than this essay gave them. The practitioner figure as the locus of the system’s structural failure is a move warranted by the analysis but not exhaustively defended.
The condition is the one the foundational essay set out, and applying the move to this essay is what the foundational essay asks of every essay including itself. The analysis went one or two layers below where the programme form is usually examined. That is the work it was for.
Close
The programme form was the form the adoption of Agile took because it was the form that could take that shape. Each part of the system that produced it was operating inside its own incentive structures, locally rationally, often with genuine belief in what it was producing. The aggregate of those locally rational decisions is a system whose four actors, taken together, jointly require the stance to be reducible to a form the stance is constituted by the absence of.
The system continues to function at the level of the actors. Programmes continue to be commissioned, frameworks continue to be deployed, certifications continue to be issued, engagements continue to conclude and be invoiced. Whether what the system is producing is what was being bought is a question that the structure of the participating roles does not, on its own, equip the participants to ask. The practitioner inside the buying organisation, several links down from the commission, is the one who lives with the answer regardless of whether the question is asked.
Whether the route the organisation is on is the route that delivers what was meant to be delivered is the question the analysis has tried to make visible. The reader is in a position to ask it, of their own situation, with what the analysis has made available. That is what the move is for. It does not settle anything. It makes the weight visible before it collapses, which is the most that careful thinking, at this level of distance from the work, can do.
A note on what this concludes. The assumption-analysis arc the series has been running ends here. The foundational essay set out the move. The essay on Agile-as-stance applied it to the disposition the Manifesto authors articulated. The pair on Scrum and Kanban applied it to the family of practices that grew up around the disposition [1, 2]. This essay has applied it to the organisational programme that grew up to deliver both. What the three meanings have in common, and where they part company, the move has now had the chance to show. Other essays might follow on other questions. This thread closes here.
