What Survived — An Inventory of Twenty-Five Years of Management Fashions, Audited
Attribution is a symptom of the fashion phase; invisibility is a symptom of absorption.
Executive Summary
This essay conducts a fashion-by-fashion audit of the major management ideas that have swept through organisational life since the turn of the millennium — Business Process Reengineering, Six Sigma, agile, big data, design thinking, OKRs — and asks not whether they worked, but what they left behind. The answer requires a decomposition lens that separates every fashion into three layers: its ritual (the ceremonies, certifications, and consulting engagements that die with the fashion’s consultants), its vocabulary (the jargon that dates like slang within a decade), and its mechanism (the core operational discipline that, if it was ever real, gets absorbed anonymously into how we work and stops being attributed at all). Applied fashion by fashion, the lens reveals a remarkably consistent pattern: rituals compost first, vocabulary second, and mechanisms — when they exist — endure precisely by becoming invisible. The predictive close turns the lens forward onto the AI-era management fashions now cresting, sorting the prompt-engineering priesthood and agent-maturity models from the eval discipline and workflow decomposition that look, by every indicator, like mechanisms. The lens itself is the essay’s real contribution — a tool the reader can carry forward and apply to whatever the next decade sells them.
The Programme Review Where Nobody Remembered
We were standing in a programme review — one of those quarterly rituals where the portfolio is spread across the wall and the senior team walks the board. The programme director was tracing a SIPOC diagram to frame a supplier-handoff problem that had been bleeding three weeks per cycle. A product owner was pointing at a kanban board, walking the group through the bottleneck where work-in-progress had been piling up between two teams who had never spoken to each other before someone made the queue visible. A data analyst had prepared a decision framework that began — as every useful one does — not with the data that happened to be available but with the question the business actually needed to answer before the next funding gate.
Three disciplines from three different decades, all in the same room, all being used unselfconsciously by people who could not have named their origins. The process map descends from Business Process Reengineering. The kanban board is agile’s grandchild. The decision-first framing is what survived the big-data gold rush. None of the practitioners in that room would have described themselves as followers of any of those movements. The movements are dead; the mechanisms are in daily use.
That room is the starting observation of this essay, and the question it raised is the one I want to work through: when a management fashion fades — and they all fade — what exactly is left behind? Not whether it “worked,” which is the question we always ask and which is almost always unanswerable at the level of a whole movement, but what specific operational deposit it left in the sediment of organisational practice. The answer, it turns out, depends on learning to see what a fashion is actually made of.
The Decomposition Lens
Every management fashion, on examination, turns out to be a composite of three layers. They arrive bundled — the consulting pitch never separates them — but they decompose at different rates, and the differences are what make the audit possible.
Ritual is the ceremonial apparatus: the certifications, the belt colours, the workshop formats, the maturity assessments, the two-day offsites with their prescribed exercises. Ritual is what the consulting industry sells, because ritual is packageable, trainable, and billable. It is also the first layer to die. When the fashion’s moment passes, the certifications stop being demanded in job descriptions, the workshops stop being booked, and the consulting practices pivot to the next wave. Ritual dies with its consultants.
Vocabulary is the jargon layer: the specialised language that marks insiders from outsiders. Reengineering. Defects per million opportunities. Sprint velocity. Data lake. Empathy map. Vocabulary serves a social function during the fashion’s peak — it signals membership, competence, currency with the moment. But vocabulary dates like slang. Within a decade, the terms either enter common usage so thoroughly that they lose their fashionable charge — nobody thinks of “process map” as BPR vocabulary any more — or they become embarrassing markers of a bygone era. Nobody says “reengineering” with a straight face in a board meeting in 2026.
Mechanism is the operational discipline at the core — if there is one. The feedback loop. The baseline measurement. The user observation. The outcome definition. Not every fashion contains a genuine mechanism; some are pure ritual and vocabulary, which is why they leave behind precisely nothing when they fade. But where a mechanism exists, it has a distinctive fate: it gets absorbed into the organisation’s working practice so completely that it stops being attributed to the fashion that introduced it. It becomes invisible. It becomes just how we work.
The decomposition lens: every management fashion is a composite of ritual, vocabulary, and mechanism. Ritual dies with its consultants. Vocabulary dates like slang. Mechanism — when it is real — survives by becoming invisible.
This is the lens I want to apply, fashion by fashion, to twenty-five years of organisational ideas. But the lens is also the essay’s real payload — a reusable tool the reader can carry forward and apply to whatever the next decade sells them.
The Audit
Business Process Reengineering
BPR arrived in the early 1990s and was, by the turn of the millennium, already deep in its composting phase. The ritual was dramatic: the clean-sheet redesign, the cross-functional team locked in a room for weeks, the consultancy engagement that promised transformation and frequently delivered redundancies dressed in process language. The vocabulary — “reengineering,” “clean sheet,” “process owner” — peaked and faded; “reengineering” is now used almost exclusively as a cautionary example in business school case studies.
But the mechanism survived completely. The discipline of drawing the process before changing it — of making the flow visible, of identifying where handoffs create delay and where rework loops hide — is now so embedded in operational practice that no one thinks of it as BPR. Process mapping is universal. Every programme manager reaching for a swimlane diagram is using BPR’s mechanism without knowing it. The reengineering theology died; the discipline of making the invisible visible before acting on it lives everywhere. It is the oldest mechanism in this audit, and arguably the ancestor of those that follow — each of the surviving disciplines we will encounter is, in its own way, a variation on the same instinct to see before acting.
Six Sigma
Six Sigma’s ritual was the most elaborate of any fashion in this inventory: the belt system — Green, Black, Master Black Belt — the certification industry, the project charters, the tollgate reviews, the dense statistical methodology that most practitioners never fully mastered and many quietly faked. At its peak in the early 2000s, a Fortune 500 company without a Six Sigma programme was an outlier. By 2015, the belt system was visibly composting — the certifications still existed but had lost their currency, and the elaborate ceremony of DMAIC tollgates had become a bureaucratic overhead that most organisations had quietly simplified or abandoned.
The vocabulary composted on schedule. Nobody in a contemporary programme review says “defects per million opportunities” or “critical to quality” unless they are in a manufacturing context where the statistical method is still genuinely applied — and even there, the language has softened into something less branded.
But the mechanism — the instinct to measure before opining, to establish a baseline before claiming improvement, to ask what does the data actually show? before accepting a narrative — that instinct survived. Where BPR made the process visible, Six Sigma made the measurement visible: the same instinct to see before acting, applied one layer deeper. It is now so thoroughly absorbed into competent management practice that it has no label. When a programme director asks to see the data before making a decision, she is not performing Six Sigma; she is doing what competent people do. But twenty-five years ago, the data-before-opinion instinct was not the default. Six Sigma, for all its ceremonial excess, installed it.
Agile
Agile’s audit is complicated by the fact that it is still, in 2026, visibly mid-decomposition — which makes it the most instructive example in the inventory, because we can watch the three layers separating in real time.
The ritual layer — the scaling frameworks (SAFe, LeSS, Nexus), the Scrum ceremonies (the sprint planning, the retrospective, the daily stand-up performed as ritual status report rather than genuine coordination), the Certified Scrum Master industry — is composting as we speak. Organisations that adopted SAFe wholesale are quietly dismantling the ceremonial apparatus while keeping the parts that work. The vocabulary is in transition: “sprint” and “backlog” have crossed into common usage and lost their fashionable charge; “velocity” and “burndown” are beginning to date.
The mechanism, however, is robust. Two things survived and will survive: the short feedback loop — build something small, show it to someone, learn, adjust, as against the twelve-month specification-then-build cycle it replaced — and the standing team, a stable group that owns a product or service, as against the project-staffing model where people were allocated as fractional resources across a portfolio and belonged to nobody. These two mechanisms are now embedded so deeply that organisations use them without thinking of them as agile. A team that releases fortnightly and reviews what it learned is not “doing agile”; it is doing what competent product development looks like in 2026. The feedback loop completes what BPR’s process mapping and Six Sigma’s baselining began — a trio of mechanisms that, taken together, amount to the discipline of closing the gap between action and evidence. The scaling frameworks are already sediment. The feedback loop and the standing team are infrastructure.
Big Data
The big-data fashion peaked around 2012–2015 and was remarkable for the purity of its ritual-to-mechanism ratio — an enormous apparatus surrounding a deceptively simple core insight. The ritual was vast: the data-lake architecture, the Hadoop clusters, the Chief Data Officer appointments, the breathless conferences promising that more data would automatically yield better decisions. The vocabulary — “data lake,” “data-driven,” “big data” itself — dated fast; by 2020, saying “big data” in a serious meeting marked you as behind the curve.
The mechanism that survived is easy to miss precisely because it is so simple: the decision-first question. The useful discipline that emerged from the wreckage of the big-data era was not about volume or velocity or variety — it was the hard-won insight that starting from the data you have is backwards. You start from the decision you need to make, then ask what information would change that decision, then go looking for it. This inversion — decision-first, not data-first — is the data-before-opinion instinct that Six Sigma planted, raised to the strategic level and applied not to a process but to an entire business question. It is now standard practice in competent analytics teams, and none of them attribute it to the big-data movement. They have absorbed the mechanism and forgotten the fashion.
Design Thinking
Design thinking’s ritual was the workshop: the empathy-mapping session, the Post-it note brainstorm, the prototype-and-test cycle compressed into a two-day offsite with a facilitator and a stack of coloured markers. As a consulting product it was brilliantly packaged — visual, participatory, and flattering to executives who could feel creative for an afternoon. The vocabulary — “empathy map,” “How Might We,” “prototype” — peaked around 2016–2019 and has dated fast. The workshop theatre has largely evaporated; the two-day design sprint is now a period piece.
The mechanism is watch the user. The discipline of observing how people actually use a service or product before redesigning it — of grounding decisions in observed behaviour rather than assumptions about what people want — survived and was absorbed. Where agile’s feedback loop shortened the build cycle, design thinking’s observation discipline asked whether the right thing was being built at all; the two mechanisms are, in retrospect, complementary halves of the same correction. It is now routine in competent product and service teams to spend time watching how their work is actually used. No one calls it design thinking. They call it user research, or they do not call it anything at all, because it has become invisible — which, as we have seen, is the sign that it has won.
OKRs
OKRs are the most recent fashion in this audit and are, in mid-2026, visibly mid-composting. The ritual — the quarterly OKR-setting ceremony, the cascading spreadsheets, the alignment meetings, the colour-coded dashboards tracking key results that nobody looks at between ceremonies — is dying in real time. Organisations that adopted OKRs wholesale are discovering what every fashion’s early adopters eventually discover: the ceremony is consuming more energy than the discipline is producing.
The vocabulary is still current but beginning to date. “Key results” and “objectives” as a paired term will likely sound as era-specific as “balanced scorecard” within five years.
The mechanism — if one survives — will be the outcome-over-activity instinct: the discipline of defining what success looks like in terms of outcomes rather than outputs, of asking what changed in the world? rather than what did we deliver? This is the decision-first question applied to planning rather than to analytics — a sibling mechanism from a different fashion. It is a genuine discipline, and it is likely to survive the composting of OKRs’ ceremonial apparatus. But the composting is not yet complete, so the audit here is necessarily provisional.
| Fashion | Peak | Ritual (composted) | Vocabulary (dated) | Mechanism (survived) | ||
|---|---|---|---|---|---|---|
| BPR | 1990s | Clean-sheet redesigns, cross-functional lock-ins | “Reengineering,” “process owner” | Process mapping — draw the flow before changing it | ||
| Six Sigma | Early 2000s | Belt certifications, DMAIC tollgates | “DPMO,” “critical to quality” | Data-before-opinion — measure before claiming | ||
| Agile | 2010s | SAFe, Scrum ceremonies, CSM industry | “Velocity,” “burndown” | Short feedback loop; standing team | ||
| Big Data | 2012–2015 | Data lakes, Hadoop clusters, CDO appointments | “Data-driven,” “big data” | Decision-first question | ||
| Design Thinking | 2016–2019 | Empathy-mapping workshops, design sprints | “How Might We,” “empathy map” | Watch the user | ||
| OKRs | 2020s | Cascading spreadsheets, quarterly ceremonies | “Key results” (still current) | Outcome-over-activity instinct (provisional) |
What the Pattern Tells Us
Run the audit across all six fashions and a consistent pattern emerges — one that says something important about how organisations actually learn.
First, the survival rate is remarkably consistent. Every fashion in this inventory contributed at most one mechanism to lasting practice. Not three, not five — one. BPR contributed process mapping. Six Sigma contributed the data-before-opinion instinct. Agile contributed the feedback loop and the standing team — arguably two, but they function as a pair and neither works without the other. The rest of the apparatus — the certifications, the consulting frameworks, the vocabulary, the ceremonies — composted on schedule. This is not a failure of the fashions; it is how organisational learning works. The fashion is the vehicle; the mechanism is the cargo. The vehicle is disposable.
Second, the mechanism survives precisely by becoming invisible. This is the most counterintuitive finding in the audit. We tend to think of survival as continued visibility — a fashion “survives” if people still talk about it, if it still commands conference keynotes and consulting revenue. But the opposite is true. The sign that a mechanism has truly taken root is that nobody attributes it any more. The moment process mapping stopped being called BPR, it had won. The moment the data-before-opinion instinct stopped being called Six Sigma, it had won. Attribution is a symptom of the fashion phase; invisibility is a symptom of absorption.
Third, the mechanisms form a family. This was not visible at the time — each fashion arrived as a self-contained revolution — but the surviving mechanisms, laid out together, tell a coherent story. Process mapping made the work visible. Baselining made the measurement visible. The feedback loop closed the gap between action and evidence. The decision-first question applied that discipline to strategy. User observation grounded it in behaviour. The outcome instinct pointed it at results. Taken as a set, they amount to a single cumulative lesson: see clearly before acting, and check what happened after you did. Twenty-five years of management fashion, and the deposit is one sentence.
Fourth, not every fashion contains a mechanism. Some fashions are pure ritual and vocabulary — they generate enormous consulting revenue, produce a wave of certifications, introduce a new jargon set, and leave behind precisely nothing when they recede. They compost completely. I have excluded these from the audit not to be kind but because the audit only works on fashions that had something real at their core. The reader can supply their own examples of the ones that did not.
The surviving mechanisms, laid out together, tell a single cumulative story: see clearly before acting, and check what happened after you did. Twenty-five years of management fashion, and the deposit is one sentence.
The Predictive Close — Reading the Fashions Arriving Now
The decomposition lens is not only retrospective. If the pattern holds — and twenty-five years of consistent evidence suggests it does — then it should be possible to look at the management fashions currently cresting and sort them, in advance, into their ritual, vocabulary, and mechanism layers.
The fashions of the moment are the AI-era management ideas: the organisational doctrines forming around how to adopt, govern, and scale artificial intelligence in the enterprise. They are arriving with the full apparatus — the conferences, the certifications, the consulting frameworks, the new C-suite titles, the breathless vocabulary. We are at peak fashion. Which means the decomposition has already begun, even if it is not yet visible to most participants.
The ritual layer is already forming and will compost first. The prompt-engineering priesthood — the specialist role, the certification programmes, the consulting practices built around the craft of talking to language models — has the signature characteristics of ritual. It is packageable, trainable, billable, and almost certainly temporary. As the models improve and the interfaces mature, the need for a specialised prompt-engineering function will fade, just as the need for a specialised “webmaster” faded in the 2000s. The agent-maturity models — the four-stage or five-stage frameworks that assess an organisation’s “AI readiness” — are maturity assessments, and maturity assessments are pure ritual. They are the AI era’s equivalent of Six Sigma’s belt levels: useful for selling consulting engagements, useless for building capability.
The vocabulary is peaking. “AI-native.” “Copilot.” “Agentic.” “Human in the loop.” “Responsible AI.” Some of these terms will cross into common usage and lose their charge; others will date within five years. The vocabulary layer always looks permanent from inside the fashion’s peak, and it never is.
The mechanisms — if the pattern holds — are the disciplines that will become invisible. Three candidates stand out, and they share the signature qualities we have seen in every surviving mechanism: they are genuinely operational, they are hard, they are not especially glamorous, and they resist packaging as a consulting product.
The eval discipline — the practice of defining what good looks like for an AI-assisted output before deploying it, of building systematic evaluation into the development cycle rather than relying on anecdote and demo — has the signature of a mechanism. It is the data-before-opinion instinct applied to a new medium. It will likely become standard practice without ever being attributed to the AI era that forced it into existence.
Workflow decomposition — the practice of breaking a process into its component tasks and asking, for each task, whether it is better performed by a human, a machine, or some hybrid, rather than applying AI wholesale to an entire process — is BPR’s mechanism reborn. Draw the flow before changing it, then decide what changes. It has the same characteristics: genuinely useful, not glamorous, likely to be absorbed into standard operating practice within a decade.
The autonomy envelope — the practice of defining, for any AI-assisted process, what decisions the system can make alone, what decisions require human review, and where the boundary sits — is a governance mechanism, not a governance framework. The frameworks, with their matrices and RAG ratings and maturity levels, are ritual and will compost on schedule. The autonomy envelope solves a real operational problem that does not go away when the fashion passes.
The reader equipped with the ritual/vocabulary/mechanism lens does not need this essay to tell them what survives. The lens works on its own. That is the point.
What Remains
“Every fashion decomposes. The question is whether you can see the mechanism before the ritual composts, and keep it.”
Twenty-five years is long enough to see the full cycle several times over. Long enough to watch the conference keynote become the consulting engagement become the corporate programme become the background practice become the thing nobody remembers choosing. The pattern is not cynical — it is, if anything, reassuring. Organisations do learn. They learn slowly, they learn wastefully, they learn through a process that involves enormous expenditure on ritual and vocabulary that will not last. But the mechanisms, when they are real, survive. They survive by becoming invisible. They survive by becoming just how we work.
The task for those of us who lead organisations through these cycles is not to resist management fashions — they are unavoidable, and the best of them carry real mechanisms that are worth the price of the ritual. The task is to learn to see which layer is which while the fashion is still in its peak, when the ritual, the vocabulary, and the mechanism are still bundled together and the consulting pitch makes no distinction between them. The decomposition lens is the tool for that work. It is, if the argument of this essay holds, one of those mechanisms itself — a way of seeing that outlasts the moment that produced it.
The fashions will keep arriving. They always do. The question is only ever what they leave behind.