The Consulting Dependency Cycle
Dependency is a choice an organisation makes and then mistakes for a fact about the world.
Executive Summary
A large programme finishes. The integrator’s people pack their laptops, the final invoice clears, and within a year the organisation is quietly drafting a brief for the next firm. The capability it paid so handsomely to acquire has walked out of the building inside the heads of people who were never on the payroll. This essay examines why that pattern — the consulting dependency cycle — recurs across organisations that are neither naive nor short of money, and why it survives every earnest resolution to build the capability in-house this time.
The argument is that dependency is manufactured, not inherited. It is produced by a handful of structural forces: the asymmetry of risk between a permanent hire and a terminable contract; the career incentives that reward visible, decisive spending over patient capability-building; the knowledge-transfer clause that both parties sign and neither enforces; the atrophy of the internal muscles required even to specify the work; and the plain asymmetry of memory between a firm that has run this play forty times and a client running it once. None of these is a failure of intelligence. Each is a rational local choice that aggregates into an irrational whole.
The essay also refuses the comfortable opposite conclusion — that the remedy is to bring everything in-house. There are sound reasons to buy help, and the fantasy of total self-sufficiency is its own expensive trap. The distinction that matters is not between buying and building but between purchasing expertise in a way that feeds your own judgement and purchasing it in a way that starves it. What follows sets out the machinery of the cycle, illustrates it with a costed composite, and proposes four structural remedies: buy the transfer of capability as the deliverable rather than as a hope appended to it; keep the client as the general contractor rather than a subcontractor to its own integrator; ring-fence the reasoning behind decisions as an owned asset; and measure the firm on the client’s independence at the point of exit, not on delivery alone.
The Morning After Go-Live
Picture the morning after a large enterprise system goes live. The war room, which for eighteen months ran on cold coffee and burndown charts, is being dismantled. The offshore development pod is already being redeployed to another account. The programme director — who chaired every steering meeting, managed the political weather, and defended the budget through two re-baselines — walks the floor with a curious feeling: the thing is done, and he does not understand how it works. He never wrote a configuration decision. He never sat in the design authority when the awkward trade-offs were settled. Those decisions were made competently, at pace, by people who are now on a plane.
Three weeks later the first serious production incident arrives. An overnight batch fails, a reconciliation does not balance, and a business process that used to take four hours now takes the whole day. The internal team gathers, and the same sentence is spoken in a dozen different ways: nobody here knows why it was built like this. The design intent — the reasoning, the rejected alternatives, the small compromises that explain the system’s odd corners — was never an artefact the organisation owned. It lived in conversation, and the conversation has left. So a call is made to the firm, a change request is raised, a rate is agreed, and the people who built it return to explain it. The dependency has begun, and it began not at some later moment of weakness but at go-live, at the very peak of apparent success.
This is the observation the essay turns on. We are inclined to narrate consulting dependency as a story of decline — an organisation that lost its edge and now leans on outsiders. In practice the reliance is engineered into the shape of the engagement from the first day. The programme was structured so that outsiders would hold the knowledge, and the predictable consequence is that outsiders continue to be needed. Dependency is not what happens when the plan fails. It is what happens when the plan succeeds on its own terms.
The Honest Case for Buying
Before dissecting the cycle it is worth defending the thing being criticised, because the case for buying help is stronger than its critics allow, and an argument that cannot survive its opponent’s best version is not worth making.
There are at least four honourable reasons to bring in a firm. The first is genuine surge: a transformation is, by definition, a temporary spike of effort, and it would be perverse to hire a hundred permanent people for a two-year push and then carry them, idle and demoralised, for a decade afterwards. The second is scarce specialism: some skills are used so rarely that no single organisation can keep them sharp — the deep tax structuring, the one enterprise package configured a certain way, the regulatory remediation that comes round once in a career. It is entirely rational to rent expertise you would never exercise often enough to retain. The third is independence: an outsider can say to a chief executive what no subordinate dares, and the abrasion of a challenging external voice is worth paying for precisely because it cannot be produced from inside the hierarchy. The fourth is memory across industries: a firm that has run the same class of programme for a dozen clients carries pattern-recognition that no first-time client can possibly hold.
Set against these, the fantasy of total self-sufficiency is its own trap. The organisation that insources everything on principle tends to end up with a comfortable population of generalists, no external benchmark, and a dangerous belief in its own competence. Isolation breeds a particular kind of mediocrity — fluent in its own conventions, blind to how far it has drifted from the field. So the sin is not buying help.
The problem is not that organisations buy expertise. It is that they routinely buy it in a form that guarantees they will have to buy it again — and again — for the same class of problem.
The remainder of this essay is therefore concerned with a narrower and more useful question than build or buy. It asks why the buying so reliably takes the dependency-producing form, and what a different form would look like.
The Machinery of the Cycle
If dependency were merely foolish it would not be so durable. Its persistence tells us it is being produced by forces that make it locally rational for almost everyone involved. Five are worth naming.
| Force | How it looks in the room | Why it endures |
|---|---|---|
| Risk asymmetry | “Let’s not add headcount for a temporary programme” | A contract can be terminated; a permanent hire is a fixed cost and a political liability if the programme fails |
| Career incentive | “The board wants to see we’ve brought in serious help” | Hiring a brand-name firm signals decisiveness; building a quiet internal team produces nothing visible this quarter |
| Transfer theatre | “Knowledge transfer is in the statement of work” | Neither side’s incentives reward real transfer, so it becomes a line item that is cut the moment the timeline slips |
| Capability atrophy | “We’ll need them to help us scope the next phase” | Once the muscle to specify and challenge the work is lost, outsiders are needed even to decide what to buy |
| Memory asymmetry | “They’ve done this so many times before” | The firm remembers across engagements; the client forgets between them, so the firm’s relative advantage compounds |
The first, risk asymmetry, is the quiet engine of the whole thing. A permanent hire is a fixed cost that sits on the establishment forever and becomes a redundancy problem if the work dries up; worse, if the programme fails, the executive who built the team owns that failure personally and permanently. A contract is a variable cost that can be switched off, and if the programme fails, the firm can be blamed and dismissed. Faced with this, a rational executive is not really buying capability at all. He is buying insurance — transferring career and financial risk to a third party — and insurance, by its nature, is something you keep renewing.
The second force, career incentive, compounds the first. Consider two directors. One spends eighteen months recruiting, training, and retaining a small internal team that will still be delivering long after she has moved on; her reward is a flat headcount line and no dramatic story to tell. The other engages a marquee firm, is seen to have “taken decisive action,” and can point to a recognisable name on the steering committee as evidence of seriousness. In most organisations the second director is promoted first. We reward the purchase of reassurance over the construction of capability, and then wonder why capability is never constructed.
The third force is the most cynical and the most common. Every serious contract now contains a knowledge-transfer clause, and almost none of them work, because the incentives on both sides are aligned against them. The firm’s economic model depends on the follow-on engagement; a client that genuinely absorbs the capability is a client that does not come back. The client’s own people, meanwhile, are the very individuals kept too busy running the business to sit alongside the consultants and learn. So transfer becomes theatre: a documentation workstream, a handover deck, a fortnight of “shadowing” scheduled for the end — which is to say, scheduled for exactly the moment the timeline has slipped and the cheapest thing to cut is the part with no external deadline. The knowledge does not transfer because nothing in the machine is built to make it transfer.
The fourth force, atrophy, is what turns a single engagement into a cycle. Capability, once outsourced, does not sit dormant waiting to be reclaimed; it wastes. An organisation that has not designed a system in a decade loses not only the skill to design one but the subtler skill of specifying one — of writing a brief good enough to buy against, of telling a good proposal from a plausible one, of knowing when a firm is guiding it well and when it is being led by the nose. And so it reaches the point that defines true dependency: it needs consultants to tell it which consultants to hire. The muscle that has wasted most is the one that mattered most — the capacity to be a discerning buyer.
The fifth force, memory asymmetry, ensures the imbalance widens over time. Each engagement teaches the firm more about this class of problem and about this client; each engagement, for the client, is a discrete event that ends, disperses, and is half-forgotten by the time the next one begins. The firm is playing a long, continuous game and the client a series of disconnected ones. In any repeated game, the player who remembers beats the player who forgets.
A Programme, Costed
Abstractions conceal; a figure exposes. Consider a composite that will be recognisable to anyone who has lived through a mid-sized enterprise programme in recent years.
An organisation embarks on a two-year core-systems replacement. The delivery is staffed at roughly forty external people through the peak — a blend of onshore consultants and an offshore build pod — at a blended day rate of about eight hundred pounds. Over eighteen active months that is on the order of eleven to twelve million pounds in fees alone, before software and internal costs. Inside the statement of work sits a knowledge-transfer workstream valued at a little over half a million — call it five per cent of the fee. On paper, capability is being bought.
Now watch what happens under pressure. The programme hits the usual mid-course turbulence: a data-migration problem that was underestimated, a regulatory reporting change that must be absorbed, a testing phase that runs long. The timeline compresses. Something has to give, and the something is always the same. The client’s own analysts, who were meant to be seconded to the programme to learn the new system, are pulled back to “keep the lights on” because the day job cannot stop. The transfer workstream, having no external go-live date of its own, is deferred, then trimmed, then reduced to a documentation exercise completed by the consultants themselves — which means the artefacts describing the system are written by the people leaving, for an audience that will not be in the room to ask questions.
The programme goes live, more or less on time, and is declared a success. Twelve months later the organisation has spent a further two million pounds with the same firm on “post-implementation support,” “optimisation,” and “phase two scoping.” None of this was waste in the narrow sense — real work was done — but almost all of it was avoidable in the structural sense. The five per cent that was cut under pressure was the five per cent that would have prevented the twenty per cent spent afterwards. The cheapest line in the budget was the one that governed the cost of every budget that followed.
“The five per cent that is cut under pressure is almost always the five per cent that would have prevented the twenty per cent spent afterwards.”
Dependency Is Not Purchase
It is tempting, having laid this out, to conclude that consultants are the problem and their reduction the solution. That would be to repeat the error in reverse. The honourable reasons to buy — surge, scarce specialism, independence, cross-industry memory — do not disappear because dependency is real. The task is to separate two things that ordinarily travel together and to hold them apart.
The distinction is this. A legitimate purchase is a discrete exchange: you have a defined need, you buy the expertise to meet it, and at the end you are more capable — or at least no less capable — of meeting the next such need yourself. Dependency is the opposite: each purchase leaves you a little less able to specify, challenge, or do the work, so that the need does not close but deepens. The first is buying a meal; the second is buying a habit. The same firm, the same day rate, and the same deliverable can be either, depending entirely on whether the engagement is structured to feed the client’s judgement or to starve it.
Dependency is not the use of consultants; it is the slow surrender of the capacity to be a discerning buyer of them. An organisation that retains that capacity can use the largest firms heavily and remain entirely free, because it can always specify precisely what it wants, judge honestly whether it is getting it, and walk away credibly if it is not. An organisation that has lost it can use consultants lightly and still be wholly captured, because it no longer knows what to ask for or how to tell whether the answer is any good. The measure of freedom is not how little help you buy. It is whether you could stop.
Buying So That You Need Less
If dependency is manufactured, it can be manufactured differently. Four structural changes, none of them exotic, tilt an engagement from producing reliance to producing capability. They are demanding precisely because each runs against one of the forces described above.
- Buy the transfer, not the hope of it. Make capability transfer a contracted deliverable with its own acceptance criteria, its own milestones tied to payment, and its own place on the critical path — not a workstream appended to the end where it can be quietly cut. If a named internal owner cannot independently operate, explain, and extend the thing after the firm leaves, the deliverable is not complete and the final tranche of the fee is not due. This puts real money behind transfer and so defeats the theatre.
- Stay the general contractor. In too many programmes the integrator is the general contractor and the client a subcontractor to its own transformation, holding the risk but not the reins. Reverse it. The client owns the architecture, the sequencing, and the design authority; firms are engaged into that structure to deliver defined components against the client’s decisions. This is harder and slower, and it requires keeping enough senior internal capability to actually hold the centre — which is exactly the point.
- Ring-fence the “why” as an owned asset. The deliverable that matters least on the day and most over time is the reasoning: the design decisions, the rejected alternatives, the trade-offs and the assumptions behind them. Insist that this be captured as it is made — a living decision log owned by the client — rather than reconstructed at the end by the people departing. A system whose design intent the organisation owns can be maintained, argued with, and evolved. One whose intent left with the builders can only be operated on faith, and faith is expensive.
- Measure the firm on your independence at exit. The final question at the close of an engagement should not be only did they deliver? but are we more able to do without them than when they arrived? A firm rewarded, in reputation and in repeat business, for leaving the client stronger behaves very differently from one rewarded for leaving the client hooked. This asks the client to value the right thing, which returns us to the deepest of the forces — the incentive to buy reassurance rather than capability.
Notice that each remedy demands the organisation keep something in-house: an owner able to receive the transfer, a centre able to hold the contract, a discipline able to capture the reasoning, a judgement able to assess independence. None of this is free, and that is the honest catch. Escaping dependency is not cheaper in the short run. It is only cheaper across the arc of a decade.
The Spine You Keep
The mature posture, then, is neither the naive one that buys help and hopes the capability sticks, nor the proud one that insists on doing everything itself. It is to keep a thin, expensive spine of internal capability — deliberately small, deliberately senior — whose job is not to do all the work but to specify it, challenge it, own the reasoning behind it, and remain able to walk away. Around that spine an organisation can buy as heavily as it likes and stay free, because it never loses the capacity to be a discerning buyer. Let the spine wither and no volume of governance will save it; the reliance will deepen whatever the contracts say.
The uncomfortable conclusion is that the consulting dependency cycle is not something done to organisations by rapacious firms. Firms respond rationally to how they are bought; a client that pays for capture will be captured, and a client that pays for capability can be served capably by the very same people. Dependency is a choice an organisation makes and then mistakes for a fact about the world. The work of escaping it is not to buy less help. It is to become, and remain, the kind of buyer who could.