Offshoring the Problem — Why Exporting Cost Means Importing Risk
We did not reduce the cost of the work; we relocated it, and paid a second time for the distance.
Executive Summary
For three or four years now the offshore development centre has been the reflexive answer to almost any question about cost. A budget overruns; a board asks why the work is not being done somewhere cheaper. The logic seems unarguable: the same task, a third of the price. This essay examines why that logic, so clean on the page, so often disappoints in the ledger — and why the pattern repeats even among organisations that have watched it disappoint before.
The argument is simple to state. Offshoring, as it is usually justified, does not reduce the cost of work so much as relocate it, and in relocating it imports a new and less visible set of costs — coordination, rework, governance, and the slow erosion of knowledge — that the original business case never counted. We export a number we can see and import risks we cannot, and we book the saving before we have earned it. The forces that sustain the pattern are not stupidity or fashion. They are structural: an accountability separated in time and in person from its consequences, a profit-and-loss account that rewards the legible unit rate over the diffuse system cost, and a persistent confusion between moving activity and moving capability.
None of this is an argument against offshoring. The organisations that have made it work — and some have made it work superbly — are the proof that geography is not the problem. But they succeeded by refusing the very framing that makes the business case easy to sign: they bought capability rather than hours, costed the whole system rather than the rate, and treated the transition as the project rather than the overhead. What separates those who thrive from those who quietly unwind their offshore programmes two years later is almost never where the work is done. It is what they believed they were doing when they sent it there.
The saving that was booked before it was earned
There is a particular slide that has appeared, with minor variations, in a great many board papers of late. On the left, the cost of a developer or an analyst here: a day rate of perhaps five hundred pounds, fully loaded. On the right, the same role in an offshore centre: a hundred and fifty, perhaps a hundred and eighty. Beneath the two columns, a single arresting figure — a saving of some sixty-five per cent — multiplied across a few hundred roles into a number large enough to change the shape of a budget. The slide is not dishonest. Every figure on it is real. And yet within eighteen months a good proportion of the programmes it launched will be delivering nothing like the saving it promised, and a few will be delivering none at all.
The reason is not that the rates were wrong. It is that a rate is not a cost. The business case captures the wage differential — the one part of the equation that is legible, boardable, and available in advance — and quietly omits almost everything that will actually determine the outcome. Consider what has to be added back once the work is genuinely under way. There is the transition itself: six to nine months in which experienced people here stop delivering in order to document, explain, and supervise, while the new team offshore is paid to learn. There is a retained layer of coordination — architects, business analysts, and managers who no longer build anything but who exist to specify, review, and translate — that commonly runs to fifteen or twenty-five per cent of the offshore cost. There is rework, because the first cut of anything built at a distance from the people who understand the problem tends to need a second cut. There is travel, telecommunications, and the standing overhead of managing a contract in another jurisdiction. And there is attrition, of which more below.
Add these back and the sixty-five per cent has a habit of collapsing to twenty or thirty, and in the first year, once transition is counted honestly, to nothing at all. The programme is not cheaper yet; it is merely committed. We did not reduce the cost of the work; we relocated it, and paid a second time for the distance. The saving was booked in the business case and will now have to be earned back, painfully, over the years in which the people who signed the case have moved on to other things.
What actually crosses the water
It helps to be precise about what is being exchanged. What crosses the water outbound is cost — specifically, the portion of cost that is wages, and more specifically still, the visible headline rate. What crosses the water inbound, in return, is risk. That is the trade the headline slide never draws, and it is the whole of the matter.
The risks imported are not exotic. They are the ordinary frictions of doing complex, ambiguous work at distance, through a contract, across a cultural and time-zone gap — but they are real, they are cumulative, and unlike the wage saving they do not appear on any slide:
- Coordination cost. Every clarification that would have been a two-minute exchange at a neighbouring desk becomes a written specification, a query, and a twenty-four-hour round trip across a four- or five-hour time difference. Multiplied across a programme, this is not a nuisance; it is a tax.
- Quality variance. Work built without immersion in the domain arrives correct to the letter of the specification and wrong to its intent. The gap is closed by rework, and rework is paid for twice.
- Knowledge attrition. The understanding of your systems accumulates in people, and offshore those people leave — often a quarter to a third of them each year. You are renting comprehension, not building it.
- Governance and control overhead. Managing a supplier at arm’s length in another jurisdiction demands contracts, service levels, audits, and a retained management layer — and, since Sarbanes-Oxley and under the Data Protection Act, a fresh set of questions about where data physically sits and who may touch it.
- Dependency. Once the capability to do the work has left the building, the option of doing it yourself has left with it. The negotiating leverage shifts, quietly and permanently, to the party that now knows how your systems actually work.
| What the business case counted | What it did not |
|---|---|
| The onshore day rate | The retained coordination layer here |
| The offshore day rate | Transition and knowledge-transfer time |
| The headline percentage saving | Rework from specification-at-a-distance |
| Contract value | Attrition and the re-learning it forces |
| Travel, connectivity, and audit | |
| The lost ability to do the work yourself |
The business case is a study in what economists would call selection on observability. We counted what was easy to count and called the total the cost. The uncounted items were not small; they were merely inconvenient to a decision that had, in most cases, already been made.
The coordination tax
Of the imported risks, coordination is the one most consistently underestimated, because on paper it looks like a communication problem and communication problems feel soluble. They are not soluble; they are structural. Distance imposes a fixed levy on every unit of shared understanding, and no amount of goodwill repeals it.
I have watched a programme in which a change that any two colleagues would have settled in a corridor took nine days: a day to notice the ambiguity, a day for the question to reach the offshore team through the retained analyst, a day for it to be misunderstood, a day for the clarification, and the balance lost to the time difference and a public holiday that fell in one country and not the other. Nothing in that sequence was anyone’s fault. It was simply what the corridor had cost, made visible because the corridor was now five thousand miles long.
The organisation’s instinct, when coordination falters, is to add governance — more analysts here, more managers there, more forums, more documentation. This is the coordination tax compounding. Each layer added to bridge the distance is itself a cost that the original slide did not carry, and each makes the retained onshore group larger, slower, and more expensive, until the very team that was meant to be a lean point of control becomes the bottleneck through which all understanding must pass.
The paradox of offshoring for cost is that the cheaper you make the hands, the more you must spend on the head. The wage saving is real; so is the coordination overhead it summons into being, and the second grows with the first.
The knowledge half-life
The most expensive of the imported risks is the one that does its damage slowest. Knowledge of how a particular organisation’s systems, processes, and exceptions actually work is not written down; it lives in people, and it decays when they leave. Offshore, in the labour markets of this decade, they leave at a rate that would be a crisis if it happened here — attrition of twenty-five to thirty-five per cent a year is discussed as ordinary, and wage inflation in the major offshore centres is already beginning to narrow the very gap the whole exercise was built to exploit.
Set a thirty per cent annual attrition against a two-year system and the arithmetic is sobering: the people who learned your platform in the first year are largely gone by the third, and what they knew has gone with them unless it was deliberately captured — which, under cost pressure, it rarely is. The understanding you paid to build offshore has a half-life, and you are on a treadmill of re-teaching, paying again and again for comprehension you thought you had bought once.
Meanwhile the retained organisation here has hollowed. The experienced people who once built the systems were redeployed or released; those who remain specify and review but no longer make, and within a few years they have lost the deep fluency required to know when the vendor is wrong. This is the quiet endgame of the arbitrage: not that the work is done badly, but that no one on your side of the contract retains the standing to say so.
The remedy is not mysterious, but it is expensive and therefore usually skipped. It is to treat transition itself as the project rather than as a regrettable overhead on the way to the saving:
- Map the knowledge honestly before anything moves — what is documented, what lives only in people, and what would be catastrophic to lose.
- Move the work in a sequence that builds understanding, not in one lift sized to hit a run-rate target by year end.
- Retain deliberately here: architecture, domain judgement, and the capacity to specify and to challenge.
- Fund the re-learning that attrition guarantees, and count it in the case, because it is not a risk but a certainty.
An organisation that does these things can offshore well. An organisation that treats them as friction on the road to a headline number is not offshoring its work; it is offshoring its memory.
Why the pattern persists
If the shortfall were merely an error of arithmetic it would have been corrected by now; business cases would carry the loadings and the disappointments would be rare. That they are not rare tells us the cause is not ignorance but structure. Four forces keep the pattern alive, and each will survive any individual’s resolve to do better.
- Accountability is separated from consequence. The executive who signs the case is measured on this year’s cost line; the coordination overhead, the rework, and the knowledge erosion land two and three years out, on a successor. The decision and its reckoning are divided in time and in person, and a cost that arrives after you have gone is, for practical purposes, free.
- The legible number defeats the diffuse one. A wage differential is a single figure, available in advance, defensible in a meeting. Coordination cost, quality variance, and attrition are diffuse, lagged, and hard to attribute to any one decision. We optimise what we can measure, and the offshore case is a monument to the tyranny of the measurable over the material.
- The unit rate seduces. A day rate is one clean number and it invites a comparison — theirs against ours — that feels like analysis but compares the wrong things. Capability is not a scalar. Two people at the same rate are not the same purchase, and a rate card conceals precisely the differences that determine whether the work will actually get done.
- Activity is mistaken for capability. The deepest error of all is to believe that because the work has moved, the ability to do the work has moved with it. Relocating activity is a logistics exercise; transferring capability is a years-long investment in people and understanding. The first can be done in a quarter; the second cannot be rushed, and the case that funds only the first is buying a shell.
“We optimise what we can measure, and the offshore case is a monument to the tyranny of the measurable over the material.”
These forces are why intelligent organisations, staffed by people who have seen offshoring disappoint before, offshore again on the same terms. It is not that they have forgotten. It is that the incentives that produced the first decision are still in place to produce the second.
The strongest case for offshoring — and what it gets right
It would be easy, and wrong, to end there. The honest objection to everything above is that offshoring plainly works, and the evidence is not thin. Some organisations have built offshore capability that is now a genuine strategic asset rather than a cost dodge — captive centres that began as back offices and have matured into places where real engineering and real analysis are done, at scale, by people who stay. The offshore services industry itself has climbed the value chain within living professional memory: the vendors that a decade ago supplied contract programmers by the head now arrive with process-maturity accreditations and quality disciplines that many of their clients cannot match. Access to a deep and growing pool of scarce talent, the ability to scale a team in weeks rather than quarters, the capacity to run work around the clock across time zones — these are real advantages, and they are not available at home at any price.
All of this is true, and none of it rescues the arbitrage. Look closely at the organisations that have made offshoring a durable strength and the same facts appear every time. They did not frame it as a cost saving; they framed it as building a global operating model, and cost followed as a consequence rather than leading as the goal. They invested in transition instead of writing it off. They retained real capability at home rather than hollowing it out. They managed attrition as a first-order problem rather than an unpleasant surprise. And crucially, they did not book the saving in advance and spend the years afterward defending a number they had promised before they understood it.
The lesson is therefore not that offshoring fails. It is that offshoring for arbitrage fails, and offshoring for capability can succeed — and that the two are almost indistinguishable on the day the business case is signed, which is exactly why so many organisations set out to do the second and find, three years later, that they did the first.
From arbitrage to operating model
If the arbitrage framing is the disease, the cure is to change what the decision is about before it is taken. That is a harder thing than adjusting a spreadsheet, because it asks the organisation to want something other than a headline saving — but where it is done, the difference is decisive.
- Buy capability, not hours. Ask what enduring ability you are trying to build, not what this year’s rate card can shave. If the honest answer is only “a lower number,” the programme will deliver a lower number and a pile of imported risk, and little else.
- Cost the whole system. Put the coordination layer, the transition, the rework, and the re-learning into the case at the start. A case that survives honest loading is worth backing; one that only survives by omission was never really an argument, only a wish.
- Keep the head, move the hands with care. Decide deliberately what intelligence must remain here — architecture, domain knowledge, the standing to challenge a supplier — and defend it against the temptation to cut it once the visible saving is under pressure.
- Make transition the project. Treat the movement of knowledge as the thing you are actually doing, resourced and sequenced as such, rather than as an overhead to be minimised on the way to a run-rate target.
- Govern for a horizon longer than a tenure. If the person who signs the case will not be present for its reckoning, build the governance that carries the accountability forward — or accept that the decision is being made by someone who will never pay for being wrong.
None of these steps is novel and none is secret. They are skipped not because they are unknown but because each of them lowers the headline saving that made the case attractive in the first place. Which is the whole point: the honest version of offshoring is less exciting on the slide and far more likely to be true in the ledger.
Coda: what the ledger finally shows
Offshoring is not, in the end, a question about geography, and the organisations that keep relitigating it as though it were will keep being surprised. The ledger settles the argument the same way every time. The saving booked in the first year is repaid, with interest, in the second and third — in coordination, in rework, in the knowledge that walked out of an office you never visited. Sometimes the balance still comes out ahead, and where it does you will find an organisation that costed the whole system and built real capability. Sometimes it comes out behind, and there you will find a slide, filed and forgotten, with two columns and a very large percentage between them.
The question worth asking, before any of this is set in motion, is not the one the slide answers. It is not “is it cheaper there?” It is “what, precisely, are we moving — and what are we determined to keep?” An organisation that can answer that has a chance of exporting cost without importing more risk than the saving is worth. An organisation that cannot is not reducing the cost of its work. It is only moving the problem, and paying the freight.