Why departments keep merging digital into technology

In summer 2020 the Department for Education put its digital and technology functions under one leader. The department’s own account of it, published on 21 April 2021 by its chief digital and technology officer Emma Stace, is brisk about the fact and more revealing about the reason. Legacy technology, processes and services are named as one of the department’s greatest challenges, and described as a side effect of how it funds, buys and implements systems.

That sentence is the whole argument for the merger, and it explains why the same reorganisation keeps happening in other departments.

Two functions, two budgets, one gap between them

Split the work and you get a predictable division of labour. The digital function builds new services, usually against a policy commitment with a date attached and money that expires. The technology function runs the estate, funded from an operating budget that is reviewed annually and squeezed whenever something else goes wrong.

Neither owns the transition. The new service goes live, the old system it was supposed to replace stays up because three other things still depend on it, and the responsibility for retiring it belongs to whichever function has the least leverage in that year’s budget round. Repeat over a decade and you have an estate where new things have been added continuously and almost nothing has been removed.

The DfE post is unusually direct about what it intended to do instead: reduce the highest priority legacy risks, stop services where the risk outweighs their value, treat thorough maintenance of software as valuable work rather than overhead, propose new funding models, and set principles that prevent tomorrow’s legacy. Stopping a service and rewarding maintenance are both decisions that a build-oriented function cannot take and a run-oriented function is never asked to propose.

The scale of what the merger was aimed at

The national picture arrived four years later. Published on 21 January 2025, the State of digital government review put the legacy share of departmental systems at 28% in 2024, against 26% in 2023. The number of legacy systems increased by 26% from 2023, and 22% of those were red-rated. Keeping such a system alive, the review notes, often runs to three or four times the cost of a modern equivalent.

Read the rise carefully. A 26% increase in the number of counted legacy systems in one year is partly better counting and partly genuine accumulation, and the review does not fully separate them. Either way, an estate on which the legacy share went up while a national modernisation programme was under way is an estate where things enter the category faster than they leave it.

The other half of the DfE diagnosis, about how systems are funded and bought, shows up nationally too. Roughly a fifth of digital headcount in the public sector is contingent labour, and that fifth costs three times what a permanent equivalent does. A function that acquires its knowledge on fixed-term contracts is structurally unable to keep custody of a system for fifteen years, which is how long systems actually live.

The same move, one level up

Whitehall made the equivalent reorganisation for itself in January 2025, unifying into one digital centre inside DSIT the office that led the digital and data function, the service that builds and runs the government’s shared products, the unit piloting artificial intelligence across the public sector, and parts of two others. The blueprint published on 21 January 2025 promised to raise the status of the Government Chief Digital Officer role to second permanent secretary level.

That last part did not happen. The role was replaced with two director general posts, one for the products GDS builds and runs and one for digital transformation across departments, both reporting to the DSIT permanent secretary, with interim appointments made to both in January 2026. The Public Accounts Committee, reporting on 27 March 2026, called that a shortcoming and recommended a chief digital officer at permanent secretary level with the authority to act across government.

It is worth noticing what those two director general titles are. One is products, one is transformation. Having merged digital and technology at departmental level for a decade, the centre has just split them again at its own.

Does merging work

Structure alone does not close a funding gap, and the honest position is that the effect is unproven in either direction. Nobody publishes a comparison of legacy trajectories in departments that merged against those that did not, and the accounting is not consistent enough between departments to make one easy.

What can be said is what a merger makes possible rather than what it guarantees. A single leader can trade a new service against a retirement in the same conversation, which two leaders cannot. That is the entire mechanism, and it only pays if someone actually makes the trade.

The evidence on whether they do is a proxy: seniority. Only four central departments seat a digital leader on their executive committee. Across NHS trusts the equivalent gap is between having a chief digital information officer, which most now do, and that person holding a board seat, which around half do not. A merged function whose head is not in the room when the budget is allocated has consolidated two weak positions into one.

There is a further complication that the DfE post flags and most commentary skips. The department funds, buys and implements systems in a way that produces legacy as a side effect, and merging two directorates does not alter how the money arrives. Capital for a new service and revenue for running the old one still come from different places, on different cycles, with different rules about carrying anything forward. A single leader can now see both sides of that. Whether they can act on it depends on a Treasury settlement none of them wrote.

The more promising part of the DfE strategy is therefore the least structural: the commitment to long-lived multidisciplinary teams and to recruiting civil servants rather than leaning on contractors. Continuity of people is the only mechanism that reliably keeps custody of a system across a decade, and it is the one thing an organisation chart cannot deliver by itself.

One line in the blueprint carries a deadline. Two named posts have to exist in every public body by the end of 2026: someone senior enough for the top table, and a board member holding the same brief from outside the organisation. The appointments have to be made public, which is what makes the commitment markable at all. The roadmap published on 20 January 2026 tracks the blueprint but sets no milestone against this one, so the original wording is the only standard there is to judge it by. Most of the year has gone.

Why now, then, is the wrong question about the DfE. It merged in 2020 for reasons the department wrote down and which the national evidence has since confirmed. The better question is why the centre of government, holding that same evidence, has just moved the other way.

The leadership question is followed further in the job that replaced the transformation CTO, and the wider estate is covered under public sector technology.