Steve Tessier spent over 40 years at the Ministry of Defence. He did a stint at NATO. Four months into retirement, he has still not received a single penny of his pension. He has had no contact from the government. No contact from Capita. That’s the company awarded a £239 million, seven-year contract to administer the civil service pension scheme for its 1.7 million members. He is living off his savings. He wrote to the Prime Minister.
“There should be no doubt that I am incandescent with the way that I, and tens of thousands of others, are being treated,” he told Keir Starmer.
Tessier is not opposed to outsourcing in principle. But he is clear about what has gone wrong.
“We’ve become reliant on outsourcing companies,” he said. “We’ve lost control and we’ve lost accountability over the services those companies provide.”
At the same moment Tessier was writing that letter, the government was busy announcing its plans for a digital ID system. It’s one of the most sensitive, complex and consequential technology projects in British public life.
And it is about to make the same mistakes.
A False Choice
Speaking to the Home Affairs Committee earlier this month, Darren Jones, chief secretary to the Prime Minister, explained the government wanted to build the digital ID system in-house. His reasoning was sound. As far as it went.
“I would like to think about how we build the technology ourselves, in-house, as opposed to outsourcing it to, say, a big American tech company,” he said.
Jones was right that handing a foundational national identity infrastructure to a US corporation carries risks that go well beyond cost. Data sovereignty, political exposure, security, and trust, the list is long. But the alternative he is proposing, a bespoke government build, carries its own risks. Whitehall is not short of legacy IT disasters. And the government’s own consultation document, published this week, contains a remarkable admission. On page 83 of a 91-page document, the Cabinet Office makes an admission. “At this stage of development, it is not possible to estimate the cost to government of developing and running the digital ID system.”
That is not a reassuring starting point.
A Programme in Search of Leadership
This programme has burned through one minister. Josh Simons, who led the digital ID effort, resigned earlier this month. James Frith has stepped into the role. At the time of writing, Digital Identity does not appear on his official government page.
A programme without stable political leadership, without a cost estimate, and without a settled vision of how government will build it? That sounds like a programme in trouble before developers have written the first line of code.
“I do not want a huge, stand-alone, multibillion-pound, ten-year piece of government IT.”
The consultation closes on May 5. A People’s Panel, 100 to 120 randomly recruited members of the public, will then convene to develop recommendations, reporting by June 21. There is something admirable about involving citizens in decisions that affect them. And buy-in for a scheme like this is important. But this is one of the most technically and legally complex infrastructure decisions the government has ever faced. The stakes, for privacy, for security, for public trust, could not be higher. Randomly selected members of the public, however well-intentioned, are not the people who should be making the final call on federated data architecture.
Elephant in the Room
Jones has described what he wants with precision. A federated system, where different types of data sits securely in different government departments. A system where citizens give permission for that data to be pulled together to deliver a service. Banking-level security. Two-factor authentication. No central database. Built to integrate with existing infrastructure, not replace it wholesale.
What Jones described is what SAP does.
SAP is not an American company. It is German, or European, in the language that matters to a government anxious about data sovereignty. It is the backbone of enterprise data management for the world’s most complex organisations. It handles identity, permissions, federated data, security and legacy integration at scale, across many jurisdictions, every day. It underpins significant amounts of government back-office infrastructure globally. It is a known quantity with a known cost model.
And yet it does not appear to be part of the conversation.
Why Nobody Is Saying It
The resistance to off-the-shelf solutions in government technology is well documented. The argument runs that government is too complex, too unique, too politically sensitive for standard software. Bespoke builds are commissioned. Customisations multiply. Costs escalate. Deadlines slip. Trouble-shooters are appointed. Apologies are issued.
The big systems integrators, Capita, Infosys, TCS and their peers, do not thrive on simplicity. They thrive on complexity. Long contracts. Custom architectures. Solutions that create dependency and guarantee renewal. Recommending standard SAP, implemented without customisation, would collapse that model. So they are not recommending it.
This is not a conspiracy. It is an incentive structure. And the government, lacking the in-house expertise to challenge it, keeps walking into the same room and signing the same contract.
Meanwhile, the NHS has reportedly refused to share data with the digital ID programme. And the Department for Education has resisted integrating its own records. These are not technology problems. They are trust and governance problems. The kind that a federated architecture with genuine departmental control is designed to solve. The kind SAP handles as a matter of course.
What Good Looks Like
Jones wants parliamentary oversight. He wants accountability. He wants a system that does not become, in his words, “a huge, stand-alone, multibillion-pound, ten-year and huge bill-driven piece of government IT.”
None of that is incompatible with SAP. It is more achievable with proven, standard enterprise software than with a bespoke government build that no one outside the original contractor can maintain, extend or audit.
The lesson of Capita, and of the Post Office, and of countless other public sector IT failures, is not that outsourcing is wrong. It is that outsourcing to the wrong partner, for the wrong reasons, without genuine accountability, is catastrophic.
Steve Tessier understood that better than most.
“You ought to be doing a proper due diligence process,” he told the Prime Minister, “where you make a judgment as to the ability of a potential contractor to do the work to the total satisfactory standard.”
You do not need to build this from scratch, and you do not need to hand it to an American tech giant. There is a third option. It exists. It works. It is sitting right there.
