The British press is having another collective meltdown over a tech executive recusing themselves from a government contract. This time, it is the NHS England Federation platform director stepping back from the £330 million Federated Data Platform (FDP) decision due to historic ties to Palantir. The predictable chorus of tech-sceptics, privacy fundamentalists, and armchair ethicists are celebrating this as a win for "accountability."
They are entirely wrong.
What the mainstream narrative frames as a triumph of public sector integrity is actually a symptom of a deep, systemic disease within government procurement. We have created a system that actively punishes competence and rewards ignorance. By demanding absolute insulation from the industries they regulate, public sector bodies ensure that the people making multi-million-pound technology decisions are the least qualified to do so.
The Compounding Cost of the "Pure" Bureaucrat
The lazy consensus states that any overlap between private sector employment and public sector decision-making is an inherent corrupting force. To avoid even the "appearance of bias," we demand that officials recuse themselves at the first sign of domain expertise.
Let us look at the mechanics of what actually happens when a top official steps down from a major procurement project like the NHS Palantir contract.
- The Expertise Vacuum: The decision-making power shifts to career civil servants who have never deployed a enterprise data architecture in their lives. They understand procurement policy, not data pipelines.
- The Vendor Advantage: When the buyer does not understand the technology, the vendor wins. Sophisticated tech firms can easily out-negotiate a committee of generalist bureaucrats who cannot tell the difference between a legacy relational database and a modern semantic data layer.
- The Delay Tax: Recusals trigger internal reviews, audits, and handovers. In healthcare, a six-month delay in data integration does not just cost money; it costs lives through inefficient resource allocation, missed diagnostic patterns, and backlogged elective surgeries.
I have spent two decades watching public sector organizations blow tens of millions on failed IT transformations. The root cause is almost never "corruption." It is sheer, unadulterated incompetence born out of a terror of conflict of interest. We have optimized for a clean paper trail instead of a working system.
Dismantling the Premise: The Misconceptions Driving the Outcry
The public discourse surrounding the NHS Federated Data Platform is built on a foundation of fundamental misunderstandings about software, data ownership, and procurement economics.
"Palantir is Buying Access to NHS Patient Data"
This is the most pervasive myth in the British media. Software vendors providing data integration platforms are data processors, not data controllers. Palantir’s Foundry platform is an operating system; it does not own the data fed into it any more than Microsoft owns the text written in Word. The NHS retains total sovereignty over its data governance. Conflating infrastructure providers with data brokers is a failure of basic technical literacy.
"We Should Build This In-House to Save Money"
The argument for bespoke, state-built software sounds noble but is fiscally illiterate. The public sector cannot compete with Silicon Valley for top-tier engineering talent. When the state tries to build proprietary enterprise software, you get disasters like the original NHS National Programme for IT (NPfIT)—a failure that cost UK taxpayers over £10 billion. Buying commercial off-the-shelf (COTS) software and configuring it is the only viable path to modernization.
"Total Neutrality Guarantees Total Fairness"
Neutrality is often just a euphemism for lack of perspective. A procurement board entirely scrubbed of anyone who has ever worked for, with, or against a major tech vendor is a board of amateurs. They are blind to the subtle traps hidden in software licensing agreements, vendor lock-in strategies, and implementation roadmaps.
How to Actually Fix Government Tech Procurement
Stop trying to fix the conflict-of-interest guidelines by making them stricter. That is a dead end that will only further deplete the talent pool. Instead, we need to completely redefine how the public sector interacts with private expertise.
1. Shift from Recusal to Radical Disclosure
Instead of forcing experts out of the room when a decision matters, keep them in the room but mandate absolute transparency. Record every meeting. Log every rationale. If a director favors a specific vendor, force them to defend that choice using hard technical and financial metrics against a panel of independent peer reviewers. We need accountability of outcome, not accountability of association.
2. Implement "Skin in the Game" Contracts
The real issue isn't where an official used to work; it's where the incentives lie. Government tech contracts should be tied to performance milestones rather than upfront block payments. If the Federated Data Platform fails to reduce patient waiting lists or optimize bed capacity by a verifiable percentage within 18 months, the vendor shouldn't get paid full freight.
3. Build a Dedicated Technical Commando Unit
The civil service needs a highly compensated, elite tier of technology buyers who rotate between the private sector and government. Treat them like military reservists. Accept that they will have industry ties, and design a governance framework that utilizes that network rather than criminalizing it.
The Trade-offs of the Pragmatic Approach
To be fair, abandoning the illusion of absolute neutrality has its downsides. It requires the public to accept a level of nuance that does not fit neatly into a newspaper headline.
| The Status Quo (Bureaucratic Purity) | The Pragmatic Model (Domain Expertise) |
|---|---|
| Low risk of scandal, exceptionally high risk of project failure. | Higher risk of political optics, vastly higher probability of technical success. |
| Decisions made by generalists relying on external consultants. | Decisions made by insiders who understand vendor tactics. |
| Multi-year delays caused by procedural paralysis. | Accelerated deployment cycles driven by execution capability. |
If you choose bureaucratic purity, you are choosing slow, broken, and expensive software. You are choosing a system where the process is defended while the patient suffers.
The NHS handles over a million patients every 24 hours. Its data architecture is a fragmented, archaic mess of legacy systems that cannot talk to one another. Fixing that requires the sharpest technical minds available—people who have actually built, scaled, and managed enterprise platforms.
If those people happen to have worked for the dominant players in the industry, that is not a disqualifying scandal. It is a resume.
Demanding that the NHS modernize its infrastructure while banning anyone who understands how to do it from the decision-making table is a form of institutional self-sabotage. It is time to stop playing politics with data procurement and start prioritizing competence over optics.
