The expectations Covid set, and who still meets them

The Coronavirus Job Retention Scheme was announced on 20 March 2020. Employers could claim from 20 April. HMRC’s IT staff told the National Audit Office they completed the work in four weeks, against the 18 to 21 months they said major IT projects normally take. Once live, 99.5% of claims were processed within three working days.

That is the number every subsequent conversation about public service expectations runs into. Not because it should be the standard, but because it happened, it was audited, and several million people experienced it.

What the four weeks actually cost

The audit is careful about the conditions, and the conditions matter more than the headline.

HMRC used an existing supplier contract rather than running a procurement. It had a single, clearly bounded purpose with a calculation the Treasury deliberately kept simple. Staff worked long hours, weekends and through Easter. Record-keeping of key documents such as process maps was not always maintained, which the department itself identified as a consequence of the pace.

None of that is available to a permanent service. A benefit that has to interact with three existing systems, satisfy a statutory appeals process and survive a change of minister cannot be built by suspending the constraints, because the constraints are the service. What Covid demonstrated was not that government can always move in four weeks. It demonstrated the size of the gap between government’s technical capacity and its normal operating conditions, which is a different and more useful finding.

The expectation stuck. The estate did not move.

Less than five years later the State of digital government review of January 2025 recorded that 47% of central government services still lack a digital pathway, and 45% of NHS services do too. HMRC takes approximately 100,000 calls each day. DVLA processes around 45,000 letters daily. Defra runs over 500 paper form-based services.

The same review found that 70% of the top 75 government services sit below private sector benchmarks, that 24 of those 75 failed to meet basic accessibility standards in the first quarter of 2024, and that satisfaction with public services fell from 79% to 68%. It also estimated that the average UK adult spends a week and a half a year dealing with government bureaucracy.

Put those beside the furlough scheme and the shape of the problem becomes clear. The gap is not between what government can build and what citizens want. It is between the handful of services that receive emergency treatment and the several hundred that do not.

The counterweight that gets forgotten in the rush

Every argument for moving faster online meets the same fact, and it has not shrunk as much as the sector assumes. The government’s one year on report on its digital inclusion plan, published on 5 August 2026, records 1.6 million people with no internet connection at all, and 1.7 million households without a laptop, tablet or desktop computer.

Those figures are the reason a service that exists only online is not a completed service. They are also why the interesting design question after Covid is not whether to digitise but what the assisted route looks like once the digital route becomes the default. HMRC monitored potential claim rates among the digitally excluded during the pandemic using data on people who file self assessment returns on paper, which is a crude proxy and considerably better than not looking.

The response since has been mostly partnership-shaped rather than structural. A £11.9 million innovation fund supporting 85 projects in England, an IT reuse charter with 43 signatories that moved 22,200 devices between June and December 2025, and industry pledges the department counts as having connected or supported over a million people. That is real and it is small against 1.6 million people offline and a state that keeps closing counters.

What has actually been built since

The GOV.UK app is the flagship of the current programme and the Government Digital Service reports it had been downloaded 316,000 times by the end of December 2025. The One Login app now carries government’s first digital credential, the veteran card, launched on 17 October 2025 for nearly two million former service personnel. GOV.UK One Login had passed three million accounts by October 2024, replacing a landscape that in 2021 contained 44 different ways to prove who you are, with monetised benefits put at around £1.75 billion over five years against an investment of about £305 million.

Those are respectable numbers for platform work and they are not what citizens compare government with. A person renewing a licence compares it to their bank, and the bank did not have to consolidate 44 identity systems first.

The comparison that gets made instead

There is a version of the post-Covid argument that says citizens now expect government to behave like Amazon. It is a lazy comparison and it is worth taking apart, because the useful part of it survives.

A retailer optimises one transaction against one measure and can decline any customer it finds unprofitable. A public body has to serve everyone, cannot decline, must apply a statute that changes annually, and has to explain its decision on request. Those constraints are not inefficiency. They are the service.

What does transfer is narrower and harder to dismiss. People have learned what a well-designed status update looks like. They know a system can tell them where their thing is, and that being told nothing is a choice rather than a technical limit. Most complaints about public services that get characterised as impatience are complaints about the absence of that. Progress information is also the cheapest thing on the list to build, which is why its absence is the most revealing.

The Covid schemes did that part well, as it happens, and the audit notes the reason: HMRC mapped the customer journey comprehensively, tested it internally and with tax agents, and wrote guidance and training before launch. None of that required four weeks of heroics. It required somebody deciding it was in scope.

The honest conclusion

The pandemic did not raise expectations of government so much as prove that a specific expectation was achievable. Everyone who now says a service should be faster is, whether they know it or not, pointing at the furlough scheme.

What the audit record shows is that the four weeks was bought by suspending procurement, documentation, scope negotiation and competing priorities, and that the bill for the first two arrived later. The reasonable ask is not permanent emergency conditions. It is that the machinery around a service should not add eighteen months to work that takes four weeks when nobody is allowed to slow it down.

That is a governance question, examined further in digital is not what holds the public sector back, and it sits alongside the rest of the public sector technology coverage.