Ensure continued business justification: the Emergency Services Network and the milestone that meant nothing

PRINCE2 7 changed the first principle by one word. “Continued business justification” became “ensure continued business justification.”

That word does real work. Justification is not something a project has; it is something somebody must actively confirm, at every stage boundary, for as long as the project runs. The moment nobody is obliged to re-test it, a business case becomes a historical document — an account of what people believed at initiation, filed and never opened again.

The clearest case I know of runs across four National Audit Office reports, a Public Accounts Committee inquiry, an Accounting Officer Assessment and a business case published in July 2026, and it is still going.

What ESN was for

The Emergency Services Network was to replace Airwave, the TETRA radio network used by police, fire and ambulance services across Great Britain, with mission-critical voice and data carried over a commercial 4G network. The commercial model was a disaggregated prime arrangement: Motorola Solutions for User Services, EE for the network, and the Home Office itself as integrator, on fixed-price contracts that allocated delivery risk to suppliers.

The 2015 development cost estimate was £2.565 billion, covering 2015 to 2023. Airwave was to be switched off in January 2020.

The dates

Roughly a ten-year slip against the original plan, and the current target is three years away as I write.

The money, carefully

The cost figures need handling honestly, because they use different baselines and it is very easy to produce a misleading “overrun” by subtracting one from another.

• £2.565 billion — the 2015 estimate of ESN development cost, 2015 to 2023.
• ~£2.0 billion — spent by March 2023, at which point the NAO recorded the programme as “a long way from having a functioning network.”
• £11 billion — a 2021 estimate of lifetime cost, covering ESN operation plus continued Airwave maintenance plus commercial mobile.
• At least £250 million a year — the cost of continuing to run Airwave in the meantime.
• £19.1 billion — the whole-life budget to 2043-44 in the summary business case published on 27 July 2026, with £6.1 billion spent to date.

Those are not four versions of the same number. The first is a build cost; the third and fifth include the legacy network the programme exists to retire. Anyone quoting a single “ESN overrun” figure is almost certainly comparing incompatible baselines, and I would rather say that plainly than produce a dramatic number.

There is also a cost that does not appear in the programme’s own accounts at all. The Public Accounts Committee recorded that police forces had spent £125 million on Airwave devices since 2018 plus £5 million on transition teams; fire services £6 million on transition preparation plus £2 million on superseded early versions; and ambulance services £9.5 million through the Ambulance Radio Programme. Money spent by other people, preparing for something that did not arrive.

The finding that should change how you write contracts

Of everything in the official record, one sentence has stayed with me:

“Motorola had been paid for meeting 85%, by value, of its contractual milestones” — while the network remained undelivered.

Eighty-five per cent of the value of the milestones. No working mission-critical network.

That is not fraud and it is not incompetence by the supplier. It is a contract in which the milestones measured activity rather than capability. Each one was presumably delivered exactly as specified. The specification was the problem.

I have seen the same structure in commercial contracts many times, usually expressed as design documents, environments stood up, modules configured, test cycles completed. Every one is verifiable. None of them, individually or together, guarantees that anything works end to end for a user.

The fix is not complicated, and it is unpopular with suppliers for obvious reasons: at least one substantial payment milestone should be conditional on an outcome a user can demonstrate, in the user’s own environment, doing the user’s actual task. Not a demonstration prepared for the milestone. A police officer, on a real shift, using the radio.

If a supplier will not accept a single outcome-based milestone anywhere in a contract of this size, that refusal is itself information about how confident they are.

The other structural problem: the incumbent was the supplier

Motorola Solutions owned Airwave — the network ESN was contracted to replace — while also supplying ESN. The Home Office did not resolve that conflict commercially. In the end the Competition and Markets Authority had to intervene on Airwave pricing.

The Public Accounts Committee’s assessment was that the Department “did not resolve this situation through astute contract management,” relying instead on the CMA to force change.

This is worth generalising. Whenever a supplier’s revenue from the current state exceeds its revenue from the target state, you have created a structural incentive against delivery. It does not require anyone to behave badly for that incentive to shape outcomes — through prioritisation, through resourcing, through how hard anyone pushes on a slipping date.

The mitigation is not suspicion. It is to name the conflict in the risk register at contract stage, with an owner, and to design commercial terms that neutralise it: transition incentives, legacy price caps, or simply not awarding both.

What the official record says about the rest

The NAO found that “the Home Office did not fully test early ESN products as well as it could have, contributing to the delays.” By late 2021, on mission-critical push-to-talk software, “limited confidence existed on both sides in completing software.”

Control-room upgrades proceeded in parallel without alignment: “The Home Office had no way to ensure that this work aligned with ESN plans.”

The Department accepted it had been “too optimistic about both the progress it has made and the challenges ahead.” The NAO recorded that “the Home Office recognises its commercial approach, which has contributed to challenges across the programme, as suboptimal but decided to continue this to avoid further programme delays.”

And on the 2018 attempt to fix it, the NAO’s summary is four words: “A 2018 reset did not work.”

Where continued business justification actually bites

Here is the difficult part, and I want to be fair to the people involved rather than glib.

The Accounting Officer’s September 2024 assessment — the one that deferred Airwave shutdown to December 2029 — concluded that the programme “conforms to the value for money standard” and “continues to demonstrate a good use of public resources.”

That may well be correct. Once £6 billion has been spent and Airwave costs at least £250 million a year to keep running, the forward-looking case for finishing can be genuinely strong even though the original case has long since evaporated. Economically, that is the right way to evaluate a decision: sunk costs are sunk, and what matters is the cost and benefit from here.

But it exposes something that every project board should understand about its own business case. A forward-looking justification can remain positive indefinitely while the original justification is comprehensively false. Each incremental decision is rational. The aggregate outcome is a programme ten years late at several times its original cost, which nobody would have approved had they known.

This is why “ensure continued business justification” has to mean two things, not one:

Is it still worth finishing from here? The forward-looking test. Necessary, and the one most boards do.

And would we start this today, knowing what we now know? The counterfactual test. Almost nobody does this one, and it is the one that produces uncomfortable, useful answers.

If the answer to the second is no, that does not automatically mean stop. It means the programme should be re-scoped, re-baselined and re-approved as the thing it has actually become, by whoever would have had authority to approve that thing at the outset. What you must not do is continue on an authorisation granted for a different project.

The National Audit Office’s report on HS2 put the same idea in one sentence: “had the current estimated costs been known when work began in 2020, this would have resulted in a benefit–cost ratio of between 0.3 and 0.4, which is categorised as ‘poor value for money’.”

Six questions for your next stage boundary

1. Would we approve this today, at today’s cost, for today’s scope? Ask it explicitly. Minute the answer.

2. What proportion of contract value have we paid, and what proportion of usable capability do we have? If those two numbers have diverged, your milestones are measuring the wrong thing, and now is when you find out rather than at the end.

3. Which of our milestones could a user demonstrate? If none, add one to the next contract variation.

4. What are we spending on the legacy thing while we wait? ESN’s Airwave bill is at least £250 million a year. That number belongs in the programme’s reporting, not in a different budget line, because it is a cost of the delay.

5. What are other people spending because of us? Police, fire and ambulance services spent nearly £150 million preparing for a network that did not arrive. Your programme has downstream equivalents — teams building interfaces, functions training staff, partners changing processes. Count it.

6. Does any supplier earn more from the current state than the target state? If yes, name it in the risk register with an owner.

The point of the principle

ESN is not a story about incompetence. The National Audit Office does not present it that way and neither will I. It is a story about a business case that was written once, in 2015, and then had to carry ten years of decisions that were each defensible in isolation.

PRINCE2’s answer is structurally simple. The Executive is personally accountable for the business case remaining valid. It is reviewed at every stage boundary. And the project may be stopped — not as a failure, but as a decision the method explicitly provides for.

The stopping decision is the one almost nobody exercises. In forty years I have seen a great many projects fail and very few projects stopped. Those are not the same thing, and knowing the difference is most of what a project board is for.

What a good milestone looks like

Since the ESN finding turns on milestones that measured activity rather than capability, it is worth being concrete about the difference. Both columns below describe the same piece of work.

The left column is verifiable, auditable and easy to write into a contract. It is also compatible with delivering nothing that works.

The right column is harder to negotiate, harder to schedule and much harder for a supplier to accept, precisely because it transfers real risk. That resistance is the point. A supplier who will accept several capability milestones is telling you something about their confidence; one who will accept none is telling you something else.

You do not need to convert every milestone. Converting one substantial payment milestone per major phase changes the incentive structure of an entire contract, because it means that at least once per phase, somebody has to produce something that works in front of a person who did not build it.

On stopping

The hardest thing a project board does is stop a project, and almost none of them ever do.

I have watched the reasons up close for four decades. The person who stops a project is visibly associated with its failure, while everyone who lets it continue is merely present. The sunk cost is emotionally real even when it is economically irrelevant. Careers are attached to programme names. And there is always a plausible next step that costs less than the decision to stop.

PRINCE2 provides for premature closure explicitly, as an authorised outcome of the Directing a Project process rather than as a failure. That framing matters more than it sounds. If stopping is a recognised outcome that the method anticipates, the person proposing it is following the method rather than breaking ranks.

Two mechanics make it more likely to happen in practice. First, ask the counterfactual question — would we start this today? — as a standing agenda item at every stage boundary, so that asking it is routine rather than an accusation. Second, when a programme is re-scoped or re-baselined significantly, require re-approval by whoever would have had authority to approve the new thing at the outset. A programme that has quadrupled in cost is no longer authorised by the person who approved it, and saying so is a governance statement rather than a criticism.

If your organisation has stopped no projects in three years, it is not because your project selection is excellent.

Sources

• National Audit Office, Progress delivering the Emergency Services Network (2023) and earlier ESN reports
• House of Commons Committee of Public Accounts, The Emergency Services Network
• Home Office, Accounting Officer Assessment on the Emergency Services Network, 25 September 2024
• GOV.UK, Emergency Services Network summary business case, 27 July 2026 (whole-life budget and spend to date)
• Competition and Markets Authority, Airwave market investigation
• National Audit Office, HS2 (June 2026) — benefit–cost ratio finding, cited for comparison
• Purple Griffon, What’s new in PRINCE2 7 (principle wording) — purplegriffon.com