Cyber Resiliency Board Briefing 15: The Degraded Operations Problem - Why the Business Must Keep Trading Without the Systems It Depends On

Executive Summary

Fourteen briefings have examined how an organisation recovers its systems to a trusted state. None has addressed what the business does while that work is going on. Recovery at enterprise scale is measured in weeks, and occasionally in months, and the company has to keep trading throughout. Most continuity plans specify what comes back first and say almost nothing about how the organisation operates in the meantime. Degraded operation is a capability in its own right, it decays quietly when nobody maintains it, and the Board is usually shown a recovery plan without ever being shown the trading plan that has to sit alongside it.

Weeks, Not Days

The public record has become hard to argue with. Jaguar Land Rover halted production on 1 September 2025 and did not return to normal output until the middle of November, with press estimates of around £50 million a week and £196 million of attack-related costs booked in the quarter to September. Marks and Spencer suspended online ordering for roughly six weeks in 2025 and told the market that disruption would run into July at an estimated cost of some £300 million.

The British Library is the most instructive case, because it published its own review. Services were "severely restricted in the first two months", onsite access ran at around half of normal volume, and the restoration programme was scoped at eighteen months with a six-month phase devoted simply to adapting.

Boards tend to hold a recovery time objective in their heads as the answer to how long this lasts. The honest answer for a destructive attack on a large estate is longer than any objective on the page.

The Fallback That Was Decommissioned

Thirty years of efficiency work has removed the manual alternative from most processes. Paper order books went, then the people who remembered using them, then the suppliers willing to accept an order by telephone. Warehouse staff who once worked from printed pick lists now follow a handheld terminal that cannot function without the network. Finance teams can no longer raise a payment outside the workflow that approves it.

The British Library found that many staff "have been unable to perform significant parts of their roles or, where they have, have been required to follow more onerous manual processes in order to do so", and recorded the effect on morale. That is the honest texture of degraded running. It is slower, it is demoralising, and it depends on knowledge the organisation has usually allowed to lapse.

The test for management is a simple one: has anybody currently employed actually performed the manual process? A procedure nobody has run is a hypothesis.

Borrowed From Aviation

Commercial aviation settled this problem decades ago with the Minimum Equipment List. An aircraft may be dispatched with specified equipment inoperative, because the regulator and the manufacturer agreed in advance which items the aircraft can fly without, under what conditions, and for how long. The deferral periods are explicit. Category B allows three days, Category C ten days, Category D one hundred and twenty.

Nobody improvises this at the gate. The judgement was made in daylight, approved by an authority, and written down with conditions attached. The crew's job is to check the list, apply the conditions and record the deferral.

Few organisations hold the equivalent document. Briefing 5 made the case for deciding recovery priorities before the crisis through a Minimum Viable Company. The mirror of that decision is the one almost nobody takes: what the company can trade without, for how long, and on what terms.

Degraded Running Creates Its Own Risk

Manual workarounds break controls that were designed assuming the system would be there. Payments get authorised by email because the approval workflow is down. Segregation of duties collapses onto whoever is on site. Customer records are captured on spreadsheets and personal devices, and staff coordinate over consumer messaging apps because the intranet has gone, as happened at the British Library.

Each of those is a rational decision under pressure and a reportable control failure afterwards. The organisations that come through this well have decided in advance which controls may be suspended, by whom, with what compensating measure, and logged each suspension as it happens. The ones that do not find themselves explaining to an auditor why six weeks of payments have no approval trail.

The Reconciliation Debt

Everything transacted by hand has to be brought back into the systems of record. Orders taken on paper, payments made outside the workflow, stock movements tracked on a spreadsheet, customer commitments agreed verbally. This is a second project, it lands after the recovery is declared complete, and it is routinely absent from the plan and the budget.

It is also where financial reporting risk concentrates. Six weeks of manual trading produces a reconciliation exercise of real size, under audit scrutiny, performed by the same exhausted people who have just finished the recovery. Briefing 9 made the case that recovery capacity is bounded by people. The reconciliation is the bill for degraded running, and it falls due when there is least appetite left to pay it.

Where the Limit Sits

Degraded operation has an expiry date, and it is rarely set by technology. Contractual service levels start breaching. Covenants that depend on reporting become difficult to evidence. Regulated activities have licence conditions that assume working systems. Customers who tolerated two weeks begin placing orders elsewhere in week five. JLR's suppliers were laying off staff while its own production was paused, which is a reminder that the limit may be reached by someone else's balance sheet before it is reached by yours.

The Board should know what that date is, as a date, and what happens on the day after it.

What Good Looks Like

Cyber resilient organisations treat trading through an outage as a planned capability rather than an act of improvisation. They hold an explicit statement of what the business can operate without and for how long, built with the people who run those processes rather than written for them, and they review it when processes change. For each critical business service they know the manual alternative, who can perform it, how much throughput it sustains compared with normal, and the point at which it stops being viable.

They keep the manual fallback alive. The procedures are current, the forms exist in printable form, a small number of people have actually run them inside the last year, and the dependencies that would defeat them, such as a supplier who only accepts electronic orders, have been identified and addressed in the contract.

They decide in advance which controls may be relaxed during degraded running, who holds the authority to relax each one, what compensating measure applies, and how each suspension is recorded while it is in force. They plan and fund the reconciliation as a named workstream with its own resource, rather than discovering it afterwards. They know which contractual, covenant and licence conditions begin to bite as an outage lengthens, and they have established the date on which degraded trading becomes a solvency or regulatory question rather than an operational one.

Most importantly, they rehearse the business running without its systems, with the people who would have to do it, rather than rehearsing only the restoration of the systems themselves.

Questions to Ask Your Executive Team

  • For our three most important business services, what is the manual alternative, and has anyone currently employed performed it in the last twelve months?

  • How much of normal volume can we sustain manually, and for how long before the business becomes unviable rather than inconvenienced?

  • Which controls would we suspend during degraded running, who authorises that, and how is each suspension recorded?

  • Who owns the reconciliation of manually transacted business back into the systems of record, and is it resourced in the plan?

  • On what date in an extended outage do contractual, covenant or licence conditions begin to breach, and what happens the day after?

  • Which of our suppliers and customers would fail before we do, and have we discussed it with them?

Key Takeaways for the Board

  • Recovery at scale takes weeks. The business has to trade through all of them, and that trading has usually never been planned.

  • Decades of efficiency work removed the manual alternative. What remains is often documented but unpractised.

  • Manual workarounds break controls by design. Unplanned, they become findings; planned, they become compensating measures.

  • The reconciliation of manually transacted business is a second project, and it is almost always unbudgeted.

  • The limit on degraded trading is set by contracts, covenants and customers rather than by technology.

The Board's Role

The Board does not need to know how an order is picked without a handheld terminal or which form replaces a workflow screen. It should, however, expect management to be able to describe how the company trades while recovery is under way. Management should be able to set out what the business can operate without and for how long, which manual alternatives exist and who is competent to run them, what throughput those alternatives sustain, which controls would be relaxed and under whose authority, how the resulting reconciliation is owned and funded, and the point at which lengthening disruption turns into a contractual or regulatory problem.

The Board should also challenge continuity metrics that stop at restoration targets. A recovery time objective describes an ambition for the technology. It says nothing about whether the company can take an order, pay a supplier, or answer a customer on day twenty, which is the question the market, the regulator and the workforce will actually be asking. A more useful view comes from management demonstrating which business services have a tested manual mode, what capacity that mode delivers, and how long it can be sustained.

The fundamental Board-level question is this: "if our systems were unavailable for six weeks, could this company still trade, lawfully and solvently, and can management show us how?"

Closing Thought

Every briefing in this series has treated recovery as the restoration of trust in systems. This one is about the weeks in between, when the organisation has to be a business without them. Survivability is decided in that gap. A company that recovers beautifully in week seven, having lost its customers in week three, has solved the wrong problem.

Next Briefing

Cyber Resiliency Board Briefing 17: The Funding Problem — why recovery is a cashflow event before it is a technical one, and why the insurance policy rarely pays at the speed that recovery demands.

Next
Next

What the Microsoft Outage Teaches Us About Communicating Through a Crisis