Put AI to work.

Get practical ideas, expert guidance and workshop updates in your inbox.

You can unsubscribe at any time. Read our Privacy Policy.

Business insights

What Business Stress Testing Reveals About Resilience and Cyber Security

Most organisations respond to resilience worries by commissioning another risk register, refreshing the incident response policy, or asking the IT team to confirm that the backups ran. All of that…

VC

Laurens de Rooij

Velox Consilium

17 November 2026

Reading time

What Business Stress Testing Reveals About Resilience and Cyber Security

Most organisations respond to resilience worries by commissioning another risk register, refreshing the incident response policy, or asking the IT team to confirm that the backups ran. All of that is reasonable. None of it tells you what happens on a Tuesday morning when the systems are down, the phones are ringing, and nobody is sure who is allowed to decide anything.

Business stress testing is the practice of finding that out before Tuesday arrives. You take a realistic disruption, such as a cyber incident, a supplier outage, a technology failure, or the sudden loss of two key people, and walk the people who would have to respond through it. You watch what they do, what they assume, and where they get stuck. It is a rehearsal, not an audit.

A plan on paper is a promise. A stress test shows you whether the organisation can keep it.

Disruption doesn’t stay in its lane

Risk is usually assessed in departments. Cyber risk sits with technology, supplier risk with procurement, continuity with operations, and regulatory exposure with compliance or legal. Each assessment may be perfectly competent. Each is a torch showing one patch of a dark room.

Real disruption wanders across the whole room. Take a cyber incident at a technology supplier that your customer service team relies on. Your own systems might be untouched, yet within hours the supplier is struggling, your agents can’t see customer records, complaints are building, and someone senior is asking whether this triggers a duty to notify a regulator. Meanwhile the people who would normally decide are working from partial information, and at least two of them are in different buildings.

That isn’t a technology problem, a supplier problem, or a communications problem. It is all three at once, and none of the separate assessments modelled it.

The timing pressure is real, too. The Cyber Security and Resilience Bill, which has been working its way through Parliament, proposes tightening incident reporting to an initial notification within 24 hours and a full report within 72 hours, and introduces oversight of critical suppliers. Even if your organisation is not directly in scope, a larger customer that is may start asking you how quickly you could tell them something had gone wrong. If nobody has practised that conversation, 24 hours is not very long.

The government’s Cyber Security Breaches Survey 2025/2026 found that 43% of UK businesses, roughly 612,000 organisations, identified a breach or attack in the previous 12 months. Only 25% had a formal incident response plan. I wouldn’t read that as a reason to panic. I would read it as a reason to stop assuming that having a plan is the unusual part. The unusual part is having a plan that someone has actually tried to use.

There is a simple way plans fail, and it resembles keeping the spare key in the locked car. The incident plan is stored on the file share that is encrypted. The call list lives in the email system that is down. The recovery procedure assumes the administrator is available, and the administrator is on a flight to Lisbon. The plan is complete. It is just unreachable at the moment it is needed.

A plan that depends on the thing that has failed is not a plan. It is a document.

What stress testing is not

There are two common misreadings, and they both make the exercise less useful.

The first is that stress testing is prediction. It isn’t. Nobody in the room is claiming this exact event will happen on this exact date. If the scenario is “a key supplier is unavailable for two days”, the point is not that you’ve forecast the outage. The point is that the same weaknesses would show up whether the cause were ransomware, a failed software update, or a flooded data centre. The scenario is a way in. The findings are general.

The second is that stress testing is a control check. A control test asks whether something worked. Did the backup complete? Is multi factor authentication switched on? Does the firewall rule exist? Those are worthwhile questions, and a good security review answers them. A stress test asks a different set of questions about people and decisions.

Checking a control confirms that the backup ran. Stress testing asks whether anyone can find it, who decides to restore from it, and what customers hear while that decision is being made.

Checking a policy confirms that an escalation path exists. Stress testing asks whether the person at the top of it would answer a call at 2am, and whether they’d know what they were being asked to approve.

What you are really doing is surfacing assumptions, and organisations carry a surprising number of them without ever saying them out loud. Five come up repeatedly:

  • Which services are genuinely critical, as opposed to merely important or politically loud
  • How long each one could be unavailable before the damage becomes serious, in customers lost, contracts breached, or cash not collected
  • Who has the authority to make decisions, and whether that authority survives their absence
  • What information leaders would need to decide, and where it would come from if the usual dashboards were dark
  • Where recovery depends on another organisation, and whether that organisation knows

That last one tends to bruise. Many organisations list their suppliers carefully and assess them annually on a questionnaire. Far fewer have ever asked what they would actually do if the supplier were silent for two days, and whether the supplier’s incident process includes telling them promptly.

A practical starting point: the two day supplier outage

I would resist the urge to build a grand programme of scenarios. One realistic scenario, run properly, teaches more than ten run superficially.

Here is a good one. A technology supplier that supports a customer facing service goes offline following a cyber incident. They tell you they expect to be back in two days, though they don’t sound certain. Nothing in your own estate has been compromised, as far as you know.

Put that in front of the right people and ask a sequence of plain questions.

What does the customer feel, and when?

Not what the system diagram says, but what a customer experiences. Orders that can’t be placed, queries that can’t be answered, payments that don’t arrive, deliveries nobody can track. Then ask which of those hurt in the first hour, which in the first day, and which only on day two.

What can we do by hand?

Most organisations have workarounds in someone’s head. A spreadsheet that can take orders, a phone number that can take bookings, a colleague who remembers how the old process worked. The exercise exposes whether those workarounds are real, who knows about them, and whether they scale beyond a quiet afternoon.

Who can approve them?

Switching to a manual process may need spending money, bypassing a control, or telling customers something awkward. Who says yes? If the answer is “the managing director”, ask what happens when the managing director is unavailable. If the answer is “it depends”, you have found something.

What do we tell customers, and who tells them?

Communications is often the thing leaders forget until the second day, by which time customers have drawn their own conclusions. A holding statement that has been agreed in advance is worth more than a brilliant one drafted at midnight.

What could stop us recovering?

This is the question that surfaces the real blockers. A contract that gives you no right to be told promptly. A supplier support line that closes at 5pm. A restore process that nobody has tested. A spare key in the locked car.

The room matters as much as the questions. Operations, technology, cyber, risk, and communications should all be there, together with anyone who owns a customer relationship. If you can get the supplier in, do. If you can’t, someone should play the supplier and be required to answer honestly about what the supplier would probably say. (Having a colleague pretend to be an uncooperative vendor is, I’m told, rather easy to cast.)

Run it to a clock. Give out information in stages, so that at hour two the team learns that the supplier’s recovery has slipped, and at hour six that a customer has posted about it online. Real events don’t arrive as a tidy briefing. They arrive in pieces, with some of the pieces wrong.

A few practical rules help. Keep the session to two or three hours. Record decisions and assumptions as they appear, because the best material is often something somebody says in passing. And resist fixing things in the room. The temptation to start redesigning the process halfway through is strong, but the value of the session is in seeing what happens unprepared. Fixes come afterwards.

This is also a safe place to get things wrong. A leadership team that discovers it has no clear decision owner during a two hour exercise has lost an afternoon. One that discovers it during a real incident may lose customers, money, or a reputation that took years to build.

What you get out of it

The output is not a thick report. It is a short list of things that would hurt, ranked by operational impact, and a clearer view of what to do about each one.

Prioritisation improves first. Security and resilience budgets tend to follow whatever feels most alarming, or whatever the last audit highlighted. A stress test gives you a different basis for deciding: the weaknesses that would stop a customer facing service, ranked by how quickly they would do it. Sometimes that points to a technical fix. Often it points to something much less glamorous, such as an unclear owner, an out of date contact list, or a supplier clause that needs renegotiating.

Ownership gets sharper. After a good exercise, someone has been named for each decision that stalled, and the delegation has been agreed for when they’re not there. That alone removes a surprising amount of risk.

Coordination improves, because people who have worked a problem together once tend to work it faster the second time. Technology and communications will have met each other before the real thing, and operations will have heard what the cyber team actually worries about.

And leaders gain something that is hard to get from a policy review: grounded confidence. Not a feeling that nothing will go wrong, which would be foolish, but a reasonable belief that when something does, the organisation can keep serving customers while it sorts out the rest.

Resilience is the ability to keep serving customers through disruption, and you only find out whether you have it by testing it.

If you’re setting priorities for 2027 over the next few weeks, a stress test is a sensible thing to put on the list. It costs hours, not months.

Start with one service

You don’t need to commission anything to begin. Choose one service your customers would notice losing within a day. Then ask three questions, and ask them of the people who actually run it rather than the people who describe it in meetings.

What could stop it operating tomorrow? How long could we manage without it? Who would lead the response?

If the answers come quickly and the people agree with each other, good. If they don’t, you’ve just found the first thing to fix, for the price of a meeting.

And when you write the answers down, keep a copy somewhere that doesn’t depend on the system you’re asking about. A printed page in a drawer is wonderfully unfashionable and has never once been encrypted by ransomware.

About the author

Laurens de Rooij

Velox Consilium brings business understanding and technology together to help organisations explore practical AI and automation solutions.

About Velox Consilium

Put insight into practice

Have a business challenge
worth solving?

Start with the problem you want to address. Let’s explore the practical next steps together.

Book a Consultation

Continue exploring.

All resources

Verified by MonsterInsights