The 2026 Allianz Risk Barometer, which surveyed 3,338 risk management experts across 97 countries, placed business interruption in 3rd place among the risks businesses cite most worldwide — behind only cyber incidents and the risk tied to artificial intelligence. That is one spot lower than in 2025, but the underlying pattern repeats every year: stopping operations is always near the top of what keeps business leaders up at night.
The problem is that "downtime is worrying" is not a number. And without a number, a proposal to invest in prevention — monitoring, backup, a second internet line, a generator — competes for priority against everything else that already arrives with a dollar value attached: hiring another salesperson, replacing the fleet, renewing a contract. Prevention almost always loses that fight, because it is the only proposal that shows up without a figure next to it.
This article does not hand you a ready-made number — no two businesses are alike enough for that to work. Instead, it shows how to arrive at your own: the line items that make up the cost of an hour of downtime, and where each one hides inside the business.
The exercise applies just as much to an industrial plant with hundreds of employees as to a mid-sized company anywhere in the country. The difference is not whether the cost exists — it is whether anyone inside the business has ever taken the time to add it up.
The price nobody puts on a spreadsheet
When a system goes down, a business measures time — how many minutes or hours until it is back — and rarely measures money. That is not carelessness: the cost of an hour of downtime is not a single line. It is the sum of line items that live in different departments and never meet on the same spreadsheet.
In Brazilian industry, the pattern is recurring. A survey by Brazil's National Confederation of Industry (CNI), which polled 1,002 industrial executives in April 2024, found that 79% of industrial companies had production affected by power outages in the prior year, and more than half faced over five such interruptions. CNI itself estimates these failures cost the country's industrial sector roughly R$65 billion a year.
The picture repeats among organizations that run their own data centers and IT infrastructure. In the Annual Outage Analysis 2025, from the Uptime Institute, 54% of organizations that suffered a severe outage in 2024 said the incident cost more than US$100,000, and one in five reported costs above US$1 million.
None of those numbers is yours. They show that the bill is real and measurable — not something a business should import from outside. The right exercise is to open up your own line items, not borrow another sector's average.
The usual way of handling it
The usual way is to feel the problem, not measure it. Someone in finance mentions "today was a bad day" after a system outage; the owner vaguely recalls a customer who complained; nobody adds anything up. With no line items to work from, the memory of an outage turns into a feeling — and a feeling does not convince whoever has to sign off on a prevention budget.
Another common approach is buying redundancy or insurance sized to the company, not to the actual problem. A business interruption policy, for instance, is usually based on the company's average daily revenue — but if nobody built that average carefully, the coverage purchased can fall well short of what an outage would actually cost.
In smaller companies, the usual way has a familiar name: leaving IT to whichever employee "knows more about computers," or to a provider who only shows up after the system has already gone down. With nobody watching the operation before it breaks, the only benchmark left is comparing it to the memory of a previous outage — one that also went unrecorded.
There is also the most common approach of all: waiting for the next outage to happen to find out its price, again. Every incident that goes unrecorded is a missed chance to turn that episode into an argument for the next budget cycle.
What has to be in place
Calculating the cost of your own downtime means separating six line items. Most of them already exist somewhere in the business — they have just never been added up in one place.
Revenue per hour of the most critical shift. Not the yearly average: the value billed during the specific hour and shift when downtime typically hits. It is not the same math for a retail store on a Saturday morning as for a factory floor at 3 a.m.
The cost of idle payroll. While the system is down, the business keeps paying the salary and benefits of people who cannot work, even though nobody has left the building.
Unbilled orders. Different from the vague feeling of "selling less": it is the specific order that came in, could not be processed, and never came back — some customers wait, others buy from a competitor in that same hour.
Penalty clauses read before the incident. Contracts with suppliers or customers that include penalties for delay or unavailability need to be reread outside of a crisis. In the middle of an incident, nobody has the presence of mind to find the exact wording of the clause.
Rework to rebuild what was lost. Bringing a system back online is not the end of the cost. A Veeam survey of 900 organizations hit by ransomware attacks in 2024 found that 57% recovered less than half of their own data on the first attempt — the rest turned into manual rebuilding, entry by entry.
The customer who does not wait. Some customers tolerate a delay; others never come back. It is worth mapping, by customer type, who cancels a contract or switches providers after an outage — that cost only shows up months later, without looking connected to the incident.
This is how Skills IT works: watching the network before it fails and regularly testing backup recovery, so that by the time someone adds up these line items, most of the bill has already gotten smaller.
What the business gains by knowing the number
The first gain is decision-making. A proposal for monitoring, redundant backup, or a second internet line stops competing on "priority" and starts competing on numbers: keeping it active all year costs less than a single hour of downtime would. It is a comparison any director can make alone, without needing an IT translator.
The second gain is budgeting. Once a business knows which line item weighs the most — idle payroll, lost orders, contract penalties — it invests first where the risk is greatest, instead of buying whatever product came up last in a vendor meeting.
The third shows up during the incident itself: a team that already knows what each hour costs decides faster what to prioritize — which system comes back first, which customer gets notified first — instead of arguing about it mid-crisis, with the clock running.
There is also an effect on the customer relationship. A business that knows its own number can explain, during an incident, what is being prioritized and why, instead of leaving a silence the customer reads as indifference. Communication backed by real cost helps keep a contract that would otherwise be lost at the first major failure.
A roadmap for finding your own number
- Ask finance what the last recorded outage actually cost. If the answer is "we don't know," that is already the first finding: the business has no habit of recording this.
- Work out revenue per hour for the busiest shift, not the yearly average. That is the number that belongs in the math, because downtime rarely picks the quietest hour.
- Add up what payroll costs per hour while the system is down. Include benefits, not just salary — it is people the business keeps paying even while idle.
- Reread contracts with penalty clauses for delay or unavailability. Do this outside of a crisis, with time to understand the actual value of the penalty.
- Ask whoever handles customers how many complain or cancel after an outage. That is the data point that rarely reaches whoever decides the IT budget.
