Per Outage Duration

What does an 8-hour outage cost? A full work day of downtime, priced.

Eight hours is the length of a standard work shift. It is the length at which payroll burn for the affected headcount becomes a line you cannot omit from the business case. Mid-size enterprises lose around $2.4 million in this window. Large enterprises run from $8 million to $40 million. The British Airways May 2017 IT failure, the most-cited reference event in this duration class, ran 12 hours of full ground-stop and was disclosed at around £80 million.

Headline Numbers

8-hour outage cost by company size

The figures below multiply the most-cited per-hour benchmarks by eight. They are a defensible floor, not the worst case. Past hour four, productivity loss approaches 100% of affected headcount as people stop working around the outage and start sitting in coordination meetings. Past hour six, customer-comms cost from executive time and PR support typically doubles relative to the first half of the incident.

SegmentPer hour8-hour costSource
Small business (under 50 staff)$25,620$204,960Ponemon 2022
Mid-size (50 to 500 staff)$300,000$2,400,000ITIC 2024
Large enterprise (500+ staff)$1,000,000$8,000,000ITIC 2024 lower band
Large enterprise, top quartile$5,000,000$40,000,000ITIC 2024 upper band
Finance, large banks (peak)$9,300,000$74,400,000ITIC 2024

Sources: ITIC 2024 Hourly Cost of Downtime Survey; Ponemon Institute 2022 Cost of Unplanned Datacenter Outages. Figures in USD, as of 2026-05-18.

The Payroll Lens

Payroll burn alone during an 8-hour outage

For knowledge-worker organisations, the payroll-burn calculation is the cleanest line on the page because it is independent of revenue model, customer mix, and contingent assumptions. Number of affected employees, fully-loaded hourly cost (salary plus benefits plus overhead, typically 1.3 to 1.5x base pay), times eight. This is the floor of what the outage cost the business in dollars that already left the building.

Affected headcountFully loaded $/hr8-hour payroll burn
50$75$30,000
200$75$120,000
500$75$300,000
1,000$100$800,000
5,000$100$4,000,000

Two refinements often come up in board reviews. First, not 100% of headcount is fully blocked by every outage. A finance team might still work spreadsheets when CRM is down. Apply a percentage (40% to 80% is typical for partial outages). Second, productivity does not fully resume the moment systems come back. Context switching costs roughly 15 to 30 minutes per person to return to focus, per Gloria Mark's research on attention recovery. Add 10 to 20% to the headline burn to capture this.

Reference Event

British Airways, 27 May 2017

The reference event for this duration class is the British Airways IT failure of 27 May 2017. A power surge at a data centre near Heathrow took down most of BA's passenger-facing systems for approximately 12 hours, with knock-on effects for the entire weekend (a Saturday before a UK bank holiday).

The numbers BA disclosed

  • Approximately 75,000 passengers stranded over the bank holiday weekend
  • ~672 flights cancelled at Heathrow and Gatwick
  • £80 million estimated cost disclosed by parent IAG in subsequent reporting (IAG investor materials)
  • EU261 compensation claims drove the bulk of the direct cost (passenger compensation, hotels, rebooking)
  • Customer-acquisition cost spike in the following quarter as competitors targeted BA frequent flyers

What makes BA 2017 useful as a teaching case is that the dollar cost was almost evenly split between the contingent compensation line (EU261, rebooking) and the brand damage that materialised over the following six months. The brand line is the one most internal business cases under-model. For an 8 to 12 hour customer-facing outage at a B2C brand of any scale, plan on the public-attention component being at least 30% of the eventual total.

Cost Components

What an 8-hour bill actually looks like

Direct revenue (30 to 45% of total)

Annual revenue divided by 8,760 hours times eight times the industry at-risk percentage. For ecommerce the at-risk percentage approaches 100% during front-end outages. For SaaS subscription it is closer to 5% (the contractual revenue accrues regardless), though renewal risk is real.

Productivity (20 to 30%)

The payroll-burn table above. By hour six this line is at its theoretical maximum because most affected workers have stopped attempting workarounds.

Recovery (8 to 15%)

Overtime, premium vendor support, expedited replacement hardware, the cumulative cost of every person in the war room. For an 8-hour incident the recovery line typically includes at least two shift changes of the incident response team, which doubles the headcount touched.

Reputation and churn (20 to 35%)

The line that hits in the following quarter. Customer churn from a public 8-hour B2C outage typically runs 2 to 5% incremental over 90 days. B2B SaaS sees lower headline churn but lower-band ARR impact via renewal pricing pressure.

Customer compensation

Where BA 2017 spent most of its money: EU261, ticket refunds, hotels, rebooking on other carriers. For SaaS the equivalent is SLA credits to your enterprise customers (typically 10 to 30% of monthly fee for the affected service).

Regulatory and disclosure

For public companies, the SEC 8-K materiality determination is routine for an 8-hour customer-facing outage. The disclosure itself is free, the share price reaction is not.

Frequently Asked

Common Questions

Is an 8-hour outage worth specifically planning for?
Yes. It is the longest single-shift duration, which means it caps the payroll burn for a single business day. It is also the threshold above which executive-comms cost and EU261-style customer compensation become serious line items. Most internal DR plans use 8 hours as a key planning case alongside 1 hour and 24 hours.
How does the cost split between revenue and non-revenue lines at 8 hours?
Roughly 30 to 45% direct revenue, 20 to 30% productivity (payroll burn), 8 to 15% recovery and incident response, 20 to 35% reputation and churn, with regulatory and customer-comms as smaller residuals. The split varies enormously by industry. Airline outages are payroll-and-compensation heavy. SaaS outages are reputation-and-churn heavy.
What was the cost of the British Airways May 2017 outage?
Approximately £80 million disclosed by IAG (BA's parent) in subsequent reporting. Of this, the majority went to EU261 compensation, rebooking, and hotels for the 75,000 stranded passengers over the bank-holiday weekend. The brand-damage line cost more in the following 12 months as competitors targeted BA's frequent-flyer base.
Do most data center outages last 8 hours?
No. Per Uptime Institute's Annual Outage Analysis 2025, the modal significant outage lasts 1 to 2 hours and the median is around 2 hours. Outages of 8 hours or more account for roughly 8 to 12% of significant outages, but their share of total dollar impact is disproportionately high because cost does not scale linearly with duration.
How do I cite an 8-hour figure in a budget request?
Use ITIC 2024 mid-size median ($300,000 per hour) times eight = $2.4 million. State the assumption explicitly: 'a single 8-hour customer-facing outage on our highest-revenue service'. Compare to the annual cost of the proposed mitigation. If the mitigation costs less than one expected event per year, the case is straightforward. Use Uptime Institute frequency data to estimate event probability.
Does cloud SLA credit help at the 8-hour mark?
Marginally. The deepest credit tier on AWS, Azure, and GCP is 25 to 30% of the monthly service fee. On a $20,000 monthly spend that is $5,000 to $6,000. Against an $8 million business loss the recovery is 0.07%. The credit's economic purpose is to signal that the provider takes the SLA seriously, not to compensate the business.

Related

Updated 2026-04-27