Cost of Downtime Calculator
What does an hour of outage actually cost your business?
Live widget · client-side · no data leaves your browser.
What this measures
The cost-of-downtime number is the single most cited figure in any reliability or SLA conversation, but most blog posts pull it from a Gartner whitepaper that is 15 years old and pre-cloud. This calculator does the arithmetic transparently: annual revenue divided by operating hours yields the per-hour revenue base; a disruption-factor multiplier (1× revenue lost only, 1.5-2× with SLA credits and churn factored in, 3-5× for payments or healthcare) sets the per-hour true cost; the SLA tier sets the annual downtime budget. The output is the annual cost of running at that tier — and the savings from tightening one tier, so you can weigh the engineering investment.
When to use it
- 1.Building the business case for an SRE hire or a platform investment.
- 2.Setting an SLA on a contract — and pricing the risk of falling short.
- 3.Deciding whether 99.9% or 99.99% makes sense for a new service (the engineering cost grows fast above 99.9%, the business cost of being below it grows fast too).
- 4.Post-incident review when leadership wants to know what the outage actually cost.
How it works
Each SLA tier maps to a pre-computed annual downtime budget (525,600 minutes × (1 − uptime fraction)) — 87.6 hours/year at 99%, 8.76 hours at 99.9%, 5.26 minutes at 99.999%. The hourly cost is annual revenue divided by 8,760 (24×7) or 2,080 (business-hours-only) hours, multiplied by the disruption factor. Annual downtime cost is hourly cost × tier budget. The "tightening to next tier" callout subtracts the next tier's budget cost from the current one — what you would save (or what avoidable downtime is currently costing you).
Source: The 525,600-minute year arithmetic is the canonical SLO/SLA framing in the Google SRE Workbook (O'Reilly). Disruption-factor multipliers reflect industry surveys including Uptime Institute, IDC, and Gartner.
Embed this calculator on your site
One-line iframe snippet, dark/light themes, optional UTM source, and a small "Powered by" attribution back to The Art of CTO. Free for commercial and non-commercial use.
Frequently asked questions
What is the average cost of downtime per hour?+
The widely cited figure is $5,600/minute or $300,000+/hour, originally from a 2014 Gartner study of large enterprises. That number is wildly off for most SaaS businesses. A $10M ARR SaaS at 1.5× disruption factor pays ~$1,700/hour. An enterprise payments business at 5× pays much more. This calculator does the math for your actual revenue and risk profile rather than parroting one Gartner average.
How much downtime does each SLA tier allow per year?+
99% = 87.6 hours (about 3.5 days). 99.5% = 43.8 hours. 99.9% = 8.76 hours. 99.95% = 4.38 hours. 99.99% = 52.56 minutes. 99.999% = 5.26 minutes. Each additional "nine" roughly cuts allowable downtime by 10×.
Should I aim for 99.99% (four nines)?+
Almost certainly not by default. Engineering cost grows superlinearly above 99.9% (you need multi-region failover, zero-downtime migrations, deep observability, an oncall rotation, and probably a dedicated SRE). Unless your customer contracts demand it or your business is genuinely critical-path, 99.9% is the sweet spot for most SaaS — and the calculator lets you compare the cost of getting from 99.9% to 99.95% before committing.
What does the disruption factor mean?+
A 1.0× factor means "downtime costs only the revenue lost during the outage". 1.5-2× adds SLA credits to enterprise customers, refund processing overhead, and a chunk of attributable churn. 3-5× is for businesses where downtime carries reputational or regulatory cost — payments, healthcare, identity, infrastructure. Pick honestly; the number reflects how the board would account for an outage, not how engineering would.
Want the full version?
The widget above is the quick check. The full SLA Generator tool has more inputs, richer output (anti-pattern detection, contextual benchmarks, Excel export), and integrates with your Command Center workspace.
Open the full SLA Generator →