Downtime Cost Calculator

No survey-derived "average cost of downtime" numbers here; they're measured on Fortune 500s and don't transfer. Enter your own figures and get an estimate you can actually defend in a budget discussion.

How to calculate the cost of downtime

The direct cost of an outage comes down to two things happening at once: revenue you can't earn while you're down, and the salaried time your team burns getting you back up. The formula this calculator uses is straightforward:

Total direct cost = lost revenue + response cost, where lost revenue is your hourly revenue multiplied by the share of it the outage blocks multiplied by the outage length, and response cost is the number of people responding multiplied by their loaded hourly cost multiplied by how long they're pulled in. Divide the total by the outage length in minutes and you get a per-minute figure, which is the number worth remembering, because it's what every minute of slow detection is quietly costing you.

What the formula covers, and what it can't

Lost revenue here assumes demand during the outage simply disappears. For some businesses it shifts instead: a customer who couldn't check out at 14:00 buys at 16:00, so the true loss is smaller. For others it's the opposite; a failed checkout sends the buyer to a competitor permanently, and the loss is bigger than the window suggests. Adjust the "% of revenue blocked" field to reflect which kind of business you run rather than leaving it at 100.

The response cost line only counts the people firefighting. It ignores the project work they dropped, the context they'll spend the next morning rebuilding, and any SLA credits you now owe customers. If you have contractual SLAs, our uptime SLA calculator tells you how close an incident put you to the threshold.

The costs this number leaves out

The figure above is deliberately the defensible floor, not the full picture. Several real costs are hard to put a clean number on, so they're excluded rather than guessed at:

• Customer churn. Some users who hit the outage won't come back. For a subscription business, losing a handful of customers to one bad incident can dwarf the revenue lost during the window itself.
• SLA credits. If you've promised contractual uptime, an outage may obligate you to refund a slice of the bill. That's a direct cash cost triggered by the downtime, separate from lost sales.
• Reputation and trust. A public outage, especially one your customers notice before you announce it, costs credibility that's slow and expensive to rebuild.
• Opportunity cost. Every hour your engineers spend firefighting is an hour not spent shipping. On a small team, a few bad incidents a month can quietly consume a meaningful chunk of your roadmap.

Because these vary so much between businesses, a good rule is to treat the calculator's output as the minimum, then reason about how much your specific situation adds on top.

Why "industry average" downtime costs mislead

Search for the cost of downtime and you'll find headline figures in the thousands of dollars per minute. Those come from surveys of large enterprises, where a minute of downtime really can cost that much, and they're close to useless for a small SaaS or a solo product. Averaging a bank's trading platform with a hobbyist's side project produces a number that describes neither. That's why this calculator takes no industry constant and asks only for your own revenue and team figures. An estimate built from your real numbers is one you can actually put in front of someone and defend, which a borrowed enterprise average never is.

The cheapest minutes to recover are the first ones

Every outage has two phases: nobody knows yet, and someone is fixing it. The first phase is pure waste; the meter runs and nothing improves. Five-minute check intervals plus an alert that lands in an unread inbox can easily mean 15 minutes of phase one. At the per-minute cost this calculator just showed you, tightening detection is usually the highest-return reliability spend available. If detection currently takes you 15 minutes and you cut it to under one, multiply the minutes saved by the per-minute cost above, and that's the recurring return on better monitoring, per incident.