How Much Does Website Downtime Cost? A Practical Calculator

Most downtime cost statistics quoted online come from Gartner, ITIC, or Aberdeen surveys of large enterprises. The headline numbers ($5,600 per minute, $100,000 per hour, and so on) are useful as a wake-up call and useless for actual planning. Your downtime cost depends on your traffic, your conversion rate, and the alternative customers have. This guide walks through a practical calculation that fits a one-person SaaS, an agency, or a mid-sized e-commerce site.

Why generic downtime statistics mislead

Industry surveys produce huge averages because the sample skews toward Fortune 500 companies where one minute of downtime affects thousands of transactions and dozens of internal teams. Quoting $5,600 per minute back to your CFO will get you laughed out of the room if your site does $3,000 in revenue a day.

Honest planning starts from your own numbers. Hourly revenue, conversion rate during peak hours, support ticket cost per incident, and the proportion of customers who will simply try again later versus the ones who will go to a competitor. The point of the calculation is not to produce a scary number, it is to produce a defensible one that drives a sensible monitoring spend.

The four direct costs to measure

Four cost lines apply to almost every business. Estimate each one for an hour of complete unavailability during peak traffic.

• Lost transactions: peak hourly revenue times the proportion of customers who would have bought during the outage. Subtract recovered sales for customers who simply retry later. The net is often 30 to 60% of the headline number.
• Refunded transactions: customers who completed payment but did not receive what they paid for. Higher for one-shot purchases (one-off products, event tickets) than for subscriptions.
• Support cost: number of incident-driven tickets times your support cost per ticket. A surprising single line for many businesses. A bad outage during business hours can produce 50 to 200 tickets.
• SLA credits: if you offer SLA-backed contracts, automatic credits apply. Multiply the SLA penalty per affected customer by the affected customer count.

The four indirect costs to acknowledge

Indirect costs are harder to measure but compound over time. Three businesses out of five take a real revenue hit weeks after an incident, not during it.

Customer churn from the outage. Slowed sales from new prospects who saw the downtime on a review site or social media. Engineering opportunity cost (post-incident review, postmortem, the next two weeks where the team is distracted). Brand cost, which is real but unquantifiable, and which scales with how visible your service is in the press.

A simple calculation framework

A back-of-envelope formula that produces a usable number in five minutes.

• Hourly revenue at peak (e.g. $400)
• Multiplied by conversion-attempt impact, typically 0.4 to 0.7 (e.g. 0.5)
• Plus support cost, typically $20 per ticket times 30 tickets per outage hour (e.g. $600)
• Plus SLA credits, if applicable (e.g. $0 for self-serve businesses, $500+ for enterprise contracts)
• Equals direct outage cost per hour (e.g. $400 * 0.5 + $600 = $800)

When monitoring pays for itself

A monitoring tool that catches one extra outage per year, ten minutes faster than your team would have caught it manually, pays for itself if the cost of those ten minutes exceeds the annual subscription. For most small businesses, this means the monitoring spend breaks even at the first incident of the year.

The harder question is how much monitoring to spend on. A common ratio: spend 0.5 to 1% of your annual downtime cost on monitoring infrastructure. A business with a $20,000 annual downtime-risk should budget $100 to $200 a year for monitoring, which is well within paid-plan pricing.

The decision the calculation drives

Once you have a number, three questions become easier to answer. Is paid monitoring worth it (almost always yes above $5,000 annual exposure). How fast should checks run (faster cadence is justified above $20,000 annual exposure). Do you need a status page (yes if customer-facing, regardless of revenue). The calculation does not need to be precise. It needs to be defensible enough to make those three decisions without arguing.