How to Work Out What Downtime Costs Your Business

You will find many headline numbers for the cost of downtime. Most come from surveys of large enterprises, and they say very little about a 5-person SaaS company or a shop with one warehouse.

Your own number is more useful, and it is not hard to estimate. It helps you decide how much to spend on prevention, how often to check, and whether someone needs to be on call at night.

The four parts of the cost

For one outage, add up four things:

1. Lost revenue: sales that happen during the outage window and do not come back later.
2. Credits and refunds: anything you pay back because of the outage, for example SLA credits.
3. Team time: the hours people spend fixing the problem and answering customers.
4. Customer loss: customers who leave earlier than they would have, because of the outage.

Some costs are real but hard to measure, such as trust or a missed sales demo. Leave them out of the number and mention them next to it. A clear, conservative number is more useful than a large, vague one.

Example 1: an online shop

The numbers below are an example. Replace them with your own.

• Revenue: $2,000,000 per year.
• Average revenue per hour: $2,000,000 / 8,760 hours = $228 per hour.
• Traffic is not flat. In this shop the busiest evening hours bring about 2.5 times the average, so a peak hour is worth about $570.
• Outage: 45 minutes during a peak hour. Revenue in that window: 0.75 × $570 = $428.
• Not every lost sale is lost for good. Some customers come back later. Assume 40% do not: 0.4 × $428 = $171 lost revenue.
• Team time: 2 people for 1.5 hours at $60 per hour = $180.
• Support: 20 emails at 5 minutes each, at $40 per hour = $67.

Total: about $420 for this outage.

That number may look small. Two things make it grow fast: the length of the outage and the time of day. The same outage lasting 4 hours over a busy evening costs several times more, because each extra hour costs another $570 of revenue at risk.

Example 2: a SaaS product

Again, an example with numbers to replace:

• 1,000 customers, $40,000 monthly recurring revenue, so $40 per customer per month.
• Outage: 45 minutes on a weekday afternoon.

For a subscription product, the direct revenue loss of a short outage is close to zero. Nobody's subscription is paused for 45 minutes. The costs are elsewhere:

• Team time: 3 people for 2 hours at $80 per hour = $480.
• Support: 25 tickets at 10 minutes each, at $50 per hour = $208.
• Credits: if your contracts promise uptime credits, add them here. In this example, none.
• Customer loss: assume the outage makes 0.5% of customers leave earlier than they would have. That is 5 customers. If an average customer would have stayed another 12 months: 5 × $40 × 12 = $2,400.

Total: about $3,100, and most of it is customer loss.

The customer-loss number is the most uncertain one. Use a small percentage, and be honest that it is a guess. If you have data on cancellations after past incidents, use that instead.

Where monitoring fits in

Downtime cost grows with the length of the outage. The length has two parts:

• Time to know: from the start of the failure to the moment someone knows about it.
• Time to fix: from that moment to recovery.

Monitoring shortens the first part. Without it, "time to know" is the time until a customer complains and the complaint reaches the right person. That can be hours, especially at night or on a weekend.

With monitoring, detection takes minutes. As a rough guide, the delay is about half the check interval plus the time needed to confirm the failure. For example, with checks every minute and 2 failures in a row required, you know within about 2 to 3 minutes.

After that, most of the cost is in human response time: does the alert reach someone who can act, and do they trust it enough to get up? That is why false alarms are expensive too. A team that gets several false alarms a week starts to ignore alerts, and the next real outage runs for longer.

Use the number

Once you have an estimate per hour of downtime, you can answer practical questions:

• Is a faster check interval worth its price? Compare the saved minutes with your cost per minute.
• Does someone need to be on call at night? Compare your night-time cost per hour with the cost of on-call.
• Which services matter most? Estimate the number separately for checkout, login and the marketing site. They are rarely the same.

Write the number down and update it once a year. It turns "we should improve reliability" into a decision you can price.

StatusTick shortens the "time to know" part and keeps alerts trustworthy: it confirms a failure from several regions before it alerts, and sends the alert with the facts needed to act. It launches in Q1 2027; join the free beta.