"How much does downtime cost us?" is usually answered with a scary industry average - thousands of dollars per minute. That figure describes banks and airlines. It's useless for deciding how much to invest in reliability if you run a SaaS, an agency or an online store.
Here's a formula you can fill in with your own numbers in five minutes.
The formula
Cost of an outage =
lost revenue
+ lost leads × lead value
+ wasted ad spend
+ staff time
+ SLA credits / refunds
+ long-term cost (churn, reputation, SEO)
The first five are easy to measure. The last one is real but fuzzy, so estimate it separately and don't let it dominate.
1. Lost revenue
Take the revenue that flows through the site in a normal month and divide by the hours in a month (about 730):
Revenue per hour = monthly online revenue ÷ 730
Traffic isn't flat, so weight it: an outage at peak time costs 2-3× the average hour, one at 4 a.m. costs a fraction. Some buyers come back later, so for most stores the true loss is 40-80% of the gross figure, not 100%.
Example: a store doing $60,000/month → $82/hour on average. A two-hour outage on a Saturday afternoon (2.5× peak weighting, 60% of buyers don't return) costs about $246 in sales.
2. Lost leads
For B2B and SaaS sites, sign-ups and demo requests matter more than direct sales:
Lead value = average first-year customer value × lead-to-customer rate Lost leads per hour = monthly leads ÷ 730 × peak weighting
Example: 300 trial sign-ups a month, 8% become customers worth $600 in year one → each sign-up is worth $48. A three-hour weekday outage (2× weighting) loses about 2.5 sign-ups, or $118.
3. Wasted ad spend
This is the line most people forget. Paid campaigns keep running while the landing page returns an error, and every click is paid for:
Wasted spend = daily ad budget ÷ 24 × hours down
$400 a day on Google Ads and a four-hour outage = $67 spent sending people to an error page. Google Ads may also disapprove ads whose destination returns errors, which adds a relaunch delay.
4. Staff time
Staff cost = people involved × hours each × loaded hourly cost
Count everyone: engineers fixing it, support answering "is it down?" tickets, and a manager writing the customer email. Three people for two hours at $60/hour is $360 - often more than the lost sales.
5. SLA credits and refunds
If you promise uptime in a contract, check what a breach triggers. A typical SaaS SLA gives a 10% credit when monthly uptime drops below 99.9% - that's only 43 minutes of downtime a month. Use our uptime calculator to see how much downtime your SLA actually allows, and how uptime is calculated for the fine print.
6. The long-term cost
- Churn. For subscription products, one bad outage in the first month of a trial is often the end of it.
- Trust. Customers who learn about an outage from Twitter rather than from you remember it.
- SEO. An outage of a few hours won't move rankings. Repeated or multi-day 5xx errors cause Google to crawl less, and URLs that keep failing can drop out of the index until they recover.
A full worked example
A small SaaS company: $40,000 MRR, 300 trial sign-ups a month, $300/day ads, a 99.9% SLA for its larger customers. It has a 3-hour outage on a Tuesday afternoon, found by a customer after 70 minutes.
| Item | Estimate |
|---|---|
| Lost leads (≈2.5 sign-ups × $48) | $120 |
| Wasted ad spend (3h of $300/day) | $38 |
| Staff time (4 people × 2.5h × $60) | $600 |
| SLA credits (10% for 5 enterprise accounts at $800) | $400 |
| Churn risk (1 account lost, $150/mo × 12) | $1,800 |
| Total | ≈ $2,958 |
Look at what drives the number: duration. Every item scales with how long the site was down, and 70 of those 180 minutes passed before anyone knew.
How to reduce the cost
- Detect in minutes, not hours. External uptime monitoring with 30-60 second checks turns "a customer told us" into "we knew within two minutes". In the example above, that alone cuts the outage by more than a third.
- Alert the right person, once. Route alerts to the people on call and require consecutive failures before paging, so alerts stay trusted. (How to set up alerts people actually read.)
- Tell customers. A public status page cuts "is it down?" support tickets and protects trust.
- Prevent the predictable ones. Expired SSL certificates and domains are 100% avoidable. SSL monitoring warns you 30, 14, 7 and 1 day ahead.
Rule of thumb: if one outage costs more than a year of monitoring - and it almost always does - monitoring is the cheapest reliability work you can do.
Frequently asked questions
How much does website downtime cost per hour?
It depends almost entirely on how much revenue and how many leads flow through the site per hour. A small online store doing $20,000 a month loses about $28 of sales per hour down, plus ad spend, staff time and churn. The formula in this article gives your own number.
What is the average cost of downtime?
Widely quoted industry averages run into thousands of dollars per minute, but they describe large enterprises. For small and mid-sized businesses the honest answer is to calculate it from your own revenue, lead value and team costs.
Does downtime affect SEO?
Short outages do not. Long or frequent ones can: Googlebot slows crawling when it sees server errors, and pages that return errors for days may be dropped from the index until they recover.
How can I reduce the cost of downtime?
Mostly by shortening it. Detect outages in about a minute with external monitoring instead of waiting for customers to report them, alert the right person, and keep a status page so customers know you are on it.