Plugins Plugin hub Backup SEO Caching Security Forms WooCommerce Images Best backup plugins
Hosting Hosting directory Shared Managed Cloud India
Themes & builders Compare Rank Math vs Yoast
Tools All tools Hosting finder Plugin stack builder Cost calculator Salt generator Theme detector Who is hosting this? Speed test Headline analyzer robots.txt generator Redirect generator Uptime calculator Blog name generator
Best lists Best backup plugins Deals
Guides All guides Speed up WordPress Choose hosting Back up WordPress
How we test

Statistics · Statistics

Website Downtime Statistics (2026)

Short answer

Website downtime is less frequent than in 2020, but it is still costly. In Uptime Institute's 2025 survey, 50% of operators reported a serious outage in three years, 57% said their most recent major outage cost over $100,000, and 1 in 5 put it over $1 million. A 99.9% uptime promise still allows 8.76 hours of downtime a year.

Key takeaways

  • Half of data centre operators reported a serious outage in the past three years in Uptime Institute's 2025 survey.
  • 57% of operators said their most recent major outage cost more than $100,000, and 1 in 5 said more than $1 million.
  • 99.9% uptime still allows 43.2 minutes of downtime in a 30-day month, or 8.76 hours a year.
  • Power is the leading cause of serious data centre outages, while networking faults lead IT service outages.
  • Cloudflare recorded 225 major internet disruptions globally in 2024, with over half linked to government-directed shutdowns.
  • Splunk put downtime for the Global 2000 at $600 billion a year in 2026, or $15,000 per minute.

How often do websites go down?

Websites go down less often than a few years ago, but outages still happen, and half of data centre operators reported a serious outage in the past three years as of November 2026.

Uptime Institute found that 50% of operators in its 2025 survey reported an impactful or serious outage in the past three years. That was the lowest level recorded since 2020.

The trend has improved for five years in a row as of November 2026. Uptime Institute reported 78% in 2020, 69% in 2021, 60% in 2022, and 53% in 2024.

Severity has fallen too. Only 9% of reported incidents in 2024 were classed as serious or severe, the lowest level Uptime Institute had recorded.

For a single website, the picture is different from a data centre. Splunk found that organisations in its 2026 study averaged 60 downtime or service-degradation incidents a year. Many of those were small and brief.

What does website downtime cost?

Downtime costs large companies far more than most site owners expect, with most major outages passing $100,000 as of November 2026.

Uptime Institute reported that 57% of operators said their most recent major outage cost more than $100,000. About 1 in 5 said it cost more than $1 million.

That top figure has held for two years in a row as of November 2026. Uptime Institute reported the same 1 in 5 share above $1 million in its 2025 and 2026 analyses.

ITIC asked firms about one hour of downtime in its 2024 survey of more than 1,000 firms. Over 90% of mid-size and large enterprises put that hour above $300,000.

The same ITIC survey found that 41% of enterprises put one hour at $1 million to over $5 million. Also, 97% of large enterprises with over 1,000 staff put one hour above $100,000.

Splunk and Oxford Economics put the total cost for the Global 2000 at $600 billion a year in 2026. That is $300 million per company per year, or $15,000 per minute.

These are survey answers from large companies, not a bill for a small site. Use them to judge scale, then work out your own cost from lost sales and staff time.

How much downtime does 99.9% uptime allow?

A 99.9% uptime promise still allows 8.76 hours of downtime a year, or 43.2 minutes in a 30-day month, as of November 2026.

The maths is simple: 0.1% of the period is the allowance. Google's SRE book publishes the same figures in its availability table.

Uptime levelDowntime per monthDowntime per year
99%7.2 hours3.65 days
99.9%43.2 minutes8.76 hours
99.99%4.32 minutes52.6 minutes
99.999%25.9 seconds5.26 minutes

Each extra 9 cuts the allowance by about 10 times. Moving from 99.9% to 99.99% cuts a monthly allowance from about 43 minutes to about 4 minutes.

A host can meet 99.9% for a year and still have one bad month. That is why a monthly figure matters more than a yearly average for your site.

For most small WordPress sites, 99.9% is a fair target. Learn what that level means in plain terms in our guide to what a good uptime percentage is for hosting.

What causes most website downtime?

Power is still the leading cause of serious data centre outages, while network and connectivity faults are the most common cause of IT service outages, as of November 2026.

Uptime Institute lists power as the leading cause of impactful outages in its 2026 analysis. Failures in UPS systems, transfer switches, and generators are prominent in that group.

Its 2026 resiliency survey gives a different view for day-to-day IT services. Networking and connectivity led at 23%, ahead of power at 21% and IT system and software faults at 18%.

Human error is a factor in most major outages, not a separate cause on its own. Uptime Institute found it in between two-thirds and four-fifths of major IT and data centre outages.

IT and networking issues caused 23% of impactful outages in 2024. That share rose from earlier years, even as power remained first overall.

Fibre and connectivity outages are rising too. Uptime Institute said in May 2026 that they are more likely than before to cause extended disruption.

How long do major outages usually last?

Most major outages last hours, not days, but a small share now runs past 48 hours as of November 2026.

Uptime Institute found that 55% of publicly reported incidents were resolved within 12 hours. That figure comes from its 2025 analysis of public incident reports.

The share lasting more than 48 hours has risen for two years in a row. Longer outages are rarer, but they are becoming a larger part of the total.

A single fault can also take far longer to clear fully than to fix at first. The first service may return in minutes, while dependent services take hours to recover.

That pattern showed up in the large 2025 outages below. The root cause was fixed early, yet full recovery took most of the day.

What happened in the biggest outages of 2025?

Three large outages in late 2025 showed how one small change can take down many sites at once.

Cloudflare began failing to deliver core traffic at 11:20 UTC on 18 November 2025. Core traffic was largely normal by 14:30 UTC, and all systems were normal by 17:06 UTC.

Cloudflare called that outage its worst since 2019. It lasted about 2 hours and 10 minutes at its worst, and it was not a cyber attack.

The cause was a Bot Management feature file that doubled in size after a database permissions change. Software with a size limit on that file then failed wherever the file loaded.

Cloudflare had a second outage on 5 December 2025. It began at 08:47 UTC and ended at 09:12 UTC, about 25 minutes, and it affected 28% of applications behind its network.

AWS had a longer event on 19 and 20 October 2025. It ran from 11:48 PM PDT on 19 October to 2:20 PM PDT on 20 October, about 14 and a half hours.

A race condition in DynamoDB DNS management left an empty DNS record for the regional endpoint. DynamoDB API errors ran until 2:40 AM, while EC2 launches failed until 10:36 AM.

How do you measure downtime on a WordPress site?

Downtime is the time your site does not answer requests, and you measure it with outside checks, not with your own server logs alone.

Uptime is the share of time a service was up. The formula is (total time minus downtime) divided by total time, times 100.

A monitor that checks your home page every minute from another network gives a fair result. A check from inside your own host can miss a network fault.

Set up alerts for HTTP errors, not only for a server that is off. A site that returns a 500 error page is down for visitors, even if the server is running.

Check the result over a month, not a day. One 20-minute outage in a month is about 99.95% uptime for that month, while the same outage in a week is far worse.

Your host can also affect the result. Read how hosting affects WordPress speed before you blame a slow page on downtime alone.

What do the numbers mean for a small WordPress site?

The large-company figures show the scale of the risk, but your own cost depends on your sales, your traffic, and how long you are down.

WordPress ran 38.5% of all websites as of 19 November 2026, according to W3Techs. That share was 57.4% of sites with a known content management system.

A shared fault can therefore affect many WordPress sites at once. The Cloudflare and AWS outages in 2025 took down sites that shared the same provider, not sites that shared the same software.

No published survey gives a sound hourly figure for a small business. Do not use a large-enterprise figure as your own cost.

Work out your own number instead. Add lost sales for the hour, staff time spent on the fix, and any refunds or credits you must give.

A site that earns little while it is down still loses trust. Visitors who see an error page may not return, and search engines may show a stale page for a time.

Speed and uptime are linked in visitors' eyes. See our guide to average website load time to judge what a normal, working site should feel like.

What mistakes should you avoid?

Do not quote a single downtime cost as if it applies to every site, and do not treat a 99.9% promise as a guarantee of no outages.

  • Do not use the $5,600 per minute figure from 2014. Newer surveys from ITIC, Uptime Institute, and Splunk give current, dated figures instead.
  • Do not mix per-hour and per-incident figures. ITIC reports cost per hour, while Uptime Institute reports cost per major outage.
  • Do not rely on your host's status page alone. Use an outside monitor so a network fault is not hidden from you.
  • Do not chase five nines unless you need them. Each extra 9 costs far more than the last one, and most small sites do not need it.
  • Do not ignore small, repeated outages. Ten 5-minute outages in a month use more of a 99.9% allowance than one 40-minute outage.

A simple plan helps more than a costly promise. Keep backups, keep a monitor running, and know who to call when the site goes down.

Frequently asked questions

How much downtime is normal for a website?

There is no single normal figure, because it depends on the host, the site, and the month. A 99.9% uptime target allows about 43 minutes in a 30-day month. Many small sites see far less in a good month, and far more in a bad one. Track your own site for three months before you judge it.

Is 99.9% uptime good enough for WordPress?

Yes, for most small WordPress sites, 99.9% is a fair target. It allows 8.76 hours of downtime a year, which covers most short faults. A shop that takes orders all day may want 99.99%, which allows only about 53 minutes a year. Check your sales pattern before you pay for more.

What is the most common cause of website downtime?

It depends on what you count. Power is the leading cause of serious data centre outages in Uptime Institute's 2026 analysis. For day-to-day IT services, networking and connectivity faults led at 23% in its 2026 survey. Human error also plays a part in most major outages, often through a missed step or a bad change.

How much does one hour of downtime cost?

Large firms report very high figures, but they do not apply to a small site. In ITIC's 2024 survey, over 90% of mid-size and large firms put one hour above $300,000. For a small WordPress site, work out your own figure from lost sales, staff time, and refunds. That number is the only one you should plan around.

Does downtime hurt search rankings?

A short outage rarely causes lasting harm, but a long or repeated outage can. Search engines may show a stale page or drop a page for a time if they keep finding errors. Fix the cause, check that pages return a 200 status, and watch your search console for crawl errors after the site returns.

Sources

  1. Annual Outage Analysis 2026 Uptime Institute
  2. The Hidden Costs of Downtime 2026 Splunk
  3. Cloudflare outage on November 18, 2025 Cloudflare
  4. Summary of the Amazon DynamoDB Service Disruption in the Northern Virginia (US-EAST-1) Region Amazon Web Services
  5. ITIC Reports and Survey Results ITIC
  6. Table for Service Availability Google SRE
  7. Cloudflare Radar 2024 Year in Review Cloudflare Radar
  8. Usage statistics of WordPress W3Techs

Is your product missing, or something out of date?

Tell us what to add or correct. We read every note and reply when we need a source. A message does not buy a ranking.

About the editors

WPLegible editors

Researched from the primary sources listed above and checked against our editorial rules. About the site · How we test