TL;DR
Downtime costs SMBs an average of $8,000 per hour (Datto, 2023). 41% of companies find out from their customers first — not from an alert (Splunk + Oxford Economics, 2024). For indie SaaS founders, the real damage isn't the outage itself. It's the hours-long window before you even know it happened.
You built the product. You shipped it. You told people about it. And one day — probably on a Friday evening, probably while you're away from your desk — it goes down.
Not for a second. Not for a minute. For two hours. Three. Maybe more. Nobody paged you. Nobody sent an alert. You find out because a user posts about it, or your inbox fills with "is your site broken?" tickets, or — the worst version — a churned customer replies to your welcome email with "I tried your product once and it didn't work."
This is not a hypothetical. According to Cockroach Labs' State of Resilience 2025, the average outage lasts 196 minutes before it's resolved. And for 41% of organizations, customers are the ones who detect the problem first.
So before we talk about cost, let's be clear about what "downtime" actually means for a small SaaS: it's not just your server being offline. It's the gap between when something broke and when you knew about it — and everything that leaked out during that gap.
The numbers most founders dismiss
The headline downtime cost figures are enterprise numbers, so indie founders tend to dismiss them. Gartner's widely cited baseline puts average downtime at $5,600 per minute across all organizations. ITIC's 2024 survey found 90% of mid-size and large enterprises face hourly costs exceeding $300,000. You're not a Fortune 500 company. So you move on.
The mistake is assuming these costs don't scale down. They do. Datto's 2023 State of the Channel Report surveyed small and medium businesses specifically and found the average downtime cost for an SMB is $8,000 per hour. Eight thousand dollars. Per hour.
$8,000/hr
Average downtime cost for SMBs — Datto, 2023
ITIC breaks it down further: for a micro-SMB with around 50 employees on an $80K average salary, idle labor costs alone during downtime run to $1,923 per hour — before a single dollar of lost revenue, customer churn, or recovery engineering time is counted.
The cost breakdown for small SaaS
When a SaaS goes down, the cost isn't one number. It's four categories hitting simultaneously:
Direct revenue loss
If you have paying users, every minute they can't access your product is time you're not delivering what they paid for. For a SaaS billing $29/month, that's roughly $0.04 per user per minute in undelivered value. Multiply by your user count. Now multiply by 196 minutes.
Customer churn
A 2023 Zealousys survey found 68% of SaaS customers would consider switching providers after just one major outage. Not a pattern of outages — one. The SaaS B2B average churn rate is already around 3.5% per month. An outage significantly accelerates that for affected users.
Support load
Every minute your product is down, support tickets are stacking up. PagerDuty's 2024 Customer Incidents Survey found customer-impacting incidents increased 43% year-over-year. For small teams, the equivalent is your entire Thursday disappearing into incident response instead of product work.
Trust erosion
This one is hardest to quantify, but it's the one that lingers. PagerDuty found 90% of IT leaders confirmed outages reduced customer trust in their organization. Splunk's research found that after a major incident, stock price equivalents drop and take an average of 79 days to recover. You have word-of-mouth, review scores, and renewal conversations — all affected by the same dynamic.
The detection problem is worse than the downtime
Here's the stat that should genuinely concern every indie founder:
41%
of tech companies say customers detect downtime before their internal team does — Splunk + Oxford Economics, 2024 (n=2,000)
This isn't a minor operational embarrassment. It's a structural trust problem. When your user encounters a broken app and emails you about it, they've already moved from "this is a tool I rely on" to "this is a tool I can't rely on." That mental shift is very hard to reverse.
New Relic's Observability Forecast 2025, based on 1,700 technology professionals across 16 countries, found that 41% of IT leaders identify issues through manual checks, customer complaints, or incident tickets — after the fact. Without continuous monitoring, the average discovery window for a small team is 3–6 hours. With monitoring, detection happens in under a minute.
Cockroach Labs found that only 20% of organizations describe themselves as fully prepared for an outage, and larger enterprises are 49% more likely to have continuous monitoring than smaller organizations. The monitoring gap is, almost by definition, a small company problem.
What happens to your users during an outage
Three things happen when a user hits your broken product:
They don't wait
Google's research confirms 53% of mobile users abandon a site that takes more than 3 seconds to load. An outage doesn't slow things down — it removes the product entirely. A 2025 synthesis of UX research found 88% of users won't return to a site after a negative experience.
They don't know it's temporary
Without a public status page, users have no way to distinguish "back in 5 minutes" from "this product is abandoned." The silence is almost worse than the outage. A visible, actively updated status page — even just saying "We're aware and working on it" — dramatically changes how the user interprets what's happening.
68% consider switching
From a 2023 Zealousys survey on SaaS customer behavior after outages. After one incident. The vocal ones create reputation drag. The quiet churners drain your renewal rate. The ones who post affect your trial conversion.
The hidden costs nobody counts
Beyond revenue and churn, two costs consistently go untracked:
- —Engineering time. New Relic's 2025 research found engineers spend 33% of their time — roughly 13 hours per week — addressing IT disruptions. For a solo founder or two-person team, that's not "engineering time." That's product-building time.
- —Recovery lag. Splunk's research found revenue recovery after a major outage event takes an average of 75 days. Cockroach Labs found 48% of teams work overtime and weekends to restore normal operations after downtime. The acute incident has a long tail.
Real incidents, real numbers
- —CrowdStrike, July 2024 — Faulty sensor update, ~8.5M Windows endpoints affected. Fortune 500 losses: $5.4 billion (Parametrix). Delta Air Lines alone: $500 million.
- —GitHub, 2024 — 124 incidents caused ~800 hours of degraded performance — the equivalent of 100+ working days of disruption across a platform serving 100 million developers.
- —J.Crew, Black Friday 2023 — 5-hour e-commerce outage during peak season. Estimated cost: $775,000 in lost sales.
- —AT&T, February 2024 — Network configuration error blocked 92 million calls including 25,000 emergency 911 calls. FCC settlement: $950,000.
These are large-company incidents. But the underlying failure modes — a config change without a rollback, a deploy that wasn't watched, a dependency that nobody checked — are identical to what takes down a solo SaaS founder's app at midnight.
Calculate your number
You don't need a consultant to estimate what downtime costs your SaaS. Here's the back-of-envelope math:
Hourly Revenue Loss = (MRR ÷ 730) × affected_user_percentage Churn Risk = active_users × 0.68 × ARPU × churn_probability Engineering Cost = hourly_rate × (hours_to_detect + hours_to_resolve) Total = Revenue Loss + Churn Risk + Engineering Cost
For a SaaS with $5,000 MRR and 200 active users at $29/month: a 3-hour outage with full user impact produces roughly $20 in direct lost revenue. But 68% churn consideration risk on just 5% of affected users puts $199 in future MRR at risk. Add 4 hours of engineering time at $100/hr and you're at $600 before a single user actually churns. The direct revenue number looks small. The compound number doesn't.
Frequently asked questions
How much does downtime cost a small SaaS per hour?
Based on Datto's 2023 SMB research, the average downtime cost for a small-to-medium business is $8,000 per hour. For micro-SMBs, ITIC estimates idle labor costs alone run to approximately $1,923 per hour before revenue and churn impacts are added.
How long does the average outage last?
Cockroach Labs' State of Resilience 2025 found the average outage lasts 196 minutes — just over three hours — before resolution. Organizations with continuous monitoring in place resolve incidents significantly faster.
What percentage of companies find out about downtime from customers first?
According to Splunk and Oxford Economics' 2024 research across 2,000 executives, 41% of tech companies report that customers are often or always the first to detect downtime at their organization.
Does downtime cause customer churn?
A 2023 Zealousys survey found 68% of SaaS customers would consider switching providers after a single major outage. PagerDuty's 2024 research confirmed that 90% of IT leaders say outages reduced customer trust in their organization.
What's the difference between uptime monitoring and a status page?
Uptime monitoring detects when your service goes down and alerts you immediately. A status page is the public-facing communication layer — it tells your users what's happening and shows your historical uptime. Monitoring alerts you, the status page informs your users.
SourcesDatto State of the Channel 2023 · ITIC 2024 Hourly Cost of Downtime Survey · Splunk + Oxford Economics "Hidden Costs of Downtime" (June 2024, n=2,000) · New Relic Observability Forecast 2025 (n=1,700) · Cockroach Labs State of Resilience 2025 · PagerDuty 2024 Customer Incidents Survey · Zealousys 2023 · Parametrix CrowdStrike Impact Estimate 2024 · Google mobile speed research · EnterpriseAppsToday UX synthesis 2025