An online shop that stands still sells nothing -- and loses more than just the missed revenue with every passing minute. The industry-wide average cost of downtime has risen to roughly 15,000 US dollars per minute in 2026 (Splunk and Cisco, Hidden Costs of Downtime 2026), driven by large organisations with high transaction volumes. For small and mid-sized shops the direct loss is lower, but rarely below 2,000 euros per minute during peak hours (gatling.io, Cost of Downtime 2026). One other figure is decisive: whoever notices an outage only after 30 minutes pays many times more than someone who detects it in under two minutes. This is exactly where proactive 24/7 monitoring comes in -- it reports the outage before the first customer ever sees an error page.
The Per-Minute Math: What an Outage Really Costs
The simplest form of outage cost calculation is a division: annual revenue divided by the number of minutes during which the shop actually sells. A shop with 2 million euros in annual revenue and round-the-clock sales reaches just under 3.80 euros per minute on a daily average. But this average is misleading: sales are not evenly distributed. During the evening peak, on promotional days or during a campaign, revenue per minute can reach ten to twenty times the average. An outage in exactly this window hits the shop hardest -- and outages tend to cluster under load, precisely when traffic is highest.
The official benchmarks paint a clear picture. The aggregate cost of downtime for Global 2000 companies has risen to 600 billion US dollars per year (Splunk and Cisco, Hidden Costs of Downtime 2026) -- an increase of 50 percent (Splunk and Cisco, Hidden Costs of Downtime 2026) within just two years. Per organisation that equals an average annual revenue loss of around 300 million US dollars (gatling.io, Cost of Downtime 2026). For e-commerce specifically, industry analyses estimate downtime costs of 500,000 to 1 million US dollars per hour (gatling.io, Cost of Downtime 2026) at large retailers. These figures apply to high-revenue companies, but the underlying mechanism affects every shop regardless of size.
How extreme the range is becomes clear from an analysis of the highest-revenue US online retailers: the leading 25 retailers together lose roughly 14,056 US dollars per second (Gremlin, Cost of Downtime for Top US eCommerce Sites), more than 840,000 US dollars per minute. The largest single retailer alone reaches about 220,000 US dollars per minute (Gremlin, Cost of Downtime for Top US eCommerce Sites). These figures are not the benchmark for an SMB shop -- but they show the mechanism: the higher the transaction volume, the more expensive every single minute of standstill. For your own planning, the peak figure is therefore not what matters, but the individual value derived from your own revenue and sales distribution.
How to calculate your own per-minute cost
Direct Revenue Is Only the Tip of the Iceberg
The lost order value during the outage is the most visible but usually smallest part of the damage. The indirect follow-on costs -- loss of reputation, customer churn, reduced customer lifetime value, increased support load and recovery costs -- exceed the direct revenue loss by a factor of five to ten in our project experience (project experience). These costs do not arise in the outage minute but over the weeks and months that follow, and that is precisely what makes them so insidious: they do not appear in the revenue statistic of the outage day at all.
How real these follow-on costs are is backed by current surveys. Publicly traded companies record an average stock-price drop of 3.4 percent (Splunk and Cisco, Hidden Costs of Downtime 2026) following a single major outage. Among marketing leaders, nearly 20 percent (Splunk and Cisco, Hidden Costs of Downtime 2026) report that brand health recovers only after a full quarter following an incident. And the reputational effect is measurable: 94 percent (Statista, via ReviewTrackers Online Reviews Survey) of consumers say a negative review has at some point convinced them not to buy from a provider -- and since 93 percent (BrightLocal, Local Consumer Review Survey 2025) of consumers read online reviews before a purchase, error pages and frustrated posts around an outage feed directly into the buying decision.
Lost revenue
Every order that does not materialise during the outage. Directly measurable, but usually the smallest line in the total bill.
Customer churn
Anyone hitting an error page at checkout often buys from a competitor -- and does not return. The loss includes all of that customer's future orders.
Reputational damage
Negative reviews, frustrated social-media posts and declining trust take effect for weeks. Brand recovery often takes a full quarter.
Support load
During and after an outage the request volume spikes. Every request ties up staff who are not selling or optimising.
Recovery costs
Data checks, synchronisation with inventory and payment providers, reworking orders -- operations do not resume cleanly the second the shop is back.
Compliance risk
Outages caused by security incidents can trigger reporting duties, audits and fines that exceed the direct damage many times over.
The loss of customer lifetime value is particularly underestimated. Losing a customer through a poor experience means losing not one order but that customer's entire future revenue stream. A long-standing regular who churns after a frustrating experience costs many times the value of a single basket. This calculation appears in no revenue statistic from the outage day -- it spreads invisibly over years. Anyone who wants to understand the true cost of an outage has to factor in this delayed, cumulative effect rather than looking only at the collapsed daily revenue.
Why Detection Time Determines the Scale of the Damage
The scale of the damage depends less on whether an outage happens -- because sooner or later it hits every shop -- than on how quickly it is detected and resolved. The damage figure grows almost linearly with outage duration, and during peak times even disproportionately. An outage detected and immediately escalated in under two minutes remains a marginal event. The same outage, noticed only after half an hour through a customer complaint, is a serious incident with noticeable consequences.
The problem for many shops: they do not notice outages themselves at all. The operator learns of the standstill through a phone call, an email or a glance at a suddenly collapsed revenue curve -- often hours later. In that time, hundreds of visitors have seen an error page, the search-engine crawler has logged the shop as unreachable, and some customers have already churned. Proactive monitoring reverses this relationship: the system detects the outage before the first customer reports it, giving the team a decisive head start.
| Detection | Outage duration until response | Estimated direct damage (SMB shop, peak) | Indirect follow-on costs |
|---|---|---|---|
| Proactive monitoring | under 2 minutes | under 5,000 euros | low, hardly any customers see an error page |
| Manual spot check | 15-30 minutes | 30,000-60,000 euros | medium, visible complaints |
| Customer complaint | 30-90 minutes | over 100,000 euros | high, reputation and churn effects |
The values in this table are model assumptions for a high-revenue SMB shop during peak hours and serve to illustrate the mechanism. The point is not the absolute amount but the slope: every minute that passes between outage and detection multiplies the damage. Reducing detection time from 30 minutes to under two minutes typically lowers direct outage costs by more than 90 percent (project experience). Anyone who contractually secures the emergency response time turns this mechanism into a predictable lever rather than an incalculable risk.
Speed Is Revenue -- Even Without a Full Outage
Not every expensive incident is a complete standstill. A gradual slowdown costs revenue too, often unnoticed. The data is unambiguous: the highest conversion rates -- an average of 3.05 percent (Portent, Site Speed Is Hurting Everyone's Revenue) -- occur on pages that load in one to two seconds. A page that loads in one second achieves a five times higher (Portent, Site Speed Is Hurting Everyone's Revenue) conversion rate than a page with a ten-second load time.
Bounce rate reacts just as sensitively. If load time rises from one to three seconds, the probability of a bounce already increases by 32 percent (Google / Think with Google, Mobile Page Speed Benchmarks). 53 percent (Google / Think with Google, Mobile Page Speed Benchmarks) of mobile visitors leave a page that takes longer than three seconds to load. If load time rises from one to five seconds, the probability of a bounce jumps by 90 percent (Google / Think with Google, Mobile Page Speed Benchmarks). A mobile speed improvement of just 0.1 seconds raises retail conversions by 8.4 percent (Deloitte / Google, Milliseconds Make Millions).
The invisible form of an outage
This is why continuous performance monitoring belongs to a serious outage strategy. It measures not only whether the shop is reachable but how fast -- and raises an alarm when response times deteriorate. This makes problems visible before they become noticeable as a revenue drop. A slow page is, in a sense, a partial outage in slow motion: it loses customers not all at once, but continuously and inconspicuously.
How Proactive Monitoring Limits the Damage
Professional monitoring is more than a ping test every few minutes. It combines several check layers that together ensure an outage is detected, confirmed and escalated to the right person within seconds. The value comes from the interplay of fast detection, reliable confirmation and automatic escalation.
- Check intervals of 30 to 60 seconds on homepage, category, product detail page and checkout entry -- not just the homepage, but the revenue-relevant paths.
- Content validation instead of a pure status-code check: also detects partial outages where the server delivers an error page with status 200.
- Checks from multiple geographic locations, so an alarm is only triggered when several locations confirm the outage -- this avoids false positives.
- Escalating alerting: immediate notification, then escalation to the on-call team if no one responds.
- Certificate and DNS monitoring, since expired certificates and DNS errors are among the most common avoidable outage causes.
- Performance tracking that detects gradual degradations before they become a revenue problem.
The decisive effect: the time between outage and detection drops from hours to under two minutes. In this short span, hardly any direct damage arises, and the indirect follow-on costs -- reputation, churn -- remain minimal, because few or no customers see an error page. This shifts the economic balance: the running costs for professional monitoring are usually small relative to a single avoided incident. Reactive damage control thus becomes a forward-looking safeguard for revenue.
The most expensive minute of an outage is not the first -- it is the thirtieth, when nobody has responded yet.
Knowing the Causes Means Avoiding the Outages
Monitoring limits the damage when an outage occurs. Even more economical is to defuse the most common causes from the outset. Many expensive incidents trace back to avoidable factors: expired SSL certificates, overloaded databases under load, faulty updates without prior testing, exploited security gaps in outdated components or resource bottlenecks during traffic peaks.
It is precisely the combination of monitoring and preventive maintenance that noticeably reduces outage risk. A tightly held patch window reduces the window in which known vulnerabilities can be exploited -- how tight this window should be is described in our article on the patch window for zero-day gaps. Updates tested beforehand in a staging environment lead to fewer outages in live operation. And continuous uptime monitoring ensures that every incident that occurs despite prevention becomes visible immediately. How to set up such uptime monitoring is described in detail in a dedicated guide.
These building blocks interlock. Preventive maintenance reduces the number of incidents, monitoring reduces the duration of each one, and a contractually defined emergency support ensures that qualified help is immediately available in an emergency. Only this interplay turns the threatening per-minute math into a manageable figure. Anyone who treats availability as a plannable part of the business model rather than a technical side issue protects not only revenue but also the trust of their customers.
Sources and studies