What a 99.9% Uptime Guarantee Actually Gets You
A "99.9% uptime guarantee" sounds like a promise your site will basically never go down. In practice, 99.9% allows 43.8 minutes of downtime every month -- about 8.8 hours a year -- and the "guarantee" behind it usually pays out a small prorated credit on your hosting fee, not compensation for lost sales. The number is a marketing line more than a service commitment, and the gap between 99.9% and 99.99% is the part that actually matters for a store or a lead-generating site.
Disclosure: HostingDive earns a commission when you buy through our links, at no extra cost to you. Ratings come from independent testing.
The Short Answer
An uptime guarantee is a service-level agreement (SLA) that defines how much downtime the host considers acceptable and what it owes you if it exceeds that. It is not a guarantee your site stays up. Read it as two things: a downtime allowance (the percentage) and a remedy (usually service credits you have to request). For most buyers, measured third-party uptime data matters far more than the advertised percentage.
What the Percentages Actually Mean
Each additional nine cuts the allowed downtime by a factor of ten. The jump from 99.9% to 99.99% is not a rounding difference -- it is the difference between most of a workday offline per year and a few minutes.
| Uptime | Allowed downtime/month | Allowed downtime/year | What it means for you |
|---|---|---|---|
| 99% ("two nines") | ~7.3 hours | ~3.65 days | Fine for a hobby site; unacceptable for a store |
| 99.9% ("three nines") | ~43.8 minutes | ~8.8 hours | The common shared-hosting promise; a workday offline per year |
| 99.95% | ~21.9 minutes | ~4.4 hours | Better managed and cloud tiers |
| 99.99% ("four nines") | ~4.4 minutes | ~52.6 minutes | What a revenue-generating site should target |
| 99.999% ("five nines") | ~26 seconds | ~5.3 minutes | Enterprise infrastructure; rare and expensive |
Figures are the arithmetic downtime each availability level permits, rounded. A host advertising 99.9% is telling you, in the fine print, that nearly nine hours offline in a year is within its promise.
What the Guarantee Actually Pays
When a host misses its SLA, the remedy is almost always service credits -- a prorated slice of your hosting fee for the affected period, not a refund of the business you lost. On a plan that costs $15 a month, an hour of downtime might entitle you to a credit worth a few cents. Three conditions usually apply, and they are where the guarantee loses most of its force. First, credits are not automatic: you have to notice the downtime, document it, and formally request the credit within a stated window, often 30 days. Second, credits are typically capped at a percentage of one month's fee no matter how long the outage ran. Third, the clock only counts downtime the host acknowledges from its own monitoring, which may not match what your visitors experienced.
An uptime guarantee transfers almost none of your downtime risk to the host. If an outage costs you a day of sales, the SLA credit will not come close to covering it. Treat the guarantee as a signal of how confident the host is, not as insurance.
Scheduled Maintenance and Other Exclusions
The percentage almost never counts everything. Standard SLA carve-outs exclude scheduled maintenance windows, downtime caused by your own code or plugins, outages from third-party services (a CDN, a DNS provider, a payment gateway), and force-majeure events. A host can run its advertised 99.9% while taking your site down for a two-hour maintenance window that does not count against the number, because the maintenance was announced. This is not necessarily dishonest -- maintenance is real work -- but it means the advertised figure and your lived experience can legitimately differ.
What Actually Takes Sites Down
The uptime percentage says nothing about the causes, which is worth understanding because they decide whether a higher tier would even help. On shared hosting, the most common cause is the neighbors: dozens or hundreds of sites share one server, and a traffic spike or a runaway process on someone else's site can slow or stall yours. Moving to a plan with isolated resources -- a VPS or a managed tier with per-account limits -- removes that class of problem, which is often the real reason a store outgrows shared hosting well before it outgrows the storage or bandwidth.
The next tier of causes sits with your own site: a heavy plugin, an unoptimized database, or a traffic surge that exhausts the plan's memory. These do not count against most SLAs and no uptime guarantee will fix them; a caching layer, a CDN, or a resource upgrade will. Finally there is genuine host-side failure -- a hardware fault or a data-center incident -- which is the only category the guarantee is really about, and the rarest of the three. Matching the fix to the cause matters more than chasing a bigger number: if your outages come from noisy neighbors or your own plugins, a fancier SLA changes nothing.
How to Verify Real Uptime
- Ignore the advertised number and measure it yourself: put the site on an independent uptime monitor that checks every minute from multiple locations. A free tier is enough to build a real record over a few months.
- Look for third-party status history, not a status page the host controls: a host-run status page can under-report. Independent monitoring communities and review testing carry more weight.
- Read the SLA before you buy, not after an outage: find the credit request window, the cap, and the exclusion list. If the SLA is hard to find, that itself is information.
- Weigh support response time alongside uptime: when the site does go down, how fast a human responds often matters more than the last decimal of the uptime percentage.
- Match the tier to the stakes: a brochure site can live with 99.9%; a store losing hundreds of dollars an hour should be on infrastructure that measures at 99.99% or better, regardless of what the marketing says.
Who Should Care About the Difference
- Care a lot if: your site takes orders, books appointments, or runs paid traffic -- downtime is direct lost revenue, and the 99.9%-to-99.99% gap is real money.
- Care less if: the site is a portfolio, blog, or brochure where an hour offline is an inconvenience, not a loss. For that use, 99.9% shared hosting is a reasonable choice.
Is an Uptime Guarantee Worth Anything?
It is worth something as a confidence signal and close to nothing as insurance. A host willing to publish a specific SLA with a real credit mechanism is making a checkable claim, which is better than vague "reliable hosting" language. But the guarantee will not protect your revenue, and the advertised percentage is the ceiling of the host's promise, not a description of your experience. Buy on measured uptime and support response, use the SLA percentage only to compare promises like-for-like, and size the tier to what an hour of downtime actually costs your site.
Compare measured uptime and renewal pricing across hosts on HostingDive -- independent testing data, not marketing numbers.
Frequently Asked Questions
- How much downtime does 99.9% uptime allow?
- About 43.8 minutes per month, or roughly 8.8 hours per year. Each additional nine cuts that tenfold: 99.99% allows about 4.4 minutes per month, and 99.999% allows about 26 seconds per month.
- Does an uptime guarantee mean I get my money back if my site goes down?
- Not your lost revenue. The typical remedy is a service credit -- a prorated slice of your hosting fee for the downtime period, usually capped at a percentage of one month's fee, and only if you request it within the stated window. It does not compensate for lost sales.
- Why does my site feel like it goes down more than the guarantee says?
- Standard SLAs exclude scheduled maintenance, problems caused by your own plugins or code, and third-party outages such as DNS or CDN failures. A host can hold its advertised percentage while your visitors still hit downtime that does not count against the SLA.
- What uptime level should an ecommerce site look for?
- Target measured uptime of 99.99% or better for a revenue-generating store, since the difference from 99.9% is roughly eight hours of potential downtime a year. Just as important, verify the figure with independent monitoring rather than trusting the advertised number.
- Is a host with no uptime guarantee automatically worse?
- Not necessarily, but a published SLA is a checkable commitment, which is worth more than vague reliability language. The stronger signal is measured uptime from independent monitoring. A host with great real-world numbers and a modest SLA can beat one with an impressive guarantee and poor measured performance.
- Does moving from shared hosting to a VPS improve uptime?
- Often, because it removes the most common shared-hosting failure: other sites on the same server consuming the resources yours needs. It does not fix downtime caused by your own plugins, code, or traffic outgrowing the plan, so pair the move with caching and a realistic resource tier rather than expecting the VPS alone to solve everything.