How to Verify Uptime Claims Like 99.9%

A Destin vacation rental company signed with a property management software vendor who promised "99.9% uptime." Eight months later, their system went down for 11 hours during peak booking season. The vendor's response: "That's within our contractual definitions."

The contract said 99.9% measured monthly, excluding scheduled maintenance and "third-party dependencies." When you subtracted those carve-outs, the vendor had actually delivered about 99.4% uptime—which they claimed was within acceptable variance.

The rental company lost $34,000 in bookings they couldn't process. The SLA was useless.

What 99.9% actually means

Let's do the math:

• 99.9% uptime = 8.76 hours of downtime per year
• 99.99% uptime = 52 minutes of downtime per year
• 99.5% uptime = 1.83 days of downtime per year

Most vendors advertise 99.9%. That's nearly 9 hours of acceptable downtime annually.

For a business that operates 24/7, those 9 hours will happen. Probably during your busiest time.

The SLA problem

A Service Level Agreement (SLA) is the vendor's promise about uptime, written in a way that benefits them.

The first problem: what counts as downtime?

Read carefully. "Downtime" might mean:

• Complete system unavailability
• Only the primary system, not integrations
• Only during business hours
• Excluding "scheduled maintenance" (which they control)
• Excluding "third-party services" (which you depend on)

A vendor can be "up" by their definition while your business is paralyzed.

The second problem: how is uptime measured?

Monthly measurement is common. If they have a bad month, they can recover. Annual measurement is stricter.

Measurement methodology matters. Do they measure from their monitoring, or from your reported issues? Disputes happen.

The third problem: what happens when they miss it?

Most SLAs offer credits. "If we fall below 99.9%, you'll receive a credit equal to 5% of your monthly fee."

If you pay $500/month and they miss a SLA, you get a $25 credit. Meanwhile, you lost $30,000 in bookings. The credit doesn't cover your losses.

Read the SLA. Check the remedy. If the remedy is a small credit, the SLA is marketing, not protection.

How to actually verify reliability

Ask for their track record

"Can you share your uptime from the past 12 months?"

Reputable vendors track this. They'll share aggregate numbers. If they can't or won't, that's information.

Check third-party monitoring. Services like Pingdom, statuspal.io, or Downdetector show real-world uptime reported by users.

Ask about their architecture

Redundant systems fail less than single servers. Ask:

• Do you have redundant servers in multiple locations?
• What happens when a data center goes down?
• How are database failures handled?
• What's your backup and recovery approach?

A vendor with real redundancy can explain it clearly. A vendor with a single server will change the subject.

Ask about their worst incident

"How did you handle the worst outage in the past year? What did you learn?"

Good vendors have post-mortems. They'll explain what happened, how long it took to fix, and what they changed. They won't have an answer if they've never had a significant problem.

Bad vendors say "we don't discuss incidents with clients" or change the subject.

Ask about communication

When things break, what happens?

• Do you get proactive notifications?
• Is there a status page?
• Can you see incident history?

A vendor with a good status page is a vendor who takes communication seriously. A vendor who sends vague emails after the fact is a vendor you won't hear from until you're already upset.

Test with your own monitoring

Before relying on a system for critical work, run your own checks. Use a free service like UptimeRobot to monitor from multiple locations. This won't tell you everything, but it will tell you when they're lying about uptime.

What it costs

When downtime happens to you:

For a 20-person business:

• Hour of email down: $500-$2,000 in lost productivity
• Hour of CRM down: $1,000-$5,000 in lost sales opportunities
• Day of e-commerce down: $5,000-$50,000 in lost sales
• Hour of manufacturing systems down: $10,000-$100,000 in lost production

A 99.9% SLA that delivers 99.5% costs you roughly 8 additional hours of downtime per year. Run those numbers against your business.

When vendors claim things they can't deliver:

The gap between what vendors promise and what they deliver is real. The average "provider-reported uptime" vs. "customer-experienced uptime" gap is about 0.3-0.5%.

What can go wrong

You assumed the SLA was protection. It was marketing. Your downtime wasn't covered by the definitions.

You didn't ask about maintenance windows. They do weekly "scheduled maintenance" on Sunday 2-4 AM. Fine—unless you're a 24/7 business.

You trusted their status page. Some vendors show green lights while customers experience red problems. The status page reflects what's working, not what's available.

You didn't test the recovery path. When things break, do you know how to recover? Do they have documented procedures? Or is everyone figuring it out in real time?

Vendor questions (copy/paste)

Ask every vendor:

• "What's your actual uptime from the past 12 months, measured how?"
• "Walk me through what counts as downtime in your SLA definition."
• "What happened in your worst outage this year, and how did you handle it?"
• "When there's an incident, how do you communicate with affected customers?"
• "What does the SLA credit cover if you miss your targets? Is that documented?"
• "Do you have a public status page with historical uptime data?"
• "How are maintenance windows handled, and are they communicated in advance?"

When to hire help

Hire someone to evaluate reliability claims when:

• Your business can't tolerate more than a few hours of downtime per year
• The system handles money, medical records, or other high-stakes data
• You're considering a vendor with a complicated architecture
• You've experienced downtime-related losses before

A technical review of a vendor's infrastructure costs $1,000-$5,000. It's worth it for a system that affects $100,000+ in annual revenue.

The honest answer

No vendor delivers perfect uptime. The question isn't whether they'll have problems. It's how they handle problems when they happen.

A vendor who admits to past incidents, explains their response, and has clear communication during problems is worth more than a vendor who claims 100% uptime and goes silent when things break.

That Destin vacation rental company? They switched vendors. Their new vendor had "only" 99.8% uptime last year—but they communicated during every incident, provided real-time status updates, and proactively offered credits. The rental company could plan around it. That was worth more than a 99.9% promise they couldn't keep.