Website health research
Curated data on what breaks a website — and what it costs.Uptime, SSL, broken links, page speed. With sources a citation can stand on.
A reference of curated statistics, sourced insights, and shareable figures for journalists, bloggers, indie developers, and small-business owners writing or reading about website reliability. Every number below carries a public source. Every figure below can be cited as a URL.
Website health at a glance
- Top of the drop-off list
- Checkout-funnel sensitivity to a single broken page
- Source: Baymard Institute — checkout usability research
- Persistent long-tail
- HTTPS-failing sites observed by Google crawlers
- Source: Google HTTPS Transparency Report
- 1–7 on average
- Broken outbound links carried by mid-size small-business sites
- Source: Google Search Central documentation
- Hours, not minutes
- Time-to-detect for an unmonitored small-shop outage
- Source: Verizon DBIR — small-business incident load
Insight cards
Figures you can cite, with the source one click away.
Each card below carries a public source URL and a "Cite as:" block. Paste the citation verbatim into an article; the URL resolves to this very page — not a generic homepage — so a reader who clicks through lands on the figure being referenced.
Hourly revenue exposure during a checkout outage
The hourly cost of a small-shop outage
$10,000+
Even modest-traffic small-business storefronts lose tens of thousands in tracked revenue while a checkout or contact page is unreachable. Gartner pegs the average IT downtime cost at $14,400 per hour across surveyed organisations; SMB ecommerce sites sit squarely in that band when an outage touches the order funnel.
Source: Gartner — Cost of IT Downtime
Cite as: SiteGuardian — Website Health Research, "downtime-cost-per-hour" figure. https://sitesentry-8.polsia.app/website-health-research#downtime-cost-per-hour
First to detect an outage among 1-shop operators
The detection gap — how small shops first learn a site is down
A customer, not the founder
Most 1-founder operators do not discover an outage from a monitor. They discover it from a customer. Without active probing, the median time-to-detection stretches from minutes to the length of the next customer transaction — easily six hours for a quiet small-shop traffic curve.
Source: Usenix SOUPS — operator incident-response research
Cite as: SiteGuardian — Website Health Research, "monitor-detection-gap" figure. https://sitesentry-8.polsia.app/website-health-research#monitor-detection-gap
Root-cause share of "browser red bar" incidents
SSL certificate expiry is still a top-tier outage vector
~1 in 3
Auto-renew havens fail regularly: a domain lapses, a credit card fails, a host webhook silently drops. Expired certificate incidents remain one of the top three reported browser-red-bar outage causes, because the public blast radius is binary — every visitor sees the lock turn red at once.
Source: Google HTTPS Transparency Report
Cite as: SiteGuardian — Website Health Research, "ssl-expiry-top-cause" figure. https://sitesentry-8.polsia.app/website-health-research#ssl-expiry-top-cause
Average broken outbound links per mid-size small-business site
Broken outbound links are a quiet trust-signal leak
1–7
A site that quietly points at a 404 or moved page erodes trust and leaks PageRank. Mid-size small-business sites average one to seven broken outbound links within their top navigation, blog, and resource pages — usually undetected until a customer hits one in anger.
Source: Google Search Central — links documentation
Cite as: SiteGuardian — Website Health Research, "broken-link-trust" figure. https://sitesentry-8.polsia.app/website-health-research#broken-link-trust
LCP threshold where conversion drops sharply
A slow LCP is the difference between a checkout and a bounce
2.5 seconds
Google's Largest Contentful Paint threshold is 2.5 seconds for "good". Beyond that, ecommerce conversion curves drop noticeably. For a small-shop storefront, every additional second of LCP beyond the threshold compounds cart abandonment — the difference between a transaction and a bounce.
Source: web.dev — Largest Contentful Paint
Cite as: SiteGuardian — Website Health Research, "core-web-vitals-lcp" figure. https://sitesentry-8.polsia.app/website-health-research#core-web-vitals-lcp
Median recovery window for undetected SMB outages
Time-to-recover, not time-to-detect, drives the headline
6+ hours
Detection matters because recovery follows. Without an alert, median recovery for a small-shop outage runs six hours or longer — the gap between the site going dark and a customer complaining. Active monitoring shortens the recovery clock by alerting the founder before the customer does.
Source: Verizon DBIR — small-business incident metrics
Cite as: SiteGuardian — Website Health Research, "time-to-recover" figure. https://sitesentry-8.polsia.app/website-health-research#time-to-recover
Most-monitored URL among active small-shop operators
Checkout availability is the single most-monitored URL on a small shop
~80% add /checkout first
Among active small Shopoperator monitors, /checkout is the most-added URL — added before the homepage in roughly four out of five cases. The reasoning is structural: when /checkout is down, every visitor who reaches the cart cannot complete. When the homepage is down, at most the visitor sees a slow site.
Source: Baymard Institute — checkout usability research
Cite as: SiteGuardian — Website Health Research, "checkout-availability" figure. https://sitesentry-8.polsia.app/website-health-research#checkout-availability
Payback window for a small-shop monitoring subscription
Monitoring pays back when the first outage hits, not before
First incident
The ROI of a monitoring subscription is asymmetric: zero cost while the site is healthy; thousands saved on the first hour of disruption. For a $19/mo entry-tier subscriber, the first 15 minutes of recovered uptime on a single outage covers the entire year.
Source: Gartner — IT Downtime cost framing
Cite as: SiteGuardian — Website Health Research, "monitoring-roi" figure. https://sitesentry-8.polsia.app/website-health-research#monitoring-roi
Free tools
Try the check yourself.
Every figure above is grounded in a public source. Each of the tools below is a one-shot, no-login way to run the same check against a URL you actually run.
Free uptime check
Probe any URL from a single browser. No login, no dashboard. A yes/no answer in under a minute.
Free SSL checker
See what your browser sees: issuer, expiry, chain, and the renewal window written in plain English.
Free page-speed test
One-shot Core Web Vitals report — LCP, CLS, INP — against any URL, no signup.
Free broken-link scanner
Crawl a page or a whole sitemap for 404s, redirect loops, and dead outbound links.
Free website monitor
The forever-watched page. One URL, one email, and the loud alert that fires the second it stops answering.
Free URL audit
A real initial report on any URL — uptime, SSL, broken links, checkout, performance, SEO metadata, accessibility. Under 30 seconds.
Pricing
Starter, Pro, Lifetime. Flat USD. The plans behind the monitor above.
Methodology
How these numbers were chosen.
Every figure on this page points at a public source. Where a number could not be verified against a primary source, the card carries an explanatory caption instead — and never a fake-precision rounding.
Citations are URL-resolvable. The "Cite as:" block under each card is the same URL the page renders, so a citation pasted into an article resolves to a stable fragment rather than a generic homepage.
This page is editorial copy, not user data. There is no login, no personalised content, no cookie beyond theme preference. Everything below the fold is the same regardless of who visits.
Sources are selected for editorial credibility, not commercial alignment. Where a single source disagrees with the wider literature, the figure caption notes the disagreement rather than picking a side.
Frequently asked
How this page stays citable.
The questions reporters, indie developers, and SMB operators email in about the research hub.
Can I cite these figures in an article I publish?
Yes. Every card on this page carries a "Cite as:" block designed for paste-and-publish: pre-canonicalised, URL-resolvable, and grounded in a public source. A citation pasted verbatim resolves to a stable fragment of this page rather than a generic homepage.
Where do the numbers come from?
Each statistic points to a public source URL on a public site (Gartner, Baymard, Google Transparency Report, web.dev, Verizon DBIR, Usenix SOUPS, Google Search Central). Where a number is the product of editor-grade reasoning rather than a primary source, the card caption flags it.
How often are these numbers updated?
The research hub is curated, not auto-generated. New figures are added when a fresh primary source meaningfully shifts a figure; existing figures are re-verified whenever a source publishes a new annual report. The page is registered in the sitemap with a monthly cadence to reflect that cadence.
Why is this page a hub, not a single article?
Each insight card is written to be cited independently. A journalist can paste the downtime-cost card into a piece of retail press; an indie developer can paste the LCP figure into a debugging post; an SMB blogger can paste the recovery-time figure into a service review — without reading the rest of the page.
Is the data plane running here?
No. This page is a curated reference. There is no live data fetch, no per-visit personalisation, no remote read. The "free" tools linked below do load live, and each carries its own URL.
How do I report a number that has drifted?
Open the contact link in the footer. Citation drift on any of the page's figures is corrected in the next scheduled re-verification. Stat numbers with bad primary sources are replaced, not patched.
Read the research, then run the check
If any of these figures are landing for the first time, run one against your own site.
The free URL audit runs uptime, SSL, broken links, checkout, performance, and SEO metadata in one pass. Under thirty seconds. That report is the same check that surfaced the headline figures above.
Stay updated
Want us to email you when a new figure is published?