Lab tests run on a clean machine with a perfect connection. Your users don't. Hoppla pulls your Core Web Vitals from Google's CrUX dataset: p75 LCP, INP and CLS, measured from real Chrome traffic over a real time window. Same field data Google ranks on, read straight off the source, sitting next to the uptime and SSL checks on the same monitor.
crux example.com p75 · LCP 2.9s over-good · INP 184ms ok · CLS 0.06 ok · sample 28d · WARNING
Fictional values. Not a live measurement.
Hoppla queries the Google CrUX API for your origin and reads the p75 of every Core Web Vital from real Chrome visits: LCP, INP and CLS, plus FCP and TTFB for context. Each core vital gets graded against Google's own thresholds. Once LCP passes 2500ms, INP passes 200ms, or CLS passes 0.1 at the 75th percentile, the check turns to warning. This is field data, not a synthetic run. It's what your slowest quarter of real users sat through, with the sample window stamped on it.
On each run, Hoppla calls the Google CrUX API, the Chrome User Experience Report, for your origin. CrUX aggregates real loads from real Chrome users who opted into reporting. So the response is your audience's actual experience, not a probe pretending to be a phone on a clean line.
Hoppla pulls the 75th-percentile figure for LCP, INP and CLS, the metric Google itself uses, plus FCP and TTFB for diagnosis. p75 means three of four real visits were at least this good. The slowest quarter is where the pain actually lives.
Each core vital is scored on Google's published lines: LCP good at 2500ms or under and poor past 4000ms, INP good at 200ms and poor past 500ms, CLS good at 0.1 and poor past 0.25. Any of the three past its good threshold and the check returns a warning. All three in good, it's OK.
Every result with data carries the CrUX sample date range, so you know which weeks of traffic the number describes. Origins with too little Chrome traffic to report come back OK with a no_data flag. Hoppla tells you it can't see you yet instead of inventing a green light.
A synthetic test on a fast machine reports the load it had, not the one your users get on a mid-range Android over hotel wifi. CrUX p75 is the slowest quarter of your real audience. That gap is exactly the slice Google grades, and the slice that bounces.
Google uses field Core Web Vitals as a page-experience signal, and CrUX is where that field data comes from. So a creeping LCP isn't just slow. It can cost you rank before any analytics dashboard connects the dots.
A poor CLS is a fix-this-sprint problem, not an outage to wake up for. Web Vitals stays warning-only. It surfaces in your dashboard, but it never opens a downtime incident and never fires on-call. The pager is reserved for things that are actually down.
You ran PageSpeed once, saw a green 95, moved on. Meanwhile your real visitors are getting a 3-second LCP, and the lab score never showed it. Hoppla reads your actual CrUX p75 next to your uptime check. The first time your field LCP slips past good, it's a dashboard warning you can act on, not an unexplained slide in organic traffic you only notice a quarter too late.
A client asks whether the performance work you billed for actually did anything. Lab numbers won't settle it. They were green before you started. With CrUX field data you point at the real-user p75 LCP from before the fix, then at the number once the sample window rolls forward, and show it dropped under 2500ms for the people who actually visit. That's a screenshot a client believes, on a metric you didn't get to cherry-pick.
Nobody is opening PageSpeed for three hundred clients each month. Each monitor can carry its own Web Vitals check, so you can see across the whole portfolio which origins are warning on LCP, INP or CLS, and which sit clean in good. Account managers walk into the quarterly review with real-user numbers per client, not a one-off lab run, and you catch the site drifting toward a poor INP before the client's traffic does.
Real-user performance reporting is an easy upsell, and it runs on data you don't have to collect. Hoppla is white-label, so the CrUX p75 figures, the good/poor grading and the sample window all surface under your brand and your domain. Your customers see real Chrome-user numbers on their own sites and never learn Hoppla is underneath, a new tier of value on the platform you already resell uptime and SSL on.
Import the monitors you already run in another tool. What comes across depends on the provider. Run both side by side for 24 hours or 7 days before you switch.
The importer opens in your workspace once you have access.
Get early accessPageSpeed Insights is a one-off check you run by hand, and DebugBear leads with lab results, a synthetic load on a controlled machine. A lab run tells you what a clean load looks like, not what your users get. Hoppla reads the field side: CrUX p75 from real Chrome traffic, the data Google actually ranks on. It runs on a schedule next to your uptime and SSL checks, so the real-user number is tracked instead of spot-checked when you remember. For lab-side waterfalls and synthetic traces, those tools are still the right call. This is the field-data half, on the monitor you already watch.
No. Web Vitals is advisory and warning-only. It shows up in your dashboard, but it never opens a downtime incident and never triggers on-call escalation. A slow LCP is a fix-it-this-sprint problem, not a wake-up-the-team one. The pager stays reserved for the site actually being down.
No. CrUX only reports an origin once it has enough real Chrome traffic to aggregate anonymously. Below that floor, Hoppla returns OK with a no_data flag rather than pretending everything's green or throwing a false warning. As your traffic grows, CrUX starts reporting and the vitals fill in on their own.
Because Google uses p75, and averages hide the users who hurt. The 75th percentile means three of four real visits were at least this fast. It deliberately captures the slow quarter on weaker devices and worse connections. That's the slice Google grades your page experience on and the slice most likely to bounce, so it's the honest number to watch.
Checks, alerts and status pages in one workspace. Now in early access: join the list for an invite.