# Synthetic Monitoring: Overview

Synthetic Monitoring runs Lighthouse speed tests on the pages you add, on a schedule. Each test loads the page the same way every time, so you can compare results fairly and see when a page got slower.

## What a lab test is [#what-a-lab-test-is]

A lab test is a speed test that a test browser runs, not a real visitor. Synthetic Monitoring uses Google Lighthouse. The test browser loads each page on a pretend slow phone and a slow connection, the same way every time. Many visitors are on faster connections than the test.

Because the conditions do not change, lab results are easy to compare. A lab test is the most reliable way to find a slowdown right after a release, before visitors or Google notice it. A lab test does not need traffic, so it also works for new pages and pages with few visits.

Lab tests are different from field data. Field data comes from real visits: [CrUX Monitoring](/crux-monitoring/overview) shows Google's data from Chrome visits, and [RUM Monitoring](/rum-monitoring/overview) shows your own visitors.

## Devices, locations and schedule [#devices-locations-and-schedule]

### Devices [#devices]

Synthetic Monitoring tests each page on two devices:

- **Phone**: a slow phone on a slow connection.
- **Desktop**: a desktop computer.

Most cards show **Phone** and **Desktop** side by side. The verdict and the default order of the pages use the phone test, because it is the harder test.

### Locations [#locations]

A location is the place that a test runs from. **Google's Servers** use the Google PageSpeed Insights service. Other locations are test machines of VitalSentinel. The Free plan tests from **Google's Servers** only. How many locations one page can use depends on your plan. See [Plans and limits](/workspace/plans-and-limits).

### Schedule [#schedule]

Each page has its own schedule. The options are **Every 8 hours**, **Every 12 hours** and **Daily**. Which options you can select depends on your plan. The first test runs soon after you add a page. To change the schedule of a page, see [Pages](/synthetic-monitoring/pages#edit-the-schedule-and-locations-of-a-page).

## Before you start [#before-you-start]

- Synthetic Monitoring needs no script on your site. Add the pages you want to test.
- To open it, select **Synthetic Monitoring** in the **Web performance** group of the domain sidebar.
- Workspace owners and admins can add, edit and remove pages. Members see the results read-only.
- The number of pages you can test depends on your plan. See [Plans and limits](/workspace/plans-and-limits).

## The Synthetic Monitoring Overview [#the-synthetic-monitoring-overview]

The Overview shows, from top to bottom:

1. The verdict: one sentence about the phone test scores of all your pages.
2. The Site Health banner, if Site Health Monitoring found Synthetic Monitoring issues.
3. The **Lab results over time** card.
4. The **Lab versus real world** card.
5. The **Worth testing** suggestions and the plan limit notice (owners and admins only).
6. The **Pages** card. See [Pages](/synthetic-monitoring/pages).

The filter bar holds the date range, the comparison and, when more than one location has results, the **Location**. For the filter bar, see [Using the app](/getting-started/using-the-app).

### The date range and the location [#the-date-range-and-the-location]

The Synthetic Monitoring filter bar offers **Last 7 days**, **Last 28 days** and **Last 90 days**. The default is **Last 28 days**. You can compare with an earlier period.

If your pages have results from more than one location, the filter bar shows **Location**. The default is **Google's Servers**. A lab score is only comparable inside one location, so there is no option for all locations together.

## The verdict [#the-verdict]

The verdict at the top of the Synthetic Monitoring Overview judges the average phone score of all tested pages. Lighthouse sorts a score into three bands:

| Band | Score |
|------|-------|
| **Good** | 90 to 100 |
| **Needs improvement** | 50 to 89 |
| **Poor** | under 50 |

The same bands set the status of Synthetic Monitoring on the domain [Overview](/domain/overview), in the sidebar and on the [Dashboard](/workspace/dashboard). A score from 50 to 89 shows **Needs attention**. A score under 50 shows **Critical**. With a good score, the card can still show **Minor issues**, for example for a slow LCP in the lab test.

The headline says how your pages load in the phone test, for example **Your pages load slower than they should in the phone test: 3 of 12 pages score under 50**. If you compare with an earlier period, the line under it says if the phone score went up or down. Hover the **?** to read how VitalSentinel judges a lab test.

## Lab results over time [#lab-results-over-time]

The **Lab results over time** card of Synthetic Monitoring shows the average results of all your pages. It has four tabs:

| Tab | What it shows | Good | Poor |
|-----|---------------|------|------|
| **Performance score** | The Lighthouse score, 0 to 100 | 90 or more | under 50 |
| **LCP** | When the main content appears | under 2.5 s | over 4 s |
| **TTFB** | How fast the server sends the first byte | under 600 ms | over 1.8 s |
| **CLS** | How much the page jumps while it loads | under 0.1 | over 0.25 |

Each tab shows the **Phone** and **Desktop** figure. The score is the average of the tests in the date range. For LCP, TTFB and CLS, VitalSentinel takes the value that 3 out of 4 tests of each page reach or do better than. Then it averages your pages.

The chart shows one line for **Phone** and one for **Desktop**. If you compare periods, the earlier period shows as a dashed line. If some pages have no result yet, the subtitle says so, for example **1 of 5 pages tested so far**. A tab shows **Waiting for the first test** until the first test finishes.

## Lab versus real world [#lab-versus-real-world]

The **Lab versus real world** card of Synthetic Monitoring compares the LCP of the lab test with the LCP of real visits. It shows the ten slowest pages of the site. Each row has two dots:

- The first dot is the **Phone lab test**.
- The second dot is Google's data from [CrUX Monitoring](/crux-monitoring/overview). If Google has no data for the page, the second dot is real visitors from [RUM Monitoring](/rum-monitoring/overview).

The row says how far apart the two are, for example **481 ms slower**, or **About the same**. A big gap means that the real world differs from the test. Look first at pages that are slower in the real world than in the lab.

If none of the ten slowest pages has both a lab test in the date range and real-world data, the card shows **Nothing to compare yet**. Connect RUM Monitoring or track the pages in CrUX Monitoring to fill it.

## Page states [#page-states]

If you test no pages, the Synthetic Monitoring Overview shows **Add your pages to test their speed on a schedule**. Owners and admins click **Add pages** to start. See [Pages](/synthetic-monitoring/pages#add-pages).

If the page list cannot load, the Overview shows an error with a retry button.

## The pages of Synthetic Monitoring [#the-pages-of-synthetic-monitoring]

- [Pages](/synthetic-monitoring/pages): the **Pages** card, and how to add, edit and remove pages.
- [Test details](/synthetic-monitoring/test-details): the results of one page, with what to fix and how it loaded.

## Troubleshooting [#troubleshooting]

### Why do my scores differ from PageSpeed Insights? [#why-do-my-scores-differ-from-pagespeed-insights]

Each lab test is one run, and results change a bit from run to run. Server speed, third-party scripts and changing content all have an effect. A difference of some points between two runs is normal. Look at the trend over many tests, not at one test.

### Why is the lab test slower than my real visitors? [#why-is-the-lab-test-slower-than-my-real-visitors]

The phone test uses a slow phone on a slow connection. Many visitors have faster devices and connections. If the lab is slower than the real world, this is normal. If the real world is slower than the lab, find the cause in [RUM Monitoring](/rum-monitoring/overview).

### How often do tests run? [#how-often-do-tests-run]

Each page runs on its own schedule: **Every 8 hours**, **Every 12 hours** or **Daily**. The options you can select depend on your plan. To see the schedule of a page, open it and look at the chip next to the title.

### Can I run a test now? [#can-i-run-a-test-now]

No. Synthetic Monitoring runs tests only on the schedule of each page. The first test runs soon after you add a page.

## Related [#related]

- [Pages](/synthetic-monitoring/pages) - add, edit and remove tested pages
- [Test details](/synthetic-monitoring/test-details) - what to fix on one page
- [CrUX Monitoring](/crux-monitoring/overview) - Google's field data for your pages
- [RUM Monitoring](/rum-monitoring/overview) - the speed of all your own visitors
- [Metrics and thresholds](/help/metrics-and-thresholds) - the limits of every measure