# RUM Monitoring: Engagement

The **Engagement** page of RUM Monitoring answers one question: do slow pages cost you readers, clicks and sales? It compares what visitors do on a page at a Good, Needs improvement and Poor speed.

## How the Engagement page works [#how-the-engagement-page-works]

The RUM script records what visitors do on a page. It records how far they scroll, how they click, how they use forms and how long they stay. RUM Monitoring puts each visit into a speed band, from the speed measure that you select. Then it compares the behavior of visitors in each band.

If visitors on slow pages scroll less, click in frustration or leave forms more often, slowness costs you visitors. Use this to decide which speed problem to fix first.

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

- RUM Monitoring needs the RUM script on your pages, with engagement tracking on. See [Install the RUM script](/connect/rum-script).
- In the sidebar, open **Web performance** > **RUM Monitoring** > **Engagement**.
- Select **Phone** or **Desktop** in the filter dock. The Engagement page shows the selected device only.

You can also filter by date, **URL Path**, **Country** and **Browser**, and compare with an earlier period. See [Filters and the device](/rum-monitoring/overview#filters-and-the-device).

### Visitors who ask not to be tracked [#visitors-who-ask-not-to-be-tracked]

The RUM script records no engagement for visitors whose browser sends a Do Not Track or Global Privacy Control signal. Their page views, speed and errors still count. Rates such as **Rage click rate** count these visits in the total, but never as an event. Averages such as **Time on page** use only the visits that report the measure.

## The Engagement verdict [#the-engagement-verdict]

The verdict is one sentence at the top of the Engagement page. It shows only when slowness hurts the measure that you selected. For example: "On Phone, a slow LCP hurts read-to-end rate".

The two figures next to the sentence are the value at a Good speed and at a Poor speed. The verdict needs at least 100 visits in both the Good and the Poor band. The gap must also be at least 10% of the Good value. If speed makes no clear difference, the page shows no verdict.

If Site Health Monitoring found issues or opportunities about RUM Monitoring data, a banner shows their number below the verdict. Click **Open in Site Health Monitoring** to see them.

## Tabs and controls [#tabs-and-controls]

The Engagement page has four tabs: **Reading**, **Clicks**, **Forms** and **Attention**. Each tab has a group of measures. The card at the top of each tab has two controls on the right:

- **Engagement measure:** the measure to show, for example **Read to end** or **Rage clicks**. **Reading** and **Attention** use a dropdown, because they have many measures.
- **Speed measure:** **LCP**, **CLS**, **INP**, **FCP** or **TTFB**. RUM Monitoring puts each visit in a speed band by this metric. The default is LCP.

The bands use the normal limits of the selected speed measure. See [Metrics and thresholds](/help/metrics-and-thresholds).

## The measure card [#the-measure-card]

The measure card has the name of the measure and the device, for example **Read-to-end rate on Phone**. It has three tiles:

- **Good LCP:** the value of the measure for visits with a Good speed.
- **Needs improvement LCP:** the value for visits with a speed that needs improvement.
- **Poor LCP:** the value for visits with a Poor speed.

The tile names follow the speed measure that you select. Each tile shows the number of visits in its band, and the change when compare is on. A tile shows **No visits at this speed** when the band has no visits. It shows **Not measured at this speed** when no visit in the band reported the measure.

## The Visits by chart [#the-visits-by-chart]

The **Visits by LCP** card (or the speed measure that you select) shows a chart. Each bar is the visits in one speed range, colored by its band: **Good**, **Needs improvement** or **Poor**. The line is the value of the engagement measure for those visits.

Hover a bar to see the speed range, the visits and the value of the measure. With compare on, a second line shows the earlier period.

If the line gets worse from left to right, slower visits engage less. Fix the speed range where the line drops.

## Reading tab [#reading-tab]

The **Reading** tab shows how far visitors scroll down a page. A page that few visitors read to the end can be too long, or too slow.

| Measure | Name in the card | What it counts |
|---|---|---|
| **25%** | **25% scroll rate** | The share of visits that scrolled at least a quarter of the way down |
| **50%** | **50% scroll rate** | The share of visits that scrolled at least halfway down |
| **75%** | **75% scroll rate** | The share of visits that scrolled at least three quarters of the way down |
| **90%** | **90% scroll rate** | The share of visits that scrolled almost to the end |
| **Read to end** | **Read-to-end rate** | The share of visits that scrolled to the very end. This is the default. |

## Clicks tab [#clicks-tab]

The **Clicks** tab shows how much visitors click, and where they click again and again because nothing happens.

| Measure | Name in the card | What it counts |
|---|---|---|
| **Clicks per visit** | **Clicks per visit** | How many clicks a visit makes on average. This is the default. |
| **Rage clicks** | **Rage click rate** | The share of visits with rage clicks |

A rage click is several fast clicks on one spot: 3 or more clicks on the same element within 2 seconds. It is a sign that something does not respond. More rage clicks on slow pages mean that visitors give up waiting.

### Where rage clicks happen [#where-rage-clicks-happen]

When you select **Rage clicks**, the **Where rage clicks happen** card shows below the chart. It lists the pages and elements with the most rage clicks in the period.

| Column | What it shows |
|---|---|
| **Page** | The page where the rage clicks happened |
| **Element** | The element, as the script stored it: tag, id and classes |
| **Rage clicks** | The number of rage clicks on that element |
| **Visits** | The visits that had a rage click on that element |

The card never shows the text on the page. It shows 10 spots on each page of results and has no compare. If there were none, it says "No rage clicks in this period".

Fix the spots at the top of the list first. Check that the element is a working link or button, and that it reacts fast.

## Forms tab [#forms-tab]

The **Forms** tab shows forms that visitors start, leave or send. A form that many visitors leave halfway loses you sign-ups or sales.

| Measure | Name in the card | What it counts |
|---|---|---|
| **Abandonment** | **Form abandonment rate** | The share of visits that started a form and left without sending it. This is the default. |
| **Completion** | **Form completion rate** | The share of visits that sent a form, out of the visits that started or sent one |
| **Time to complete** | **Time to complete a form** | The typical time from the first field to the form sent, for the visits that sent one |

**Time to complete a form** is the median: half of the visits were faster. A higher abandonment at a Poor speed means that slowness costs you sign-ups.

## Attention tab [#attention-tab]

The **Attention** tab shows how long visitors stay, how active they are and how many leave at once.

| Measure | Name in the card | What it counts |
|---|---|---|
| **Time on page** | **Time on page** | How long visitors stay on the page. This is the default. |
| **Attention score** | **Attention score** | How actively visitors engaged with the page |
| **Active time** | **Active time** | How long the page was open and in front of the visitor |
| **Tab switches** | **Tab switches per visit** | How often a visitor switched to another tab and back |
| **First interaction** | **Time to first interaction** | How long visitors wait before their first tap, click or key press |
| **Quick back** | **Quick back rate** | The share of visits that left within 5 seconds |

All measures except **Quick back** are averages over the visits that reported them. A visit that leaves within 5 seconds is attention that never started.

## Web Analytics link [#web-analytics-link]

At the bottom of the Engagement page, a box links to Web Analytics. It says "Web Analytics has the conversions and revenue of these visits." Click **Open Web Analytics** to see what slow pages cost in sales. If Web Analytics is not installed, the box shows **Install the Web Analytics script**. See [Web Analytics](/web-analytics/overview).

## Troubleshooting [#troubleshooting]

### Why does the Engagement page show no verdict? [#why-does-the-engagement-page-show-no-verdict]

The verdict shows only when slowness clearly hurts the measure. It needs at least 100 visits at a Good speed and at a Poor speed. If one band has fewer visits, or the gap is small, there is no verdict. Select a longer date range, or another speed measure.

### Why does a tile say "No visits at this speed"? [#why-does-a-tile-say-no-visits-at-this-speed]

No visit in the period had a speed in that band, for the selected speed measure and device. This is often true for the **Poor** band on a fast site.

### Why does a tile say "Not measured at this speed"? [#why-does-a-tile-say-not-measured-at-this-speed]

The band has visits, but none of them reported the measure. For example, no visit in the band started a form. Select another measure, or a longer date range.

### Why are the engagement numbers lower than I expected? [#why-are-the-engagement-numbers-lower-than-i-expected]

The RUM script records no engagement for visitors who send a Do Not Track or Global Privacy Control signal. Engagement tracking can also be off in the script settings. See [Install the RUM script](/connect/rum-script).

### Why does the Engagement page say "No visits on Phone in this period"? [#why-does-the-engagement-page-say-no-visits-on-phone-in-this-period]

There were no visits on the selected device in these dates and filters. Select **Desktop** in the dock, select a longer date range, or remove a filter.

## Related [#related]

- [Overview](/rum-monitoring/overview) - all Core Web Vitals, filters and page-view limits
- [INP](/rum-monitoring/inp) - slow reactions to taps often cause rage clicks
- [Web Analytics](/web-analytics/overview) - conversions and revenue of the same visits
- [Performance & engagement](/google-search-console/performance-and-engagement) - if slow pages get fewer Google clicks
- [Install the RUM script](/connect/rum-script) - engagement tracking settings