RUM Explorer

Every AEM Site, including Adobe Experience Manager Sites as a Cloud Service without Edge Delivery Services, collects RUM data by default. This data is sampled and anonymous, and used by Adobe to diagnose issues around performance, traffic acquisition, and conversion for customer sites.

With RUM Explorer, Adobe offers a simple user interface to explore the collected data for a given domain. It is a simple data visualization tool for authors, site owners, developers, and operations teams to understand how their site is doing and what can be improved.

Insights in RUM Explorer are meant to be shared, therefore for each view of RUM Explorer, you can copy, paste, and share the URL from the browser to invite team members to look over the data.

All RUM data is sampled, which means each data point represents hundreds, or even thousands of page views. This sampling means that data is always anonymous and analysis is fast, but it also means that, like in every measurement system, there is a margin of error. Hover the mouse over a number displayed to see the margin of error. Sampling also means that the uncertainty in the data rises when you filter out more of the data.

User Interface Overview

RUM Explorer is available at https://www.aem.live/tools/oversight/explorer.html and you need a domain key to access the data for most domains. Data for the website emigrationbrewing.com, a small but great Salt Lake City-based restaurant is open for everyone, so you can explore it under https://www.aem.live/tools/oversight/explorer.html?domain=emigrationbrewing.com&view=month&domainkey=open

The user interface is broken down in three parts:

1. Key metrics on the top left. These key metrics are split into business metrics (page views, visits, engagement) and performance metrics (Largest Contentful Paint, Cumulative Layout Shift, Interaction to Next Paint) and will update whenever you change the filters.
2. A chart on the bottom left that provides an at-a glance visualization of your data. The default visualization is to show page views and performance metrics over time.
3. A set of facets that allow you to drill down into the data on the right. Each facet comes with checkboxes that allow you to filter the full data set to only include matching page views and three kinds of metrics around performance, traffic acquisition, and engagement.

Date Ranges

RUM Explorer allows you to look at the data over time, to identify recent, short-term and long-term trends. Under each of the key metrics at the top of the page you can find a small green or red line telling you at a glance if there is a positive or negative trend.

Weekly View

Viewing by Week is done by changing the dropdown to the right of the URL from Month to Week. Data in the chart is broken down by hour, which gives more precision than the default Month view which is broken down by day. Please be aware of the tradeoff of zooming in more closely - the benefits of having a smaller dataset for each hour can reveal an extrapolation-related drawback if there is not much data collected in a given hour.

Yearly View

It is also possible to zoom out and see data for a full year, broken down by week.

Additional Views

The default view explored so far is the timeline view. For some data, this is not the best visualization and alternative drill down views are offered. To enter a drill down view, locate a facet on the right side of the screen, click the little arrow pointing downwards, and you'll see data for that facet alone visualized.

When you've seen enough, click the arrow pointing upward to navigate back.

Device Share

In the Device Type and Operating System facet, there is a small down arrow box. Clicking this drill down updates the chart on the screen to become a pie chart which shows how different devices use the site. The inner pie represents the distribution of RUM collection among the major device type categories - desktop, mobile, and bot. The outer pie breaks it down further by the operating system category. Click the up arrow box in the facet header to return to the original chart view.

Top URLs

You can navigate to the top URLs bar chart from the URL facet in the sidebar. For each URL, you can see the distribution of performance metrics. Pages with good performance appear green, pages with poor performance red. For some pages there is not enough data, these are shown in gray.

Flow Diagram (Sankey)

The final visualization is the flow diagram (also called a Sankey diagram, after the inventor of this style of chart). It allows you to see how visitors enter a page (on the left), consume content and engage with it (in the middle) and leave the page, or stop interacting with it (on the right)

Focus Modes

Depending on whether you are an author, a marketer, a developer, or a designer, different aspects of the data matter to you. The focus mode selector at the top right of the facet sidebar allows you to select a view that fits your needs.

Performance Focus

The default focus is the performance focus and it shows Largest Contentful Paint (LCP), Cumulative Layout Shift (CLS) and Interaction to Next Paint (INP).

Green indicators mean that this performance metric is in a good state, yellow means it needs improvement, and red means performance is poor.

Do not read "needs improvement" as "it is not great, but acceptable", our data shows that anything that is not marked green leads to significant decreases of user experience, and therefore higher bounce rate and lower engagement.

For each of the focus views, a thick border around the facet metric means that there is a significant difference in that metric compared to the baseline (all page views). These are the areas that deserve further investigation.

Traffic Acquisition Focus

It does not matter how fast a site is if it can't attract and keep visitors. This is what the traffic acquisition focus is all about.

The core metrics for traffic acquisition are:

1. Bounce Rate (BOUNCE): this is the share of traffic that enters your site, and leaves immediately without clicking anywhere on the page. Accepting a consent banner isn't counted as a click for this metric. Generally this metric should be as low as possible, but any value below 50% is great.
2. Time on page (TIME): how long does a page view take on average? Longer times indicate that visitors are deeply engaged with your content, short times mean that the page isn't able to hold their attention
3. Organic Percentage (ORGANIC): great pages are able to attract traffic organically, i.e. without having to buy ads to acquire traffic. The organic percentage is the share of traffic that enters your page without following an ad. Internal navigation isn't considered when calculating organic share.

Conversion or Engagement

This focus mode can show one of two metrics:

• Engagement Rate (ENGAGEMENT) is the percentage of page views where visitors either consumed significant amounts of content or clicked. Engagement rates can be very high, often reaching up to 90%.
• Conversion Rate (CONVERSION) requires a custom conversion definition, which allows you to declare what kind of engagement is most valuable. This rate is often in the single digits, and a 5% rate can be good for many sites.

Exploring the Data

The fastest way to drill down into your data is to type a term into the search field above the facet sidebar. Any data point that has the search term either in the URL or in the checkpoints will be considered, the match will be highlighted and you can quickly see what has been found.

The facets offer a more structured approach to drilling down on the data. Each data point is classified across multiple dimensions, and you can include or exclude any category of page views depending on their facet values.

Device Type and Operating System

RUM does not track the user agents to preserve visitor privacy, but from the information available it can discern the broad device classes Mobile, Desktop, and Bot. Bot includes crawlers that access your site and execute JavaScript. The bot traffic you will see on RUM Explorer is smaller than what you would see in your CDN statistics, as most bots do not execute JavaScript and do not try to mimic a visitor. An important exception are search engines, which do execute JavaScript and are visible in RUM explorer.

For each broad device class, the operating system of the browser is broken out, too. This allows you to distinguish between macOS and Windows (and Linux, too – although some bots disguise themselves as Linux visitors) and between iOS and Android. You will notice that the iOS traffic does not report Core Web Vitals, which is a limitation of the browsers available on iOS.

If you click a checkbox for a device class, you will see all traffic in that device class, regardless of operating system. If you check a checkbox for an operating system, you will only see the traffic of that operating system. Selecting multiple checkboxes will show all traffic that satisfies any of the checked boxes.

Use the Device Type and Operating System Facet to determine whether the operating system has influence on experience quality and visitor behavior.

If you want to include a facet, click the checkbox and only data points that fulfill the criteria are included. To exclude a facet, hold the shift key while clicking the checkbox and that category will be excluded. In the screenshot above, all Bot traffic has been excluded from the overall count.

URL

The URL facet shows the different pages that make up your site. In some cases, when URLs contain long numbers, or other types of IDs, all paths following the same pattern will be grouped together to make it easier to identify patterns.

Checkpoints

Checkpoints represent the type of activity collected. Each different checkpoint includes additional data, which can be seen by checking the box next to a checkpoint. For example, checking the Visit Entry (enter) checkpoint will load an additional facet named External Referrer. Checking the Clicked (click) checkpoint will load two additional facets - Click Source which shows CSS selectors to help identify where on the page a user clicked, and Click Target which shows where the clicks take the user. Try out the various checkpoints and see what insights are available.

Checkpoint-Specific Facets

The following facets only appear when you've selected the corresponding checkpoint.

Facet: Viewmedia

You can see the CSS selector of the image (Media Source) and a preview of the image or video that has been seen (Media Target).

Facet: Viewblock

This facet allows you to identify the blocks that visitors have seen. The CSS selector includes the block name, and can include additional information such as dialog for elements that are fixed in the viewport and can't be scrolled away.

Facet: Navigate

This facet enables us to discover internal navigation paths. For external referrers, the enter checkpoint can be used.

Facet: Enter

For external referrers, often only the domain name, or in case of Android apps, the name of the Android application is visible. The value direct subsumes visitors entering the URL directly into their address bar, following browser bookmarks, or opening the page from iOS applications.

Facet: LCP

Drilling down into the LCP (Largest Contentful Paint) checkpoint allows a developer to identify elements of the page that are slowing down page loading. It is often ideal to combine this with a URL filter, as LCP elements vary from page to page and between desktop and mobile browsers.

Facet: Click

The click checkpoint tracks where on the page visitors clicked (CSS Selector) and what links have been clicked. This is useful for identifying internal and external navigation targets.

Conversion Definition

The most in-depth analysis that you can perform with RUM Explorer is to understand conversion rates. As RUM does not require any special instrumentation of the page to identify conversions, defining what a conversion is can be done ad-hoc.

This is a three-step process:

1. Filter the traffic (using facets) down to traffic that is a valid conversion
2. Redefine the conversion definition using that filter
3. Analyze the data using the new conversion definition

Whenever facet values have been selected in the sidebar and no conversion has been defined yet, the Engagement key metric area will feature a "redefine" button. Use the filters to select traffic that matches your defined criteria. In this example, we select visitors who have scrolled down far enough to see the footer of the page, so the criteria is a combination of the viewblock checkpoint and the Block CSS Selector (target).

A
After clicking redefine, the draft conversion definition will be highlighted and a confirm button will appear. Click the confirm button to dial in the new conversion definition.

Your new conversion definition has been applied and all elements that referred to Engagement so far will now use the conversion definition. Please note that conversion rate is calculated against the baseline of visits, not against page views like the engagement rate.