End User Experience Monitoring Guide

End User Experience Monitoring (EUEM) is an essential task for any website owner. Although your team might think that the website looks great and works well, you need to test its functions in the real world. Does the site work properly for visitors accessing it from different locations in the world? Do pages load quickly enough? Are there elements in the website that don’t always load properly? Are there broken links on the site?

EUEM, also known as digital experience monitoring, provides continuous availability testing for your site. It operates primarily as a technical test of the elements of a website. However, it also provides feedback for marketers who want to know whether the site facilitates the buyer’s journey properly.

On a technical level, you don’t want to wait for users to contact you to tell you that some part of the site is not working. In many cases, first-time visitors and casual users won’t feel that is their responsibility – they will just move on to the next site on the search engine results page. Regular users might report problems. However, errors on web pages can wear down the user community’s loyalty, so it is better to spot those problems and fix them before users feel forced to contact you.

Website design is a recursive process and EUEM provides the feedback needed to assess each element on each page of a site and the site’s success overall. You can put up a survey on the site to ask visitors what they think of various aspects of the website but not everyone is going to agree to complete the survey and those that do won’t necessarily tell the truth. Looking at what real people do on your website confirms that you have got the design right or informs you where changes need to be made.

EUEM and related terms

As a cutting-edge discipline, IT is very good at creating new terminology and lots of abbreviations. Many people end up using IT acronyms in their daily speech without really understanding what those terms mean. There are a lot of new terms that sound like EUEM but not all of them relate to the field of website monitoring and testing by watching user actions. Here are some terms to look out for.

  • EUMEnd User Management is not the same as EUEM. EUM refers to the management of user accounts. Many websites encourage users to create accounts in order to buy things or save preferences. However, EUM is not part of EUEM other than checking on the ease of the process that website visitors need to complete in order to set up an account.
  • UEMUnified Endpoint Management relates to the monitoring and management of desktop computers and mobile devices that have access to a business’s network and facilities. It has nothing to do with EUEM.
  • UXUser Experience is a consideration applied to the design phase of a website. It is a recursive process that includes testing response time and ease of use. UX is a recursive process and feedback from testing may require parts of a newly developed site to be redesigned. Although not exactly the same as EUEM, UX does rely on EUEM to provide confirmation of the design once a site goes live.
  • RUMReal User Monitoring is almost synonymous with EUEM. RUM is the mechanism by which EUEM is conducted – it is like a toolkit. In the field of EUEM, RUM is termed “passive monitoring.”
  • STMSynthetic Transaction Monitoring doesn’t include any monitoring of end-users, so strictly speaking, it shouldn’t be regarded as part of EUEM. However, it is. STM and its wider field of Synthetic Monitoring (which doesn’t have an abbreviation) is known as active monitoring or proactive monitoring. It uses automated processes to simulate user actions.

Real User Monitoring & Synthetic Monitoring

EUEM requires planning. There are some aspects of website performance that need to be tested almost constantly, while others only need to be examined if regular testing identifies an area that might be experiencing problems.

Typically, EUEM involves both synthetic monitoring and real user monitoring. Synthetic monitoring is cheaper to perform than real user monitoring, so it is performed more frequently than RUM. Generally, it is more cost-effective to perform as much EUEM testing as possible with synthetic monitoring and leave RUM for website features. This is because synthetic monitoring can’t properly exercise or for marketing assessments.

See also: Synthetic Monitoring vs Real User Monitoring

Synthetic monitoring tasks

Network managers will be very familiar with synthetic monitoring techniques. The performance of automated tests on networks, servers, and applications are part of their everyday working practices. A big advantage of synthetic monitoring is that it can spot problems with sites before they are encountered by users.

Synthetic monitoring tools fall into three categories:

  • Uptime monitoring
  • Response time monitoring
  • Transaction monitoring

Of these three categories, uptime monitoring is the cheapest and is performed more frequently. Many synthetic monitoring providers categorize uptime and response time monitoring as standard tests but place transaction monitoring in an advanced test category along with real user monitoring. Advanced tests are charged at a higher rate than standard tests.

Uptime monitoring

The most frequently implemented method of synthetic monitoring is a Ping test. This only checks that the Web server is available. It is possible to set up a recursive Ping to the web server’s IP address and watch out for test failure messages. However, the user community for a website accesses the internet from many different places around the world, so it is better to perform those Ping tests from many servers.

Uptime monitoring is also known as availability monitoring and there are more sophisticated tests than Ping available in this category. Testing with an ICMP Ping gives a domain name rather than an IP address simultaneously checks that the DNS entry for the website is correct. Testing with a TCP connection request will check that the HTTP and HTTPS services on the site’s Web host are active.

The top-level of uptime monitoring is implemented by an HTTP Get command. This requests the code for the web page from the Web server. This test automatically also checks the session establishment process and will therefore spot any problems with the site’s SSL certificate on an HTTPS connection.

Response time monitoring

Response time monitoring is usually performed less frequently than uptime monitoring. As a response time test can’t be performed if a website is unavailable, a response time monitor simultaneously performs an availability test.

Some synthetic monitoring services offer test allowance bundles and will include a lower allowance for response time monitoring than uptime tests. This is one of the main reasons that website owners perform response time tests less frequently than uptime monitoring.

Effective response time monitoring requires a page to be loaded into a browser. Response time test providers will advertise tests in one specific browser by default, which is usually Google Chrome. However, many services offer the subscriber a choice of browser and it is possible to perform the same tests with many browsers – although each browser test will be charged for individually. It is also important to test the mobile version of a site.

As with uptime monitoring, response time monitoring should be performed from different locations around the globe. Edge service providers offer site optimization and content delivery networks to speed up the delivery of the site to users located far away from the Web hosting service. If a server in that network goes down unexpectedly or performs badly, part or all of your site will be slow to load in one part of the world while performing well in another country.

There are many metrics that a response time monitor can offer. The main web page events to look out for are:

  • First Contentful Paint (FCP) – This is a timestamp of when the first visible element appears in a browser and lets the user know that the site is actually being delivered. The element could be text, an image, or a graphic.
  • Largest Contentful Paint (LCP) – The time the largest element in the initial view of the page is visible. This is likely to be a headline image on the page.
  • First Input Delay (FID) – This is the moment an interactive element on the page becomes active. For example, a link. The FID response time does not wait for all interactive elements to become active.
  • Time to Interactive (TTI) – This is the moment when all the interactive elements on a page become responsive.
  • Total Blocking Time (TBT) – The amount of time that “long tasks” block the main loading activity. Long tasks are segments of JavaScript code that cause the page to freeze for more than 50 ms.
  • Cumulative Layout Shift (CLS) – The elements in a page will move as other elements become visible. This is particularly noticeable with text when a font is applied. The loading of images will also move all subsequent content down the page. The CLS is the moment that all elements are in their final position.

These response time measurements are known as Web vitals. Google defines LCP, FID, and CLS as “core Web vitals.”

Synthetic transaction monitoring

Synthetic transaction monitoring is considerably more complicated than uptime and response time monitoring. The transaction monitoring task requires a script of actions to be executed. These simulate the actions of real users and test interactive elements on a page, such as drop-down lists or video players. The aim of synthetic user monitoring is to check that combinations of elements work well together. Thus, they are more likely to be used to check on a web form.

A typical situation where a synthetic transaction monitor is useful is where the contents of one set of options are influenced by the user input in an earlier field. For example, an address finder that works on an entered postcode needs to present a list of addresses that belong to that postcode. Thus, the candidate values in the address field can only be populated once a postcode has been entered.

See also: Synthetic Monitoring

Real user monitoring

While synthetic monitoring implements a series of tests, RUM is a passive form of monitoring. It relies on data collection elements on a web page that track actual actions of real visitors to a site. Once the steps taken by a user have been recorded, data can be aggregated to provide metrics for a page. RUM stats rarely detail the actions of a single user.

The advantage of RUM is that it isn’t susceptible to tester forgetfulness. There might be combinations of actions possible on your site that you never accounted for when writing up your own test scripts. Another advantage of RUM is that it provides a much wider testing environment than a synthetic testing platform. There is a very large number of permutations in access technology when all the users in the world are taken into account. For example, there are many makes and models of laptops, not all running exactly the same version of Windows and with different update points for Chrome. There are so many unique combinations of these few factors that marketing trackers can identify the same user’s activity across different sites without relying on each person always using the same IP address.

Real user monitoring is very similar to the tracker libraries deployed by marketers. A RUM exercise requires data gathering code to be included in the HTML of a page. This is known as “JavaScript Injection.” RUM platforms are usually provided by specialist edge services. These providers supply APIs that call monitoring procedures that are resident on the monitoring service provider’s servers. These can be implemented as microservices.

Unfortunately, calls to extra functions resident on other servers can also slow down the load times of web pages.

See also: Real User Monitoring Guide

Who needs EUEM?

All job titles involved in creating and managing a website will need to use some form of EUEM. Synthetic monitoring is more suited to the information that network managers and website administrators need. RUM is essential for UX specialists in order to confirm the effectiveness of a specific feature on a web page.

Developers engaged in agile development also need to use RUM. Agile development is a staged process with some parts of a site going live before all planned features are ready. In this scenario, the introduction of a new element needs to be tested by both synthetic monitoring methods and RUM.

Marketers will be more interested in RUM because it facilitates typical market testing strategies such as A/B testing.

Whether it is in your job description to make sure that a website is working or if you earn your income by making a website stand out and appeal to consumers, you will need to use end-user experience monitoring.