Network-based applications used to be simple things. A user connected to a server through a browser. Today, that's only one part of a complex picture. Mobile applications, internal business applications, and autonomous clients are all important to success. They connect not just to on-premises servers but to cloud services and third-party resources.
To know how well its data services are performing, a business needs to monitor and measure all its client-server applications. They include the following and more:
- Applications running in the browser
- Mobile applications
- The Internet of Things: autonomous and semi-autonomous devices
- Cloud applications
- Microservices, on or off the premises
Gartner has reported that traditional performance monitoring tools leave large gaps. Application Performance Monitoring has expanded into Digital Experience Monitoring (DEM).
DEM is related to end-user monitoring but isn't limited to human users. It addresses the quality of all client-server connections, whether the client is a human or an autonomous device. According to Gartner, DEM applies to "the operational experience and behaviour of a digital agent, human or machine."
Why digital experience is important
A brick-and-mortar store can have the greatest merchandise anywhere, but unless its appearance, layout, and service match the product's quality, it won't get and keep customers. Digital experience is the equivalent for a business's digital infrastructure and public interface. Keeping its quality high is important for customer interaction, B2B transactions, and internal operations.
Digital experience is a brand differentiator. Well-known sites such as Amazon and Google stand out not just for what they offer, but for being reliably available and delivering error-free performance and ease of use.
A good digital experience holds up through periods of peak demand. Peak demand periods are when the most users are looking at you! It's important to keep applications from bogging down, even — or perhaps especially — when they're heavily loaded.
Third-party services are an important consideration. They aren't directly under your control, but an erratically performing service can drag down the usability of your applications. If that happens, you need to look into improving or replacing it, but first you have to know about it.
Digital Experience Monitoring scenarios
To get a better idea of what's involved in DEM, let's look at a few different scenarios. They range from human interaction to purely machine-to-machine communication, but they have features in common.
An e-commerce site is a common example of human interaction with an application. Measurements of the digital experience include uptime, proportion of transactions that succeed, and timing statistics. The page load time is often cited, but at least as important is "time to interactive," or TTI. This is the time till the user can start performing actions on the page, even if the images aren't all loaded. User error rates and abandoned transactions are important indicators; if they're high, there could be a design problem.
The situation is mostly similar when the client is a mobile application rather than a browser. Banking apps let users check their balances, make deposits and payments, and schedule future activity. TTI is less important when the app already resides on the device. A greater concern is the ability to deliver a good experience on an erratic mobile connection. The service needs to be tolerant of dropped packets and not require excessive bandwidth.
A wearable device for tracking exercise is an intermediate case. It does a lot of pure machine-to-machine interaction, but it has a human interface. Many of the same factors, including uptime and successful exchanges of information, come into play. As an autonomous machine client, it exchanges a lot of short packets with the server, and consistency and low latency take on more importance. There won't be user errors, but there could be anomalous data, and it should be rare.
A temperature and pressure monitor in a factory is an autonomous machine client. Humans aren't involved at all. Considerations are similar to the wearable device, but constant availability is very important. The inability to send an alert to a server could mean serious trouble. When large numbers of these devices are deployed, low latency and reliable exchange of information are critical.
How to approach DEM
Old-style application performance monitoring is server-centric. It works well when clients talk only to an on-premises server, but there are many cases where that model no longer fits. Mobile devices, cloud services, and microservices present a more complex situation. Today, full visibility requires an updated approach. It has to take into account everything the client communicates with, including services that aren't under your direct control.
When we talk about digital experience, we aren't talking about a single metric. A variety of factors contribute to the sum.
Availability and uptime: Multiple, independent components may have their own patterns.
Response time: Data rate, server performance, packet loss, and latency all contribute to this factor.
Ease of use: This applies to humans, and it's harder to measure than objective numbers. User interface design and technical features both contribute to it.
Success in use: A successful digital experience will include few failed or abandoned transactions. It just works.
Making the best use of DEM requires taking all these factors into account and looking for areas that can be improved.
Developing a Digital Experience Monitoring strategy
DEM overlaps with end-user experience monitoring (EUEM) as well as APM. Traditional APM assumes that the application is entirely in the business's possession. DEM expands on that idea to take cloud servers, microservices, and third parties into account. Instead of being server-centric, it takes the client's perspective so it can measure all the relevant factors.
DEM uses both synthetic monitoring and real-user measurement. Synthetic monitoring uses a simulated client; it has the advantage of testing in a controlled environment and catching problems before any live users touch the application. Real-user measurement (RUM) tracks how people actually use the system, and it can turn up use cases which synthetic monitoring misses.
A DEM strategy should start with an inventory of all the services that need tracking. This includes all network-based applications and services, including third-party services and strictly internal ones. The monitoring tools need to be flexible enough to work with all types of applications. Different applications run on different devices, communicate with different services, and use different protocols. Each one has its own potential issues to identify and measure. To learn about our DEM tools, contact us today.
The measure of a business's digital services comes from the experience which all of them deliver. Whether they're internal or external, whatever kind of client they run, they all need to be tracked for their quality of digital experience. When you know which services are delivering satisfactory results and which have room for improvement, you know where to focus the efforts of developers and the IT department.