To deliver the best user experience (UX), the human-centered design cycle (HCDC) serves as a well-established guideline to application developers. However, it does not yet cover network-specific requirements, which become increasingly crucial, as most applications deliver experience over the Internet. The missing network-centric view is provided by Quality of Experience (QoE), which could team up with UX towards an improved overall experience. By considering QoE aspects during the development process, it can be achieved that applications become network-aware by design. In this paper, the Quality of Experience Centered Design Cycle (QoE-CDC) is proposed, which provides guidelines on how to design applications with respect to network-specific requirements and QoE. Its practical value is showcased for popular application types and validated by outlining the design of a new smartphone application. We show that combining HCDC and QoE-CDC will result in an application design, which reaches a high UX and avoids QoE degradation.
The UX Book: Process and Guidelines for Ensuring a Quality User Experience download pdf 1
Although pure network management or application-aware networks might be in place, a joint network and application management could further enhance the experience for end users [12, 13]. However, this would additionally require network-aware applications, which also consider the network-related aspects of experience. Thus, we clearly see a chance here that QoE, which focuses on a network-centered view, teams up with UX and their human-centered view towards an improved overall experience. This can be achieved by considering QoE aspects during the software development process, such that applications become network-aware and QoE-aware by design.
Having this idea in mind, this paper proposes the Quality of Experience Centered Design Cycle (QoE-CDC). This cycle resembles and complements the HCDC, and provides guidelines on how to design applications with respect to network-specific requirements and QoE. Thus, the primary goal of the QoE-CDC is to ensure that the user experience, which was created by the HCDC, is not deteriorated by network-related issues. To show the practical value of the QoE-CDC, we will discuss past and potential improvements of several popular types of applications in this paper, which could have been resulted from employing the QoE-CDC. Moreover, we will point to open research questions and missing studies with respect to the subjectively perceived experience with these applications in this context. Finally, we will validate the QoE-CDC with an app for crowdsourced video streaming QoE studies, which has been designed from scratch applying both the HCDC and the QoE-CDC.
Network operators and communications researchers have already acknowledged that the subjectively perceived experience with networked applications is a major business factor. The introduction of the concept of Quality of Experience (QoE) [3, 6, 23] moved the focus from the system to the user, putting the user and his perceived experience to the center of the evaluation process [24]. This paradigm shift led to the advent of so called QoE studies, i.e., studies about the impact of technical parameters of systems and networks on the experience of end users, which has produced an enormous amount of findings. The results of these studies are considered by network operators to avoid QoE degradation and improve the experience with a networked application by traffic management, e.g., [10, 11]. By also considering these findings of the QoE community during the development of new or improved applications, QoE degradation due to network issues or fluctuating network conditions can be mitigated by design, which will positively affect the user experience.
In Step 1 of the Quality of Experience Centered Design Cycle, the QoE influence factors of smartphone apps have to be identified. As mentioned above, Internet connectivity is a functional requirement of most smartphone apps, thus, it is also a QoE influence factor. If the app cannot be used when there is no Internet, the users are not satisfied and their QoE will be bad. Apart from connectivity, loading times have a well-known impact on QoE of web browsing related tasks [29]. With smartphone apps, those loading times are omnipresent, especially if large amounts of data have to be downloaded from the Internet and displayed, e.g., in AR/VR apps, social network apps, or mobile gaming, see Fig. 3. However, shorter loading times might also occur in case computations are offloaded to data centers, or when mobile webpages are browsed in simple web or hybrid apps [30]. A 2015 report [31] found that mobile app users are impatient, such that 61% expected apps to start in 4 seconds or less, and 49% expected apps to respond in 2 seconds or less. Moreover, 80% of the users indicated that they will only retry an app up to three times, if they experience problems.
Next, a solution has to be designed, which meets these QoE requirements (Step 3). For this, we will mainly focus on the connectivity requirement, which is common to all smartphone apps. In this process, we have to change the perspective and consider that Internet connectivity is not always available, which can lead to delays when loading content. As the above-described studies reported, such loading times significantly reduce the QoE. To avoid users staring at a blank screen, it is advisable to use loading screens [32]. Using these, users often face a load screen, e.g., a blank screen with a spinning icon or progress bar, which indicates that users have to wait for a specific amount of time. As an alternative to loading screens, skeleton screens become increasingly popular. While loading the content, here, the outline (skeleton) of the content to come is displayed using a simplified presentation, for example, gray boxes and lines. Another possible solution to this problem would be to mask or hide waiting times from the user. If the app notices that no Internet connection is available, it could communicate this to the user and minimize itself to the background. In the background, the app would try to access the Internet and send a push notification to the user as soon as connectivity is available, and the app can be used. This way, users would not be blocked waiting in a load screen, but they could put their attention to something else in the meantime. As soon as they are notified that Internet connectivity is available, they could return and bring the app to the foreground again to continue using it. Note that the same solution could also mask or hide slow Internet connections, where users would have to wait for some content to be downloaded. As a more advanced solution, apps could monitor the mobility of the user, and notify the user, when network coverage or Internet connectivity was lost due to mobility. Then, the user could decide to go back to a place with network connectivity for some time to manually or automatically download some content for the offline phase. Note that some apps already offer the option to manually download content as a preparation for offline phases, e.g., episodes of series or even entire movies in streaming apps. A last potential design solution is caching. It allows to keep popular content in the storage of the smartphone of the user. If the user wants to access this content again, the app does not need an Internet connection because the content is already available on the local device. This technology is already used by so-called progressive web applications [33]. One step further, the app could even pre-fetch content, which is potentially interesting to the user in the future. Pre-fetching means that content is speculatively loaded during times, in which the app has access to the Internet, such that the content would be locally available if the Internet connection breaks. Such pre-fetching could be based on content popularity or typical user interaction patterns. Note that pre-fetching irrelevant content will reduce the available bandwidth and the available storage and increase the risk for exceeding data caps, which has to be considered when implementing this solution.
If streaming applications shall be improved in the QoE-centered design process, the QoE influence factors have to be investigated (Step 1). For video QoE, most works on video streaming agree that initial delay, stalling, and quality adaptation are the most dominant QoE factors [7]. Stalling, i.e., playback interruptions due to buffer depletion, is considered the worst QoE degradation [39, 40], and should be avoided, see Fig. 4. Furthermore, video streams should be played out with high visual quality [41]. In contrast, initial delay has only a small impact on the QoE [29]. For music streaming, similar trends are visible. Here again, stalling is considered as the biggest influence factor of QoE while initial delay plays only a minor role [42, 43]. Having a look beyond the streaming itself, the user satisfaction can also be degraded for increased navigation time (time between starting the app and the actual start of the audio playback) [44].
To meet the QoE requirements (Step 3) and control the delivered QoE, we could not stream the videos from the backend server to the app. Thus, we implemented a file download to transmit the final video files to the frontend app. Only after the videos are completely downloaded to the app, CroQoE allows users to proceed to watching the locally played out videos. This way, additional QoE degradation introduced by fluctuating network conditions of the users can be avoided. Note that this method of pre-download and local playout is typical for crowdsourced video studies, e.g., [61]. Full-screen mode and landscape orientation are used for the video playout. Also, users are not able to control the media during playback. When a video ends, CroQoE displays the experience survey, in which users have to submit ratings on visual quality, streaming quality, quality acceptance, and content liking. 2ff7e9595c
Comments