A significant growth of research in well-being [1] and affective health [2] has taken place in the last decade across a range of disciplines, from human-computer interaction (HCI) and science and technology studies to clinical psychology and psychiatry. This range reflects the interdisciplinary work in this space, and we argue the unique position of the HCI discipline to articulate the design knowledge required for digital well-being interventions. Such work includes novel affective interfaces intended to support real-time awareness of emotions or their regulation [3-7], novel design approaches emphasizing the importance of the human body such as soma design [8], novel technologies supporting reflection and meaning making [9], and those intended to train meditation or mindfulness skills [10] or to conceptualize meaning [11]. Other strands of HCI work have focused on ill-health, such as mobile apps for cognitive behavioral therapy [12] and empirical studies exploring ways to support vulnerable users; such as those living with depression [13], dementia [14], addiction, or the compulsive use of technology; including screen time research [15,16]. Much of such work frames mental well-being as “positive emotional, psychological and social health” [17], whereas digital well-being is broadly seen as the result of being able to use technologies in productive and healthy ways without the negative consequences of dependency, distraction, or risks to users’ privacy [18].

A specific body of work has focused on interventions supporting smartphone nonuse, for instance, by increasing interaction cost to discourage smartphone app use [19]. Both limiting phone use and increasing interaction cost can be conceptualized within the slow movement, where technology is reframed with the aim of pausing and reflecting on its use [20]. Other examples of interventions of smartphone nonuse include apps such as MyTime to make users aware of their tracked use data, which in turn prompt them to reflect upon their use and especially the problematic use [21]. In addition, Roffarello and De Russis [22] argued that current digital well-being apps’ focus on self-monitoring may not be a sufficient mechanism to change users’ behavior with smartphones. Moreover, Roffarello and De Russis [22] also pointed out the limited exploration of the effectiveness and theoretical underpinning of digital well-being apps, whereas van Velthoven et al [23] highlighted also the insufficient investigation of the positive effects of regulating problematic smartphone use with digital interventions. The nascent research exploring the effectiveness of digital well-being apps has been limited, with only 1 study focusing on the analysis of users’ qualitative reviews of commercial apps [22]. Their findings indicate that such apps are liked, especially in studying, working, sleeping, parental control, and free time contexts, albeit limited in supporting behavior change and habit formation toward more conscious smartphone use.

In addition, the theoretical underpinning of digital well-being apps has also received limited attention. In this respect, most work has looked at their adoption [24,25], leveraging, for instance, technology acceptance theories [26,27], including the more recent technology acceptance life cycle model [28]; although these models are rather generic, they are leveraged for personal or domestic technologies. Scholars such as Douglas et al [25], Lukoff et al [29], Lyngs et al [30], Kim et al [31], or Colombo et al [3] have also identified other theories more relevant to digital well-being apps, such as the uses and gratification theory [19,31], theory of planned behavior [32], dual system theory [33], nudge theory [34], framework for behavior change [35], or theories for regulation [36]. However, it is less explored how such theories could actually inform the developing of commercial well-being apps.

Given the limited research on the theoretical and evidence-based aspects of digital well-being apps [23,25], we argue that unpacking the functionalities of the most used commercial apps is an important initial step toward better designing them. The exploration of functionalities and features of mobile apps is an emerging research area, with initial HCI work focusing on digital interventions and especially development of apps for specific conditions such as depressions [37,38] or for supporting, for instance, mindfulness [39,40] or physical activity [41]. In contrast, the functionalities of digital well-being apps have been less investigated. A noticeable exception is the exploration by Roffarello and De Russis [22] of 42 digital well-being apps and their descriptions on Google Play, whose findings indicate the following as key features: (1) tracking user behavior through phone unlocks, phone and app time, and app checking; (2) data presentation through phone and app summary, charts, daily or widget recap, and social comparison; (3) phone interventions through timers and blockers; (4) app interventions through timers, blockers, and notification blockers; and (5) extra features such as motivational quotes or rewards. However, given the brevity of apps’ descriptions available on marketplaces, a richer source to identify their key functionalities is the actual use of the apps, with authors, as HCI experts, adopting the role of the user by directly interacting with the apps—a method previously used for app reviews [38,42,43].

Specific functionalities of digital well-being apps have been also explored through research prototypes usually implementing tracking and notifications [44,45], whereas others included also specific interventions for limiting use [21,46]. For instance, the Socialize [22,47] app integrates the most common functionalities of tracking, data presentation, real-time notifications, and blocking use, which were evaluated in the wild with 38 young people over 3 weeks. Findings indicate improvements in terms of problematic use, measured through the phone addiction scale, and self-regulation, measured through the general self-efficacy scale. Although this is one of the few studies involving measures to explore the effectiveness of a digital well-being app, the Socialize app itself does not appear to be novel, borrowing common functionalities of commercial apps, whose theoretical grounding is limitedly unpacked. The Focus app [48] is another research prototype that leverages Nielsen’s heuristics to support tracking phone use and the blocking of any app, indeterminately or for a limited time set by the user, with the option to unblock them at any time, and provision of educational content on digital addiction. To mitigate overuse from a broader perspective, another research prototype, the FeelHabits app [49], tracks and notifies users about their use of specific apps; albeit rather than on a smartphone alone, this apps tracks use across devices and blocks them if limits set by the user are exceeded.

Another strand of scholarly work with richer theoretical underpinning has focused on restrictive and coercive interventions intended to be stronger than persuasive interventions by supporting users to commit to self-impose limits of use while the phone is blocked [31]. The framework for influencing behavior change [35] suggests the following four types of influence: persuasive (explicit and weak); coercive (explicit and strong); seductive (implicit and weak) and decisive (implicit and strong), which are based on the influencing force (strong and weak); and salience (explicit and implicit). Inspired by this framework, Kim et al [31] designed and evaluated GoalKeeper, a smartphone app featuring both a weak lockout, that is, the phone is locked increasingly longer (eg, 1, 5, 15, 30 and 60 minutes) each time the user exceeds the time they have previously set for use, with each lockout being mitigated by a temporary 15-minute allowance time, and a strong lockout, that is, the phone is locked until midnight without any allowance. Their findings indicate that both mechanisms were more effective than mere notifications of use, with the strong lockout being the most effective, as users set longer limits for not using their phones. Although in the latter case users experienced also more frustration, this was mitigated by the flexibility of setting their own limits and one-time opportunity to modify it.

Despite this growing academic interest in digital well-being, the commercial apps far outweigh the research prototypes in terms of uptake. Thus, the increased adoption of commercial well-being apps offers an opportunity to explore their potentially richer set of functionalities, and the aim of this paper is to articulate these functionalities as well as the novel design implications informed by them to better inspire the design of technologies for well-being. To address this aim, we focused on the following research questions:

Our contributions are 3-fold. First, we unpacked richer insights about tracking and monitoring functionalities in terms of user profiling and understanding of monitoring as a complete, location-based, and flexible practice that can benefit from tailored, time-based visualizations. Second, we identified 4 interventions for limiting use including richer understanding of different types of obstacles for limiting use and specific features for less explored functionalities such as supporting awareness for reaching use limits, focused attention, and motivation to keep within set use limits. Third, grounded in our findings, we generated 6 design implications for digital well-being apps.

To identify the digital well-being apps, in winter 2019, we searched for free apps in Google Play Store using the following search terms: digital wellbeing, digital detox, detox apps, unplugging, and distraction, which is a new direction given that extensive previous work on such apps has prioritized addiction and screen time [22]. We have focused on Google Play because its apps represent the largest global market share, >2.5 greater than iOS apps [50], whereas the latter is also more restrictive in terms of available information [22]. However, future work could extend this exploration to other platforms.

For each search term, the top 250 most relevant apps returned on Google Play were retained, totaling 1250 apps, with 37 duplicates. At the screening stage, after reading their titles, summary descriptions, and main screenshots, we excluded 931 less relevant apps such as fitness, activity planner, or nondigital detox apps. The eligibility of the remaining 282 apps was assesses based on their full descriptions, with further 147 apps being excluded such as utility apps, games, and general well-being and meditation practice apps. From the remaining 135 apps, we further excluded those with less than 1000 raters and with average rating score <4, leading to 39 apps to be included in our review. The PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) diagram for the searching and screening process for digital well-being apps is shown in Figure 1. We also note that of our 39 apps, 12 (31%) are also available on the Apple Store, with 7 (58%) of them having user rating >4.2.

Our final set consisted of 39 digital well-being apps (Multimedia Appendix 1 [19,21,22,31,34,44-46,48,49,51-57]), which were analyzed through two complementary methods: first, a review of their functionalities based on their descriptions from Google Play and, second, an autoethnography with the authors (SA and CS), as HCI experts directly interacting with them in order to viscerally understand how these functionalities work and are experienced by potentially users in their daily lives. Such interactions were iterated, involving at least two sessions for each app, lasting for at least 30 minutes. For the autoethnography, we used a Samsung Galaxy Note 9 phone with an Android mobile operating system.

The first author (SA) evaluated all 39 digital well-being apps, whereas the second author (CS) evaluated 21% (8/39) of the apps. Through the use of both methods, the authors (SA and CS) iteratively revised the coding scheme over several months, a process that has followed a hybrid approach. This integrated deductive codes, informed by prior work on functionalities [22] such as tracking, data presentation, and interventions (Table 1). The inductive coding was informed the distinction between tracking and monitoring, the revision of intervention functionalities such as tracking phone or app use by setting limits and of data presentation or visualization and its subcategories such as numerical and diagrammatic format through charts, round diagrams, metaphors, heat maps, or reports. Particularly important are the new functionalities capturing 4 interventions for limiting use.

To better contextualize our review in scholarly work, we subsequently extended the list of apps with 17 digital well-being apps designed in academia, which we found through search on Google Scholar using the following keywords: digital wellbeing application or digital wellbeing app. This search returned 42 papers, which after reading their abstracts, led to 17 papers describing such apps [3,8,12,18,26,32,37,49,51,58-65]. The remaining 25 papers do not included digital well-being apps and for this reason they were excluded. We have explored the functionalities of the apps described in the 17 papers by applying the above coding system to their description, as not all of them were available to download from app marketplaces. All the tables provided in Multimedia Appendices 1-7 [19,21,22,31,34,44-46,48,49,51-57] include information on both commercial and academic apps.

The descriptive characteristics captured by our analysis include app category, target users, scientific underpinning and evidence base, and cost. Findings indicate that the top-rated digital well-being apps belong to 6 categories, albeit feature predominantly in the Productivity category (27/39, 69% apps), followed by fewer apps in the Tools (4/39, 10%), Personalization (3/39, 8%), Health and fitness (2/39, 5%), Parenting (2/39, 5%), and Lifestyle (1/39, 3%) categories. Together with their main aim of limiting phone overuse, this is an interesting outcome that can be linked to the ethical principle of nonmaleficence [2] to protect users from the negative impact of phone overuse. These can also be aligned with the ethical principle of beneficence, particularly the predominant instrumental value of digital well-being apps supporting increased productivity rather than their eudemonic value for supporting meaningful goals [66]. Arguably, the latter would further strengthen their beneficence potential.

Another important outcome, which can potentially hinder their beneficence is the limited science base of digital well-being apps, with 97% (38/39) of the apps not specifying if they are backed up by research, the only exception being the Focus To-Do app described as a science-based app. This indicates the importance of these apps unpacking in their descriptions the scientific underpinnings informing their design and any available outcomes from evaluation studies. This, in turn, will support users make more informed choices regarding their beneficence.

The target users of digital well-being apps appear to be unrestricted, with most of them available to users of all ages, which reflects the principle of justice. Indeed, all but 4 apps are rated on Google Play as “PGEI 3,” which stands for “Pan European Game Information.” The remaining four apps do not specify any age: Brain Focus Productivity Timer, Lock Me Out, SleepTown, and Sma-Phospital. Interestingly, the design of the apps does not vary with age, as we see the same functionalities for both children and adults. We also examined whether the target users include the clinical population. Findings indicate that of the 39 apps, 38 (97%) do not specify clinical user groups, whereas 1 (3%) app, that is, AppBlock, mentions its suitability for children or adults with attention deficit hyperactivity disorder. This suggests that digital well-being apps predominantly target users without specific conditions or health concerns. However, given their value for supporting attention, some of these apps may be beneficial for users with attention deficit. Future work should further explore this.

Also related to justice, the cost of the digital well-being apps is an important aspect that can increase or limit diverse users accessing them. Regarding cost, an important outcome is that although all the 39 apps are free to download, only 11 (28%) are entirely free to use, whereas 28 (72%) offer in-app purchase mostly for removing adds, unlocking premium features, or subscribing to premium versions of the apps. This is an important outcome indicating that most functionalities of these apps are freely available, making their use particularly inclusive.

Digital well-being apps have an interesting relationship with the ethical principle of autonomy. On the one hand, these apps tend to limit one’s use of phone or apps; on the other hand, consistent findings have shown that autonomy is already impaired [67] when people live with some form of addiction such as phone overuse.

This section focuses on the comparison of the functionalities of the apps developed in academia with those of commercial apps, with a specific focus on how they differ. It is not surprising that most academic apps share the tracking and monitoring functionalities available in commercial apps. For example, the lockout mechanism that blocks the phone until midnight when reaching the use limit [31] is similar to blocking apps and phone when the user exceeds the defined time limit in some commercial apps (ie, Ubhind). Similarly, blocking and scheduling blocking in the academic app Forest [48] are comparable with those in the commercial app AppBlock. An interesting distinction concerning tracking and monitoring is the new form of visualization of tracked data in academic apps, namely, timelines.

In terms of interventions for limiting use, findings indicate additional key distinctions between commercial and academic apps for digital well-being. Regarding creating obstacles to limit phone or app use, important distinctions concern the force and saliency of the created obstacles and their temporal aspect and source. With respect to force, commercial apps predominantly use strong obstacles such as phone or app blocks (14/39, 35% apps) instead of weak obstacles such as notifications or microboundary interactions (5/39, 13% apps), with only 3% (1/39) of the apps providing both strong and weak obstacles. In contrast, academic apps take a more balanced approach, using equally both strong (10/17, 59%) and weak (11/17, 65%) obstacles, with 29% (5/17) of these apps using both strong and weak obstacles. Given the nascent research exploring the effectiveness of digital well-being apps, academic work is more likely to use both types of obstacles to compare their effectiveness.

With respect to the saliency of obstacles, almost half of the commercial apps (17/39, 44%) specify saliency, with all but 1 featuring explicit obstacles (which also tend to be strong), whereas the SPACE app features implicit obstacles. In contrast, almost all academic apps (16/17, 94%) involve explicit obstacles, that is, mostly notifications. What is interesting here is the innovative use in academic apps of a new type of obstacles for restricting use through design frictions. These could involve mandatory cognitive tasks such as entering several digits as users attempt to start interacting with apps targeted for limited use [57] or entering 30- or 10-digits try [19], which, when compared with merely pressing OK, indicates that the more complex the cognitive task, the more likely that users will restrain from engaging with those apps. Commercial apps present limited such cognitive tasks, with one exception being the MMGuardian app, which requires entering a password by parents to prevent the child from removing the app or modifying the set time limit of use.

Findings also indicate differences regarding the temporal aspects of obstacles to use. Although commercial apps use these obstacles predominantly after use of the targeted apps (15/39, 38%), as opposed to during use (4/39, 10%), academic apps take a more balanced approach using such obstacles equally during (8/17, 47%) and after the use of targeted apps (8/17, 47%), with 12% (2/17) of the apps using them both during and after use. This suggests not only the value of providing flexibility and users’ choice but also the importance of real-time obstacles in limiting phone or app overuse in real time.

Regarding obstacles’ source, commercial apps use mostly obstacles set and customized by users (15/39, 38%) rather than obstacles set automatically (6/39, 15%); in contrast, academic apps feature more automatically set obstacles (10/17, 59%) than those set by users (6/17, 35%).

Scholarly work on digital well-being apps has also focused on the types of apps that users are more willing to limit use. In this respect, empirical findings indicate that users were willing to restrict the use of specific apps such as messaging ones [56], as well as social media or games apps [21]. Academic work has also explored limited use not only beyond individual devices such as phones but also across multi-devices and their context of use [46,49]. Similar work has looked, for instance, at chatbots to notify users of their smartphone use [51] or video platforms supporting preschoolers to self-manage their phone and app consumption [52].

The second intervention, intended to increase users’ awareness of reaching their limits of phone or app use, also shows differences. Although both sets of apps use mostly explicit notifications to support such awareness, academic apps do so more (8/17, 47%) than commercial apps (11/39, 28%). Interestingly, both sets of apps also used implicit notifications such as screen dimming featuring in the SPACE app and vibrations for notifying users when they exceeded their set time limit for phone use featuring in the Good Vibrations app [34].

The intervention targeting focused attention has been supported by both sets of apps through training for focused attention, with 21% (8/39) of commercial apps and 29% (5/17) of academic apps providing such training. Interestingly, commercial apps also feature white noise as a specific mechanism for supporting focused attention, whose effectiveness as part of digital well-being apps has been less explored, although a body of scholarly work has shown its value for relaxation [17,70].

Finally, the fourth intervention for supporting motivation to keep within set limits shows similar findings for the 2 sets of apps, with emphasis on rewarding user behavior when the goal of keeping within limits has been reached (9/39, 23% commercial apps; 3/17, 18% academic apps), albeit commercial apps show more diverse forms of rewarding content, usually leveraging gamification principles, as opposed to academic apps, which use merely points. In contrast, findings show much fewer apps leveraging punitive feedback when users fail to keep within set use limits for both commercial apps (4/39, 10%) and academic apps (1/17, 6%). In terms of social support, a small number of apps provide it to support cooperation (5/39, 13% commercial apps; 2/17, 12% academic apps), competition, and recognition (5/39, 13% commercial apps; 3/17, 18% academic apps).

Also unique to research on academic apps for digital well-being is the extended focus of their audience to include not only individual users as commercial apps but also groups of users. For example, such academic apps focused on enhancing self-regulation through groups of users collaborating or competing toward limiting their collective use of phones and apps [45,54], through limiting use as a family activity [45], or through providing in-app spaces for college students to restrict their phone use during class time [44].

We now revisit the research questions advanced in the Introduction section and articulate the novelty of our key findings. The first 2 research questions focused on identifying the key functionalities of the top-rated digital well-being apps and their theoretical underpinning. Our review of top-rated digital well-being apps indicates the following six main functionalities: tracking use; monitoring use against set limits; and four interventions for limiting use, namely, creating obstacles to limit use, supporting awareness of reaching the set limits, supporting focused attention, and supporting motivation for limiting use. In this section, we also theoretically position these functionalities and leverage them to articulate new implications for better designing digital well-being apps.

Findings indicate that >70% (29/39) of digital well-being apps provide tracking of use of phone or app data, visualized mostly through reports and charts. More than a third (11/29) of the apps providing the tracking functionality also support user profiling, either automatically from tracked data or through users’ entered data. This aspect of tracking has been limitedly explored in previous work [22,23]. Another key finding is that almost 30% (11/39, 28%) of the digital well-being apps do not support tracking phone or app use but support instead focused attention or tracking of offline activities. This is a key outcome with important design implications that we will revisit later.

The second functionality is monitoring phone or app use against set time limits, which is key for limiting their use. This functionality features in 64% (25/39) of our reviewed commercial apps. Interestingly, however, the remaining almost 35% (14/39) of the digital well-being apps do not support this functionality directly, albeit they monitor the time spent on offline activities, away from the phone and its apps. From the former set of apps monitoring phone and app use, most tend to target some of the apps installed on the phone, with fewer digital well-being apps monitoring the use limits of all the apps. An important implication here is designing for complete monitoring of all the apps installed on the phone and providing users with the choice of selecting the ones to monitor, as well as location-based monitoring currently limitedly supported, albeit useful for situating the monitoring behavior in a spatiotemporal context. We also suggest supporting flexible monitoring allowing circumventing the set use time limit, which can support ongoing motivation for monitoring phone consumption and regulating phone overuse behavior. Findings also indicate interesting time-based, monitoring-specific visualizations featuring in approximately half (20/39) of our reviewed apps, which are useful to be extended to all digital well-being apps.

With respect to the first intervention, almost half (16/39) of our reviewed commercial apps implement strong and explicit obstacles, such as blocking to limit phone or app use, with much fewer apps featuring weak or implicit obstacles, usually in the form of notification. Even fewer apps attempt to implement microboundary interactions using launchers as substitute home screens. Such obstacles can be generated either automatically or by the users, with only few apps tailing them to user profiles and none mapping the force of obstacles (strong or weak) to such profiles. These approaches suggest the value of using both sources, so that digital well-being apps could benefit from the customization of users’ set obstacles and potentially even more so from extending the use of automatically set obstacles. Although previous work suggested that strong obstacles, despite inducing frustration, can be preferred by users and are likely to be more effective than the less restrictive obstacles [22,46], the value of providing both strong and weak obstacles can be further explored, in terms of both effectiveness and user experience for more sustained and long-term change of one’s relation with their mobile phones. Our findings from academic apps also highlighted new explicit obstacles for restricting use through design frictions such as cognitive tasks, which, unfortunately, have been limitedly explored. However, these innovative obstacles open up an interesting design space, as frictions support users to pause before compulsively re-engaging with their phones and apps, and thus a more mindful interaction.

From the 25 apps that support monitoring, 13 (52%) support users becoming aware when they reach the set use limits of their phone or apps, mostly through explicit notification and much less through implicit ones such as screen dimming, whereas daily reminders support a high-level awareness of use patterns exceeding set limits. Academic apps also started to explore implicit notification, albeit in tactile modality, through vibrations. These implicit notifications open up a less explored design space for this intervention. Arguably, vibration-based notifications are weak obstacles, and illustrations of how nudge theory [34] can be leveraged in the design of digital well-being apps. Implicit notifications may be less intrusive and therefore more persuasive, although future work is needed to explore their specific benefits when compared with explicit ones.

An important outcome is the 2 ways of supported focused attention that digital well-being apps implement: The first is implicit support through the monitoring and limiting of phone or app overuse, and the second is the explicit training of attention by focusing on offline activities without phone use, including also exposure to white noise to support concentration, which has strong research underpinning [17,70].

A key functionality, less explored in previous research on digital well-being, is supporting motivation for keeping within limited use. For this functionality, we identified the following three mechanisms: reward and penalty leveraging gamification principles, educational and motivational content, and social support provided, however, by a limited number of apps and where cooperation among users is limited.

The theoretical underpinning of digital well-being apps has received limited attention. However, Roffarello and De Russis [22] suggested the value of grounding the design of well-being apps to support behavior change, habit formation, and self-regulation. As shown in the Introduction section, scholars have identified a range of theories that may inform the design of digital well-being apps, such as those of uses and gratification [19,31], planned behavior [32], dual system [33], nudge [34], framework for behavior change [35], or regulation [36]. However, it is less explored how such theories have been actually informing the developing of commercial well-being apps. However, the operationalization of these theories in this respect has been limited. In this section, we argue for the value of self-regulation theories.

Previous work has shown that tracking is a key functionality of digital well-being apps that captures the use of the phone and its apps [22]. However, this does not make the important distinction between the digital well-being app running in the background to collect such information and the user’s active effort to minimize phone use. The former is usually important in the early stage of digital detox when people want to understand their consumption patterns, whereas the latter follows with setting up limits to phone or app use. For this, we called the former tracking, and the latter monitoring, which is a better term for capturing or tracking data against a specific target. Most behavior-changing apps use monitoring toward specific goals such as exercising ones [41]; therefore, the link between monitoring and goal setting is crucial. We note the important alignment of monitoring functionality to the three ingredients of self-regulation as reflected in self-regulation theories: setting target standards, monitoring the current state against these targets, and activating processes to reduce any identified distance between the current state and the targets [59]. Thus, we argue that designing for monitoring functionality can benefit from theoretical grounding in self-regulation theories.

Regarding the intervention of creating obstacles to limit use, we have seen the value of both strong restrictive mechanisms and weak ones, mostly explored in academic research rather than reflected in commercial apps. We argue that weak and particularly implicit obstacles are illustrations of nudges, which nudge theory describes as persuasive attempts for behavior change that do not limit users’ choices [73,74]. Future work is needed to understand how nudge theory can be sensitively leveraged to rigorously inform such obstacles to use.

The intervention for supporting focused attention is particularly interesting, as it marks a shift away from limiting excessive use toward more mindful activities, either technologically mediated or offline, whose valuable side effect is limited use of the phone or apps. Rather than steering away from undesirable behavior, this intervention encourages engagement in meaningful and ideally desirable activities, subsequently supporting the most powerful appetitive rather than aversive motivation. We also highlight in this context the value of supporting users to understand and support their meaningful goals [66], which subsequently can address the phone overuse and the boredom often associated with it. However, goal theories have been limitedly discussed in relation to digital well-being apps.

The final intervention focuses on supporting motivation to keep within use limits. Although limitedly mentioned in relation to digital well-being apps, we suggest the value of broaden and build theory [75], where positive emotions are leveraged for increased self-awareness and behavior change. Illustrations of how this theory may be underpinning some of the identified functionalities include the provision of allowances for overruling the set use limits during monitoring. This is important for instrumental reasons both allowing the completion of some immediate tasks, and maintaining motivation in case of setbacks in meeting the set limits. In turn, this could broaden users’ resilience and more flexibly support the acknowledged high demands of self-regulation [36]. Future work is needed to explore effective ways for managing the negative emotions associated with setbacks.

We focused on Google Play, which limited our review of iOS apps not available on Google Play. Future work could extend this exploration to other platforms. Future work can also aim to further strengthen the scientific underpinning of the design principles of digital well-being apps, in terms of both their theoretical framing and evidence-based evaluation studies. Our findings indicate that despite the growing number of digital well-being apps, parts of their design space have been less explored, such as supporting awareness for reaching use limits, motivation to keep within set use limits, implicit obstacles rather than explicit ones, recommended interventions to determine the right type of obstacles according to the tracked data, and mechanisms for supporting focused attention. We encourage researchers and developers to focus on these aspects, and together with the key features identified in our study, they can significantly improve the design of digital well-being apps.

We report on a functionality review of 39 commercial and 17 academic digital well-being apps. Findings provide richer understanding of tracking and particularly monitoring functionalities, together with 4 interventions for limiting use. These provide new understanding of the different types of obstacles for limiting use, as well as of specific features for less explored functionalities such as supporting awareness for reaching use limits, focused attention, and motivation to keep within set use limits. We conclude with 6 design implications for digital well-being apps, namely, calling to move beyond screen time and support the broader focus of digital well-being; supporting meaningful use rather than limiting meaningless use; leveraging (digital) navigation in design for friction; supporting collaborative interaction to limit phone overuse; supporting explicit, time-based visualizations for monitoring functionality; and supporting ethical design of digital well-being apps.