Do Activity Trackers Really Help?

img_5597By Adam Eckart, MS, CSCS, FDN-P     

     In recent years, wearable activity tracking devices have gained widespread popularity. Research by Bravada et al. (2007) and Cadmus-Bertram, Bess, Patterson, Parker, and Morey (2015) has shown that the use of wearable activity trackers increases short-term activity levels above previous levels in some individuals. Consolvo et al. (2008) and Lin, Mamykina, Lindtner, Delajoux, and Strub (2006) suggest wearing activity tracking devices increases self awareness and provides feedback prompting a user to modify behaviors to increase activity levels. Despite the evidence showing the benefits, new research has shown that measuring activity undermines intrinsic motivation, thus, decreases the enjoyment of the physical activity performed and may lead to cessation of activity (Etkin, 2016). Additionally, a recent study by Jakicic et al. (2016) found activity tracker users lost less weight than those on a standard weight loss intervention.  Research by Clawson, Pater, Miller, Mynatt, and Mamykina (2015) and Li, Dey, and Forlezzi (2010) has shown discontinued or decreased usage of wearable activity trackers is attributable to several factors including device limitations, technological complexity, and incompatibility with changing goals.

      The study of persuasive technology for the purposes of improving health through daily tracking and feedback began in the early 2000s. Devices are considered persuasive if new goals are encouraged and goal achievement is rewarded. Since inception of wearable activity trackers, researchers have failed to show clear associations between use of activity trackers and long-term health behavior. However, researchers have shown short-term increases in activity levels with the use of activity trackers (Bravata et al., 2007). While some researchers have reported long-term tracker usage among participants, study design limitations make it difficult to draw clear associations between tracker usage and long-term behavioral change (Fritz, Huang, Murphy, & Zimmermann, 2014). The researchers of the interview study by Fritz, Huang, Murphy, and Zimmermann (2014) did not question former tracker users to gain clarity on the individual reasons for discontinued use. Similarly, other studies by Day (2016) and Li, Dey, and Forlezzi (2010) using behavioral constructs to determine the efficacy of activity tracking devices utilized qualitative analysis due to small sample sizes and yielded inconclusive data.  Though there exists a substantial body of knowledge in self-quantification theory and technology, it is not quite clear why the evidence for long-term adoption remains inadequate. A deeper look into the literature reveals that behavioral change presents a multifactorial affair and that the current state of activity tracking technology may be limiting people from long-term healthy physical activity habits. Therefore, the purpose of this paper is to analyze the efficacy of current wearable activity tracking technology on long-term behavioral change.  

The Current State of Activity Tracking May Not Contribute to Long Term Behavioral Change

      A recent Internet survey revealed that nearly one in 10 adults owns an activity tracker (Ledger & McCaffery, 2014). Despite the popularity of wearable activity tracker usage, more than 50% of consumers no longer use their device and a third stop using their device after just six months of purchase (Ledger & McCaffery, 2014). Li, Dey, and Forlezzi (2010) proposed that tracker users progress through five stages of self-quantification. These stages are preparation, collection, integration, reflection, and action.  The authors advised that a device must allow the user control in each stage and facilitate the user experience to ensure long-term usage and user success. In other words, the device must adapt to the user’s specific needs by collecting the appropriate information and providing feedback that motivates the user until goals are met. This process starts over when new goals are set by the user. Li et al. (2010) added that this process either becomes iterative or stops if barriers become too high. A recent study by Clawson, Pater, Miller, Mynatt, and Mamykina (2015) shared the sentiments of Li et al. (2010). Clawson et al. (2015) investigated the many reasons former tracker users listed their device for sale on Craigslist. Though no clear conclusions were made, the authors remarked that the following reasons were the most common: changing goals, initiation of new exercise programs, changes in domestic situations, socially-driven device switches, and device limitations. Yet, the Clawson et al. (2015) did not indict activity trackers for the failure of long-term usage by consumers. Instead, the authors proposed that the creators of the technology lack the full understanding of how personal informatics influences behavioral outcomes. Thus, the Clawson et al. (2015) concluded that the technology fell short of meeting the long-term needs of the consumer.  Other theories associated with the complexity of self-quantification models classify users based on the type of motivation and the user’s specific purpose of tracking their activities (Gimpel & Niβen, 2013; Rooksby, Rost, Morrison, & Chalmers, 2014). Yet, Jakacic et al. (2016) showed that individuals wearing activity trackers to track exercise during a weight loss program lost less weight than participants who did not use them. Participants in Jakacic et al. (2016) study were given diet and exercise recommendations, access to weekly support groups, and follow ups calls. The group that received a wearable activity tracker only lost 7.7 pounds on average compared to 13 pounds for the group that did not use a wearable device. The authors proposed that tracking exercise may have lead the group to assume that it earned the opportunity to eat more food.

      Because activity trackers attempt to affect human behavior, a number of constructs have been used to determine the efficacy of the technology. Day (2016) proposed a research model derived from established human behavior constructs to examine positive attitudes displayed by long-term Fitbit users including

  • performance expectancy, which determines the user’s perception of the usefulness of the technology;
  • effort expectancy, which determines the ease of use experienced by the user and the ubiquity of the technology;
  • social influence, which determines the influence of friends or family on usage of the activity tracker;
  • attitude, which determines the degree of motivation derived from the device;
  • and Goal Determination, which determines how a user sets goals, receives feedback, understands their progress, and how this information influences the user to set longer term goals.

      Using the previous constructs and conducting interviews,  Day (2016) revealed how some Fitbit users change the way they use their device. While 93% of participants claimed they intended to use a device in the future, users’ attitudes toward how the device shaped behavior were mixed. For users utilizing their device from one to three months, 63% indicated that usage had not changed and, for some, excitement abated. For longer term users (three to twelve months), usage shifted towards maintaining initial goals and setting new ones. It appears that these results may support Li et al. (2010) stages of useage model, yet, limitations such as self-reported usage and qualitative analysis make it difficult to draw clear associations.

      To further understand human behavior regarding self-quantification, Etkin (2016) examined how individuals felt when their activity was being measured. Etkin (2016) found that when students’ activity level was measured, students increased step count, yet, enjoyed the activity less than when the activity was not measured. These findings are supported by Kruglanski, Friedman, and Zeevi (1971) and Laran and Janiszewski (2010) establishing that activities are extrinsically motivated when output is quantified and that self-quantification tends to make common activities feel like work (Kruglanski, Friedman & Zeevi, 1971; Laran & Janiszewski, 2010). Etkin (2016) displayed that self-quantification could ultimately lead to a reduction in the activity measured and a decrease in subjective well-being. However, it should be noted that some of Etkin’s (2016) results suggested that when measurement is associated with an individual’s goal, it may not decrease enjoyment, especially when measurement is an integral part of the activity.

Activity Tracking Technology Shows Promise for the Future

      The literature regarding the influence of activity trackers on long-term activity behavior appears inconclusive. However, there is strong evidence to suggest activity trackers are consequential for helping users change short term activity levels. For example, Bravada et al. (2007) found pedometers increased activity levels by over 26% when users set a step goal. While several other studies corroborate these findings, researchers have recognized the need for long term studies to evaluate the effectiveness of prevailing activity tracking technology (Barua, Kay, & Paris, 2013). Thus far, research has focused on how the behaviors of adopters have changed due to tracking with various devices (Day, 2016; Fritz, Huang, Murphy, & Zimmermann, 2014). For example, Fritz et al. (2014) and Day (2016) learned that users made durable behavioral changes, especially individuals that used persuasive technology, such as Fitbit.

      Though these studies are limited by sample size and provide only contextual information, activity tracker developers are using similar research to improve tracking technologies. Endeavour Partners, a think tank for digital business and technologies, recognizes the need for tracking devices to manipulate the governing mechanisms of human behavior. Endeavour Partners members, Ledger and McCaffery (2014), have proposed that a wearable activity tracker must involve three factors to ensure long-term engagement and the success of the device in the market. These factors include habit formation, social motivation, and goal reinforcement. Habit formation includes cuing a behavior and rewarding the user after completion. Social motivation involves sharing goals and building a social support network. Goal reinforcement requires continuous objective feedback about progress. Currently, there are several devices on the market incorporating all three components including Polar Flow and Nike FuelBand. Interview studies by Day (2016) and Fritz et al. (2014) support the notion that long-term activity tracker users utilize social media to stay motivated. Bandura (1986) introduced social cognitive theory, a behavioral construct proposing that people make decisions about their behaviors by evaluating the behaviors of others. This theory continues to guide activity tracker makers toward improvement of social media interfaces (Ledger & McCaffery, 2014).

Wearable Activity Trackers of Tomorrow

      Despite short-term improvement in activity levels, the current state of wearable activity tracker technology appears to lack the complexity to align with factors influencing long-term behavioral change. Activity tracker makers are using behavioral constructs such as social cognitive theory and goal setting theory to improve the effectiveness of their devices (Ledger & McCaffery, 2014). Yet, these constructs do not fully explain why many people abandon a device within the first few months of purchase. One proposed solution involves the stages of change model. This theory suggests that people move through five stages change, from not perceiving the need for change to recognizing a problem, developing a solution, taking action, and maintaining progress (Velicer, Prochaska, Fava, Norman, & Redding, 1998). It is important to consider that in order for tracking devices to be effective, an individual must have a strong belief in their need for change. For progress to continue, so must the belief that progress is needed. Activity tracking devices must be part of the solution and the maintenance of progress. Research has shown that activity tracking devices can become a barrier if the device is too difficult to use, does not help the user set new goals, does not provide appropriate feedback, or does not allow social sharing as users advance their progress (Clawson, Pater, Miller, Mynatt, & Mamykina, 2015; Day, 2016; Fritz, Huang, Murphy, & Zimmermann, 2014; Li, Dey, & Forlezzi, 2010).  More research is needed to understand correlations between readiness for change and success for activity tracker users.

      Goal setting and social support are important determinants of long term use which is mediated by self-efficacy (Locke & Latham, 2002). The wearable activity tracker of tomorrow should be able to predict the behavior of the individual and help set and track new goals based on user information such as personality traits, special interests, geographical location, and social network. It appears that the more autonomously the device operates, the more effective it may be to help users achieve long-term behavioral change.



Adam Eckart, MS, CSCS, FDN-P

Co-Founder, Critical MASS Training Systems

Leave a Reply