AWS Community Day CPH - Three problems of Terraform
Back to the App: The Costs of Mobile Application Interruptions
1. Back to the App:
The Costs of Mobile Application Interruptions
Luis A. Leiva1, Matthias Böhmer2, Sven Gehring2, Antonio Krüger2
1 Universitat Politècnica de València // Institut Tecnològic d'Informàtica
2 DFKI // Ubiquitous Media Technology Lab
4. What are the costs of app interruptions?
- When switching between applications?
- When being interrupted by a phone call?
5. In this note...
‣ We present a framework for detecting interruptions from
mobile app usage logs
‣ We investigate the costs and frequency of app interruptions
- Intentional app-switching behavior
- Interruptions by incoming phone calls
‣ We discuss strategies for mitigating interruptions
6. Study Dataset
3.6K unique users
15.7K different apps
1.5 years
5.5M data samples
7. Detecting Interruptions
... app use ...
time
‣ App use contained in log file
- Timestamp of start of use
- Duration of use
‣ We keep track of usage time of apps
8. Detecting Interruptions
... app use interruption app use cont‘d ...
time
‣ Mining for interruptions
- Internal interruptions: another application
- External interruptions: incoming phone call
‣ We know the duration of the interruption
9. Measuring Interruptions
Runtime of non-interrupted app app use
time
Duration of interrupted app use app use app use cont‘d
time
Duration of interruption interruption
time
app use overhead
Duration of overhead
app use app use cont‘d
time
10. Study Results
‣ Internal interruptions are more frequent and involve more apps
external internal
Interruption data samples 776,922 970,543
Interrupted users 1,929 [1,676] 2,926 [2,609]
Interrupted Applications 1,373 [487] 4,626 [1,043]
[paired samples denoted in brackets]
‣ Our analysis is based on paired cases
11. Occurrences of App Interruptions
external internal
Daily interruptions (% usage) 3.2 (2.2) 8.3 (5.3)
per user
Interrupted applications 3.3 (2.6) 8.7 (7.2)
mean (SD)
‣ Interruptions do not happen as often as expected
- 3% of app use is interrupted by phone calls (external)
- 8% of app use is interrupted by app switching (internal)
12. Impact of App Interruptions
external internal
Regular app runtime (s) 24.8 (31.8) 18.9 (24.4)
Runtime when interrupted (s) 107.1 (121.1) 87.9 (75.5)
per app
Interruption duration (s) 12.5 (8.1) 23.7 (19.3)
Overhead duration (s) 43.2 (65.9) 34.4 (40.7)
mean (SD)
‣ If interruptions happen, resumption costs may be exceedingly high
13. Design Implications
‣ Help to regain context of deferred application when resumed,
inspired by previous approaches in field of desktop interruptions
‣ Preventive: Prepare user for being interrupted
- Postponing phone call for a few milliseconds
- Gradually overlaying phone call instead of full-screen popup
‣ Curative: Guide user back into task
- Replaying last milliseconds of UI interactions
- Leave visual cues on screen to hint where to continue
- Vanish interruptive application into point of last focus
14. Conclusion
‣ Main Findings:
- Application interruptions are happening rarely
- If interruptions happen, they are costly
- App-switching behavior is not as often as assumed
- Call interruptions add significant overhead to app use
‣ Future Work:
- Enriching data collection process with experience sampling
methods to get more insights into context of interruptions
- Model for predicting interruption overhead to inform the
means of curation
‣ Thank you :-)