We studied what notifications 15 mobile phone users received during one week, how they attended to them, and how the notifications affected them.
Highlights of our findings are:
* our participants dealt with more than 60 notifications per day, mostly from messengers, SMS, and email.
* Notifications from messengers are largely attended within minutes.
* Notifications from email clients and messengers show significant correlations with people's emotions: emails are associated with interruption and stress. Messages are related to feeling overwhelming but also connected to others.
The paper received best-paper award at ACM MobileHCI '14, which was held in September 2014 in Toronto, Canada.
The full paper is available here: http://pielot.org/pubs/Pielot2014-MobileHCI-Notifications.pdf
word2vec, node2vec, graph2vec, X2vec: Towards a Theory of Vector Embeddings o...
An In-Situ Study of Mobile Phone Notifications
1. Research
An In-Situ Study of
Mobile Phone Notifications
Martin Pielot
Telefónica
Research
Rodrigo de
Oliveira
Google Inc.*
Karen Church
Yahoo Inc.*
* this work was done while working in Telefónica Research
ACM MobileHCI ’14, Sep 2014, Toronto, Canada
(Slideshare edit)
2. Motivation
On notification, people often interrupt
current activities to check notification
…
We aimed to better understand why
Practice Cell Phone Etiquette Step 4 by Wikiphoto, via WikiHow, CC BY-NC-SA 3.0
3. In-Situ Study on Mobile Phone Notifications
(1) Quantitative analysis of notifications
and attentiveness
(2) Fusion with daily subjective feedback
on emotional impact
(3) Implications for strategies to improve
notification handling
13. E-mail
Other apps
(updates, games,
etc.)
Diary Survey – Daily for 7 Days
*
5. How much did the majority of these messages/notifications interrupt you from
your daily tasks?
Very
frequently
distracted
Frequently
distracted
Occasionally
distracted
Rarely
distracted
Very rarely or
never
distracted
NA
Mobile instant
messaging apps
(WhatsApp, SMS,
Skype, etc.)
Social Network apps
(Facebook, Twitter,
LinkedIn, etc.)
E-mail
Other apps (updates,
games, etc.)
Please tell us more about the application or notifications that cause the most distraction
14. E-mail
Other apps
(updates, games,
etc.)
Diary Survey – Daily for 7 Days
*
5. How much did the majority of these messages/notifications interrupt you from
your daily tasks?
Very
frequently
distracted
Frequently
distracted
Occasionally
distracted
Rarely
distracted
Very rarely or
never
distracted
NA
Mobile instant
messaging apps
(WhatsApp, SMS,
Skype, etc.)
Social Network apps
(Facebook, Twitter,
LinkedIn, etc.)
E-mail
Other apps (updates,
games, etc.)
Please tell us more about the application or notifications that cause the most distraction
15. E-mail
Other apps
(updates, games,
etc.)
Diary Survey – Daily for 7 Days
*
5. How much did the majority of these messages/notifications interrupt you from
your daily tasks?
Very
frequently
distracted
Frequently
distracted
Occasionally
distracted
Rarely
distracted
Very rarely or
never
distracted
NA
Mobile instant
messaging apps
(WhatsApp, SMS,
Skype, etc.)
Social Network apps
(Facebook, Twitter,
LinkedIn, etc.)
E-mail
Other apps (updates,
games, etc.)
Please tell us more about the application or notifications that cause the most distraction
16. E-mail
Other apps
(updates, games,
etc.)
Diary Survey – Daily for 7 Days
*
5. How much did the majority of these messages/notifications interrupt you from
your daily tasks?
Very
frequently
distracted
Frequently
distracted
Occasionally
distracted
Rarely
distracted
Very rarely or
never
distracted
NA
Mobile instant
messaging apps
(WhatsApp, SMS,
Skype, etc.)
Social Network apps
(Facebook, Twitter,
LinkedIn, etc.)
E-mail
Other apps (updates,
games, etc.)
Please tell us more about the application or notifications that cause the most distraction
17. E-mail
Other apps
(updates, games,
etc.)
Diary Survey – Daily for 7 Days
*
5. How much did the majority of these messages/notifications interrupt you from
your daily tasks?
Very
frequently
distracted
Frequently
distracted
Occasionally
distracted
Rarely
distracted
Very rarely or
never
distracted
NA
Morning of next day – to capture
late-night notifications
Mobile instant
messaging apps
(WhatsApp, SMS,
Skype, etc.)
Social Network apps
(Facebook, Twitter,
LinkedIn, etc.)
E-mail
No experience sampling – to not
add new notifications
Other apps (updates,
games, etc.)
Please tell us more about the application or notifications that cause the most distraction
19. 6854 notifications in 7 days
messengers
49%
email
32%
social
networks
4%
other
15%
Participants did not have
to alter notification
settings
• 2 of 15 did not receive
email notifications
• 7 of 15 did not receive
social network
notifications
20. 140
120
100
80
60
40
20
0
Notifications per Day
Messengers Email Social Networks Other
63.5 notifications per day (Median)
Weekday Day in Weekend
21. 100%
80%
60%
40%
20%
0%
MIM Email Social Other
User Responses
How many notifications have you received
on your mobile phone?
don't know
way more
more
the usual
fewer
way fewer
22. 350
300
250
200
150
100
50
0
Messages: breaks and evening
Emails: work hours
Messengers Email Social Other
0 4 8 12 16 20 24
Number of Notifications
Hour of the Day
24. Attentiveness - Expectations
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
With respect to the people you communicate most with
How fast do YOU typically respond?
Within a few days
Within 24 hours
Within a few hours
Within an hour
Within a few minutes
Immediatey
25. Attentiveness - Expectations
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
With respect to the people you communicate most with
How fast do THEY typically respond?
Within a few days
Within 24 hours
Within a few hours
Within an hour
Within a few minutes
Immediatey
26. Actual Attentiveness
220
200
180
160
140
120
100
80
60
40
20
0
Time until attended (min)
Messengers Email Social Networks Other
From: 3.5 min (messages on weekends)
To: 27.7 min (emails on weekends)
Weekday Day in Weekend
27. Ringer Mode
Vibration mode faster notification attendance
No difference between silent and normal mode
29. How frequently felt distracted?
100%
80%
60%
40%
20%
0%
Msgs Email Social Other
User Responses
NA
never
rarely
occasionally
frequenty
very frequenty
30. WhatsApp was distracting a bit in the
evening when I was out at the opera and
then for birthday drinks with friends. It
was not important to answer, but still a
matter of politeness I had to answer in a
reasonable time span.
31. Please tell us more about the application or notifications that cause the most distraction
Daily Diary
*
6. With respect to the messages/notifications you received on your mobile phone
on Tuesday March 5th, 2013.
Strongly disagree Disagree
Neither agree nor
disagree
Agree Strongly agree
I received a lot of
notifications
I felt overwhelmed
by the amount of
notifications
I felt stressed by
3/13/13 [SURVEY PREVIEW MODE] Push Message/Notification Study - Diary Survey
the notifications
www.surveymonkey.com/s.aspx?PREVIEW_MODE=DO_NOT_USE_THIS_LINK_FOR_COLLECTION&sm=%2ffVLBq9vJAztCbrM44%2bMhq0hVuR… 2/4
I felt interrupted by
the notifications
I felt annoyed by
the notifications
I felt connected
with others.
I felt that others are
thinking about me.
32. more
emails
Keri J. Email customized icon. via Flickr, Mar 3, 2010 (CC BY-ND
2.0)
33. Increased feelings of
overwhelmed
interrupted
stressed
annoyed
LaurMG. A frustrated man sitting at a desk. via Wikipedia, May 24, 2011 (CC BY-SA
3.0).
34. Email notifications are usually from
exchange server, so they are work
related. It means if I am not near my
computer, I should get to my computer
and reply
36. Increased feelings of
having to deal with
a lot of messages
and being
overwhelmed
Jhaymesisviphotography. Texting. via Flickr, Dec 10, 2011 (CC
BY 2.0)
37. WhatsApp users don’t have option to
stop other side of users knowing if you
have already read or not [...] people
tend to expect immediate answers.
38. BUT also
feeling connected
with others
Brinks Alo. Week 11/52: How Sweet It is To Be Loved by You. via Flickr, Oct 3, 2010 (CC BY-ND
2.0)
39. I was talking to my sister. She was
sad [and] I wanted to respond
immediately to comfort her.
41. Reducing number of interruptions
Rosenthal et al. 2011
e.g. by muting the ringer when
notification is predicted to be
“unimportant” or context “unsuitable”
…
Not advisable for communication
services as users will check their
phones anyway
42. Defer notification ‘til opportune moment
Horvitz et al. 2005, Iqbal and Bailey 2008, Fischer et al. 2011, Leiva et al.
2012
e.g. detecting break-points or wait until
activities are finished
Advisable for work/emails, but difficult
in case of personal communication
43. Communicating non-availability
Harper and Taylor 2009, Reynolds et al. 2013, Pielot et al. 2014
e.g. by communicating context or
predicting attentiveness
Advisable for personal communication,
to manage expectations
44. Take Aways
63.5 notifications per day, mostly from
communication apps
People attend fast to notifications, in particular
from messengers (3.5 min median) … even if
the phone is in silent mode
Notifications, in particular related to work,
correlate with negative emotions
However, personal communication also
correlates with feeling connected
Implications depend on type of notification
Other/Non-important = mute ringer
Work/Email = deliver at opportune moment
Personal/Messages= communicate availability
An In-Situ Study
of
Mobile Phone
Notifications
Martin
Pielot
Rodrigo
de Oliveira
Karen
Church
Telefónica
Research
* this work was done while
working in Telefónica
Research
ACM MobileHCI ’14, Sep 2014, Toronto,
Canada
Google
Inc.*
Yahoo
Inc.*
Editor's Notes
Owner picks phone up – check what cause the notif
Urge so strong – willing to interrupt – driving or conversations
Mix of multiple choice and free text
Send on morning of next day, because notifications arrive until late into the night
No ESM, because would add new notifications = bias the phenomenon that we intended to study
Mix of multiple choice and free text
Send on morning of next day, because notifications arrive until late into the night
No ESM, because would add new notifications = bias the phenomenon that we intended to study
Mix of multiple choice and free text
Send on morning of next day, because notifications arrive until late into the night
No ESM, because would add new notifications = bias the phenomenon that we intended to study
Mix of multiple choice and free text
Send on morning of next day, because notifications arrive until late into the night
No ESM, because would add new notifications = bias the phenomenon that we intended to study
Mix of multiple choice and free text
Send on morning of next day, because notifications arrive until late into the night
No ESM, because would add new notifications = bias the phenomenon that we intended to study
Notifications are mostly from communication services / social
So, while we have to agree that message notifications frequently distract …
We believe – on the basis of our findings – that message notifications should still be delivered immediately
People have developed very high expectations towards responsiveness in messenger communication,
So, we believe, our mission should be to help people in managing and reducing those expectations
We have to remind people that it is not always possible to comply to those expectations