Enticing Notification Text & the Impact on Engagement Kieran - - PowerPoint PPT Presentation

▶
enticing notification text the impact on engagement
SMART_READER_LITE
LIVE PREVIEW

Enticing Notification Text & the Impact on Engagement Kieran - - PowerPoint PPT Presentation

Mar 21, 2024 122 likes •363 views

UbiTtention 2020 Enticing Notification Text & the Impact on Engagement Kieran Fraser & Owen Conlan Overview 1. Introduction 2. Related Work 3. Information-Gap Theory 4. Capturing Mobile Push-notifications 5. Enticing

slide-1
SLIDE 1

Enticing Notification Text & the Impact on Engagement

Kieran Fraser & Owen Conlan UbiTtention 2020

slide-2
SLIDE 2
  • 1. Introduction
  • 2. Related Work
  • 3. Information-Gap Theory
  • 4. Capturing Mobile Push-notifications
  • 5. Enticing Push-notifications
  • 6. Push-notification Delivery Agents
  • 7. Capturing Web Push-notifications
  • 8. Enticing Titles vs Summaries
  • 9. Limitations & Future Work
  • 10. Final Thoughts

Overview

1

slide-3
SLIDE 3

Introduction

2

slide-4
SLIDE 4

Related Work

3

Growing number of notifications pushed at users (Pielot, M. et al, 2014). Large no. of incoming notifications = negative user emotions (Sahami Shirazi, A. et al, 2014). Notification delivery not smart (Mehrotra, A. et al, 2016). Unnecessary notifications may dramatically decrease productivity (Iqbal, S.

  • T. et al, 2010).
slide-5
SLIDE 5

Information-Gap Theory

4

Previous Knowledge or Experience “Zone of Curiosity” Something New

“Curiosity happens when we feel a gap in our knowledge” - Loewenstein Curiosity peaks when a small amount of information is revealed, but key parts are withheld

slide-6
SLIDE 6

Information-Gap Theory

5

John Caples, 1926 WordStream, 2017

slide-7
SLIDE 7

Capturing Mobile Push-notifications

6

WeAreUs Android App

 Experience Sampling Method  Moments of notification interest, moments of phone usage interest  Privacy-preserving text feature

slide-8
SLIDE 8

Capturing Mobile Push-notifications

7

18 participants

  • ver 3 months

35,314 notifications logged 309 questionnaire responses 7,123 smartphone usage logs

slide-9
SLIDE 9

Enticing Push-notifications

8

slide-10
SLIDE 10

Enticing Push-notifications

9

Neutral Notification Unigrams Enticing Notification Unigrams

slide-11
SLIDE 11

Enticing Push-notifications

10

slide-12
SLIDE 12

Enticing Push-notifications

11

slide-13
SLIDE 13

Enticing Push-notifications

12

Exact Sign Test: p = .012

48.44 67.51 10 20 30 40 50 60 70 80 Neutral Enticing CTR (%)

Enticing Vs Neutral Notification Click-Through Rate

slide-14
SLIDE 14

Push-notification Delivery Agents

13

OpenAI Gym Open source toolkit for “developing and comparing reinforcement learning algorithms” 1 Gym-Push Custom OpenAI Gym environment simulating push-notification overload

  • n mobile device users
  • 1. https://gym.openai.com/
slide-15
SLIDE 15

Push-notification Delivery Agents

14

  • 1. https://gym.openai.com/

Gym-Push Custom OpenAI Gym environment simulating push-notification overload

  • n mobile device users

State Privacy-preserving Text Embedding Action Open / Dismiss the notification

slide-16
SLIDE 16

Push-notification Delivery Agents

15

slide-17
SLIDE 17

Push-notification Delivery Agents

16

slide-18
SLIDE 18

Capturing Web Push-notifications

17

Subscribed Topics:

  • Breaking
  • Entertainment
  • Health
  • Beauty
  • Style
  • Science
  • Sport
  • Tech
  • World

Title Content

slide-19
SLIDE 19

Capturing Web Push-notifications

18

slide-20
SLIDE 20

Enticing Titles vs Summaries

19

slide-21
SLIDE 21

Limitations & Future Work

20

Limitations

  • Granularity of notification text in WeAreUs study
  • Lack of engagement action in Google Alert data set
  • Enticing but useful, distinguishing nuance
  • Offline simulations

Future Work

  • Larger online in-the-wild study
  • Link toward inherent participant traits e.g. curiosity or

personality

  • How can we train intelligent systems to be better than us
slide-22
SLIDE 22

Final Thoughts

21

This research is supported by the ADAPT Centre for Digital Content Technology under the SFI Research Centre Program (Grant 13/RC/2106) and is co-funded under the European Regional Development Fund.

  • Click-through Rate larger for notifications

with enticing text

  • Enticement is an important feature for

predicting subscriber engagement action

  • Summarisation has potential to reduce

the influence of enticing notification text

slide-23
SLIDE 23

www.adaptcentre.ie

22

Email: kieran.fraser@adaptcentre.ie

Thank you!