THE UNIVERSITY of NORTH CAROLINA at CHAPEL HILL Improving IM Collaboration in the Workplace Kirstin Williams COMP 790-063.

THE UNIVERSITY of NORTH CAROLINA at CHAPEL HILL

Improving IM Collaboration in the Workplace

Kirstin WilliamsCOMP 790-063

THE UNIVERSITY of NORTH CAROLINA at CHAPEL HILL

FTF Interactions

+ Visual awareness of target’s presence

+ Visual awareness of target’s workload

+ Shared sense of value of interruption

+ Immediate sense of response urgency and time demand

– Can be intrusive

– Unwanted prolonged discussions

– No multitasking

– Greater time demands

THE UNIVERSITY of NORTH CAROLINA at CHAPEL HILL

IM Collaboration

+ Less intrusive

+ Multitasking

+ Quick and spontaneous with little additional effort

+ 140M plus users

– Social Play?

THE UNIVERSITY of NORTH CAROLINA at CHAPEL HILL

IM in the Workplace

Are people actually using IM for work??

THE UNIVERSITY of NORTH CAROLINA at CHAPEL HILL

AT&T Research Study

● 437 participants– Used client for at least a week

● Spanned 16 months● Chat content analysis

– Most users at work

THE UNIVERSITY of NORTH CAROLINA at CHAPEL HILL

Nature of IM at Work

• Simple info exchange– 91.4% work-related

• Scheduling– 85.7% work-related

• No Response– Reminders

• Work is largest category– 60% of conversations Figure published in “The Character, Functions, and Styles of Instant Messaging in the Workplace”

THE UNIVERSITY of NORTH CAROLINA at CHAPEL HILL

Nature of IM at Work

• Work– Work Talk

– Work-Related Talk

– Doing Work

Figures published in “The Character, Functions, and Styles of Instant Messaging in the Workplace”

THE UNIVERSITY of NORTH CAROLINA at CHAPEL HILL

IM in the Workplace

Are people taking advantage of the ability to perform multiple tasks?

THE UNIVERSITY of NORTH CAROLINA at CHAPEL HILL

Multitasking

● 86% of all users multitasked● Analyzed focus switching

– Once every 70 seconds!!

● Activity not from multiple conversations– 77% never had overlapping conversations

THE UNIVERSITY of NORTH CAROLINA at CHAPEL HILL

IM in the Workplace

Does IM create unwanted interruptions and affect productivity?

THE UNIVERSITY of NORTH CAROLINA at CHAPEL HILL

Effects of Interruption

• Helper– Saves jumpers

– Sees image

• Seeker– Sees pieces of image

– Asks 20 ?s

• More jumpers saved with less interruptions

THE UNIVERSITY of NORTH CAROLINA at CHAPEL HILL

IM in the Workplace

How can we control IM interruptions?

THE UNIVERSITY of NORTH CAROLINA at CHAPEL HILL

Current Control Methods

• Toggle Device Status– May forget to turn off– May forget to turn back on

• Screen Incoming Messages– Negative social dynamic– Time spent assessing each

message• “Away” Status

– Forget to set– Miss valuable messages

• Proactive Management– Forget to switch between

accounts

THE UNIVERSITY of NORTH CAROLINA at CHAPEL HILL

Traditional IM Client

● Sounds for incoming/outgoing message

● Indicator of away/available● Hover produces idle time● Focus indicator

THE UNIVERSITY of NORTH CAROLINA at CHAPEL HILL

AT&T’s Hubbub

• Individual Sound ID• Location of user

THE UNIVERSITY of NORTH CAROLINA at CHAPEL HILL

IM in the Workplace

Can we improve IM to make a seeker aware of a target’s presence?

THE UNIVERSITY of NORTH CAROLINA at CHAPEL HILL

Awarenex

• Logs location

• Tracks online calendar

• Observes e-mail activity

• Updates every minute

• Presents seeker with visualization of helper’s activity

THE UNIVERSITY of NORTH CAROLINA at CHAPEL HILL

Awarenex Actogram

• Displays start, stop, and lunch trends• Can display daily trends and reoccurring meetings• Makes a seeker aware of the target’s trends in presence• Can augment with Horvitz’s work on return prediction• Target may be reachable, but not receptive…

Figure published in “Work Rhythms: Analyzing Visualizations of Awareness Histories of Distributed Groups”

THE UNIVERSITY of NORTH CAROLINA at CHAPEL HILL

Lilsys

• Presence data

• Availability detection– Sound

– Phone

– Door

• Vague status displays– Plausible deniability

• No feedback about willingness to receive a disruption…

THE UNIVERSITY of NORTH CAROLINA at CHAPEL HILL

Notification Platform

• Seeker provided with video sequence of target’s activity• Target can annotate video with busy times• Compromising privacy!

THE UNIVERSITY of NORTH CAROLINA at CHAPEL HILL

BusyBody Status

• Training phase with “busy palette”

THE UNIVERSITY of NORTH CAROLINA at CHAPEL HILL

BusyBody COI

• Assess probability helper is in state S

• Train Bayesian Network

• Build case library

• Evaluate Expected Cost of Interruption

• Display cost to seeker

• Start to infer willingness to receive an interruption

THE UNIVERSITY of NORTH CAROLINA at CHAPEL HILL

IM in the Workplace

Will a seeker pay attention to the information presented to them in an awareness display about the target’s presence and availability?

THE UNIVERSITY of NORTH CAROLINA at CHAPEL HILL

Motivation to Heed Awareness Display

• Awareness Display– None

– Abstract

– Full

• Motivation– Team

– Individual

THE UNIVERSITY of NORTH CAROLINA at CHAPEL HILL

Performance Improved!

• Helper saves more jumpers– Seeker more motivated

– Seeker more aware

– Helper given control

THE UNIVERSITY of NORTH CAROLINA at CHAPEL HILL

IM in the Workplace

How can we enable the target to help their performance even when seeker disregards awareness displays?

THE UNIVERSITY of NORTH CAROLINA at CHAPEL HILL

QnA

• Relates to Winograd’s Action Workflow

• Question and Answers are adjacency pairs

• Questions often require a quick response

• When you send a question, you are often awaiting a response

• Help target filter messages by alerting them of questions and answers

THE UNIVERSITY of NORTH CAROLINA at CHAPEL HILL

QnA

• Question matched with strings– Use ExpectingResponse if target just sent a

question

– Use IncomingQuestion if target receives a question

• Response if adjacent to question– Use IncomingResponse if a message is

received and ExpectingResponse is set

• Notify user

• Can extend this concept with Horvitz’s prioritizing work

THE UNIVERSITY of NORTH CAROLINA at CHAPEL HILL

The Ultimate IM

• Detect presence

• Monitor our social interaction

• Assess our cost of interruption

• Enable us as a target

THE UNIVERSITY of NORTH CAROLINA at CHAPEL HILL

Summary

• IM in the workplace

• IM as a distraction

• Traditional disruption control methods

• New approaches to control messages

• Seeker’s motivation

• The Ultimate IM Tool

THE UNIVERSITY of NORTH CAROLINA at CHAPEL HILL

Questions??

THE UNIVERSITY of NORTH CAROLINA at CHAPEL HILL

IM Statistics

• Heavy users– More short turns

– More threading

– Fewer conversation closings

– More likely to allow interruptions

• Light users– Longer duration

– Less threading

– Quicker 1st message response time

Figure published in “The Character, Functions, and Styles of Instant Messaging in the Workplace”

THE UNIVERSITY of NORTH CAROLINA at CHAPEL HILL

Controlling Interruption

• Helper saves more jumpers– Seeker more motivated

– Seeker more aware

– Helper given control

• Display results– No display = no awareness

– Abstract Display = good awareness

– Full Display = too busy