You’ve got email! Does it really matter to process emails now or later?

Abstract

Email consumes as much as a quarter of knowledge workers’ time in organizations today. Almost a necessity for communication, email does interrupt a worker’s other main tasks and ultimately leads to information overload. Though issues such as spam, email filtering and archiving have received much attention from industry and academia, the critical problem of the timing of email processing has not been studied much. It is common for many knowledge workers to check and respond to their email almost continuously. Though some emails may require very quick responses, checking emails almost continuously may lead to interruptions in regular knowledge work. Managing email processing can make a significant difference in an organization’s productivity. Previous research on this topic suggests that perhaps the best way to minimize the effect of interruptions is to process email frequently for example, every 45 min. In this study, we focus on studying email response timing approaches to optimize the communication times and yet reduce the interruptive effects. We investigate previous recommendations by performing a two-phase study involving rigorous simulation experiments. Models were developed for identifying efficient and effective email processing policies by comparing various ways to reduce interruptions for different types of knowledge workers. In contrast to earlier research findings, results indicate that significant productivity improvements could be achieved through the use of some email processing policies while helping attain a balance between email response time and task completion time. Findings also suggest that the best policy may be to respond to email two to four times a day instead of every 45 min or continuously, as is common with many knowledge workers. We conclude by presenting many research opportunities for analytical and organizational IS researchers.

Appendices

Appendix I

Table A.1 Email types, processing time of email and primary task

Appendix II. The mathematical model

Table A.2 Parameters used in experiment phase I

2.1 Notation

i	type of email message	i = 1,2..I
		{i = 1 for SPAM,
		i = 2 for Irrelevant,
		i = 3 for Read only,
		i = 4 for Reply}
j	urgency (priority) of message	j = 1,2..J
		{j = 1 for Urgent (Priority-1),
		j = 2 for within Business Day (Priority-2),
		j = 3 for within 1 week (Priority-3),
		j = 4 for Irrelevant}
k	category of processing need	k = 1,2..K
		{k = 1 for <1 min,
		k = 2 for 1–10 min,
		k = 3 for >10 min}
d	day	d = 1,2..D
s	sequence number	s = 1,2..S
t	time period of day	t = 1,2..T
		{t = 1 for 8:00 a.m. until 10:00 a.m.,
		t = 2 for 10:00 a.m. until 12:00 p.m.,
		t = 3 for 12:00 p.m. until 2:00 p.m.,
		t = 4 for 2:00 p.m. until 4:00 p.m.,
		t = 5 for 4:00 p.m. until 6:00 p.m.,
		t = 6 for 6:00 p.m. until 8:00 a.m.}
X	email processing strategy employed
λ ijkt	arrival rate of email messages of type i, urgency j, processing need k, occurring during time period t
P kds	random variable that represents the processing time required for email of type k, occurring on day d, having sequence number s
p kds	E(P kds )
${f_{kds}}^P\left( {\hbox{x}} \right)$	probability density function (pdf) of P kds
R ds	random variable that represents the resumption lag occurring on day d, sequence number s
r ds	E(R ds )
${f_{ds}}^R\left( {\hbox{x}} \right)$	pdf of R ds
L ds	random variable that represents the interruption lag occurring on day d, sequence number s
l ds	E(L ds )
${f_{ds}}^L\left( {\hbox{x}} \right)$	pdf of L ds
Q d	random variable that represents the threshold of productive work (email processing and primary work) to be completed on day d
q d	E(Q d )
${f_d}^Q\left( {\hbox{x}} \right)$	pdf of Q d
Wq js	email’s wait in the queue (time spent waiting for the knowledge worker’s attention) for email of urgency j having sequence number s
Ws js	email’s wait in the system (email resolution time) for email of urgency j having sequence number s
	$W{s_{js}} = W{q_{js}} + {P_{kds}}$
$\overline W {s_{js}}$	mean email resolution time for email of urgency j
	$\overline W {s_j} = {\sum_s}W{s_{js}}/S$
Y d	total email processing occurring on day d
	${Y_d} = {\sum_k}{\sum_s}{P_{kds}}$
Z d	total amount of primary work completed on day d
G d	total lag time occurring on day d
	${G_{\rm{d}}} = {\sum_{\rm{s}}}{L_{ds}} + {\sum_{\rm{s}}}{R_{ds}}$
H d	total hours worked by the knowledge worker on day d
	${H_d} = {Y_d} + {Z_d} + {G_d}$
$\overline H$	mean hours worked by the knowledge worker
	${\sum_d}{H_d}/D$
E d	knowledge worker efficiency occurring on day d
	${E_d} = \left( {{Y_d} + {Z_d}} \right)/{H_d}$
$\overline E$	mean knowledge worker efficiency
	${\sum_d}{E_d}/D$
Q d	threshold of productive work (email processing and primary work) to be completed on day d
	${Q_d} < = {Y_d} + {Z_d}$