TASK-TECHNOLOGY FIT - Human-Computer Interaction and Management Information Systems: Foundations - page 180

How Does Technology Affect Task Performance?
Technology can be thought of as affecting task performance in two ways. The first is by improv-
ing some non-human portion of the task, even while the task appears the same to the task doer
(same process, same inputs and outputs, same interface). For example, the new technology might
return an answer in less time, or give a higher-quality answer. In this case, the technology would
not change the sequence of actions taken by the task doer (what might be called the execution
sequence), but would result in better performance overall.
The second way technology affects task performance is by changing the possible execution
sequences presented to the task doer. Each technology applied to a task enables a set of available
execution sequences that could be used to carry out the task. These possible execution sequences
are the different approaches or strategies that can be used for addressing the underlying task (e.g.,
Newell and Simon, 1972; Payne et al., 1996; Todd and Benbasat, 1999). To illustrate this point
with an example, consider a task doer who is asked to report “sales minus returns and discounts”
(an aggregate of several pieces of information from the organization's database). Even holding the
technology constant, a task doer might opt either to create several simple queries, each one return-
ing a small portion of the desired information (one possible execution sequence), to create fewer
and more sophisticated queries, each one returning a greater portion of the desired information
(a second possible execution sequence), or to guess at the approximate answer based on the task
doer's general knowledge (a third possible execution sequence).
Parenthetically, we note that one could argue that there is a third way a technology could affect
performance: It could completely carry out the task and eliminate the need for the task doer.
However, there is always some other higher-level task doer in the background, carrying out some
higher-level task and deciding to use the technology as opposed to a human. The technology has
changed the task as presented to this latter task doer.
Changing the available technology generally means adding new possible execution sequences,
removing old execution sequences, or changing the characteristics of existing execution sequences.
Continuing with the example above, if the required “sales minus returns and discounts” information
spanned several divisions and the available databases were not semantically integrated across the
divisions, it would not be an option to use the execution sequence of retrieving all sales data for all
divisions with one query, since fields might be defined differently for the different divisions. Rather,
separate queries would be needed for each semantically distinct database. Switching to a semanti-
cally integrated database would change the execution sequences available to task doers, since now
the entire task could be accomplished with a single more sophisticated query. Even so, there would
still be several possible execution sequences beside the single large query. Regardless of which tech-
nology is available or which execution sequence is chosen, the underlying task of retrieving infor-
mation on sales minus returns and discounts across divisions would remain the same.
This general relationship is shown in Figure 9.4. A change from technology 1 to technology 2
changes the possible execution sequences presented to the task doer for carrying out the same
underlying task. In the figure, execution sequences A, B, and C are available with technology 1,
while execution sequences C, D, and E are available with technology 2.
Presumably task doers choose a particular execution sequence based on their assessment of its
“attractiveness” to them. In other words, they might take into account cognitive cost benefit consid-
erations (Payne, 1982; Smith et al., 1982; Creyer et al., 1990) such as differences in expected per-
formance and differences in effort required, as well as other non-cognitive costs such as the
mechanical difficulties of walking to a different building to access a different database or bureaucratic
difficulties of getting access to data. Finally, they may be affected by habit, social norms, or