IdeaWatch
RE M OTE WORK
Do We Make More
Mistakes When
Working from Home?
When the Covid-19 pandemic forced
many workers to go remote, managers
fretted about possible declines in
productivity. More than two years later,
studies suggest that their fears were
largely unfounded. But what about
remote workers’ cognitive performance?
To investigate, a research team turned
to an unusual empirical setting: profes-
sional chess.
The researchers analyzed a set of
online tournaments organized during
the pandemic by the reigning world
champion, Magnus Carlsen, which
were structured to mirror traditional
offline contests. Because most par-
ticipants had competed in at least
one recent World Chess Federation
tournament, the researchers were able
to compare individuals’ online and
in-person performances, using the AI
in a leading chess engine to assess more
than 200,000 moves and associated
errors. They found that the quality of
play was 7.5% lower, on average, when
participants competed online. Although
the researchers couldn’t determine
why that happened, they say that the
absence of peers probably contributed.
“Cognitive tasks are important in
many modern professional, managerial,
technical, and creative occupations,”
the researchers note, and so perfor-
mance in other remote settings might
likewise take a hit. However, the decline
in chess players’ performance was
steepest in the first two tournaments
22 Harvard Business Review
January–February
of the series—suggesting that people
adapt over time.
ABOUT THE RESEARCH “Cognitive
Performance in Remote Work:
Evidence from Professional Chess,” by
Steffen Künn, Christian Seel, and Dainis
Zegners (The Economic Journal, 2022)
GENDER
How Women
in the C-Suite
Boost Financial
Performance
Women now account for roughly one in
four executives (though very few CEOs)
in the S&P 500. What does that mean for
their firms’ financial performance? A
new study finds that female executives
focus on customer relationships more
than their male counterparts do—and
that bolsters long-term business results.
The researchers studied 389 publicly
traded Fortune 500 firms from 2007 to
2015. To measure female influence on
the top management team, they consid-
ered four factors: the share of positions
on the team that were held by women, the
rank of the highest female member, the
ranks of all female members, and the
number of responsibilities on their
dockets. The researchers calculated
the share of customer-oriented words
in each firm’s annual reports and used
Tobin’s q (the market value of a firm
divided by the replacement value of
its assets) to assess long-term financial
performance. The greater the female
influence, the higher the firm’s cus-
tomer orientation and Tobin’s q.
Women generally operate from a
so-called interdependent self-construal,
the researchers explain: They are more
likely than men to see things in terms of
relationships and to consider the per-
spectives of others. So when in positions
of influence in the C-suite, they often
promote strategic decisions that reflect
a higher focus on customers—which
previous research has linked to higher
Tobin’s qs.
The results varied according to
several characteristics of the firm and
the management team. The effect of
female influence was stronger when
the C-suite had a high degree of control
over strategy—for instance, when it was
relatively unburdened by regulations.
It was also stronger when the board had
THE TOGGLING TAX
Employees at three Fortune 500 companies to led among apps and websites nearly 1,200 times a day,
on average. They spent just under four hours a week—the equivalent of five working weeks annually—
reorienting themselves after each switch. “How Much Time and Energy Do We Waste To ling Between
Applications?” by Rohan Narayana Murty, Sandeep Dadlani, and Rajath B. Das

robust female representation or had
directors with marketing experience. It
was weaker when firms were contending
with unpredictable customer prefer-
ences, lots of technological change, or
strong competition, or when they had a
high degree of family ownership.
Overall, the findings challenge prior
thinking, the researchers say. “Many
studies…suggest that female executives
engage in reduced risk-taking, but cus-
tomer orientation may actually result
in female executives pursuing riskier
strategies,” they write, noting that under
Mary Barra’s leadership, GM surprised
many observers by abandoning several
models of sedans in favor of increas-
ingly popular SUVs, thus fundamentally
changing the firm’s strategic direction.
ABOUT THE RESEARCH “Customer
Orientation and Financial Performance:
Women in Top Management Teams Matter!”
by Chandra Srivastava, Saim Kashmiri,
and Vijay Mahajan (Journal of Marketing,
forthcoming)
similar ticket had been sold to another
customer for the same price; the others
were told that another customer had got-
ten a better deal. Half the participants
in each group interacted with a human
ticket agent, while the other half inter-
acted with a bot. Among those offered
the similarly priced ticket, the type of
agent made no difference to acceptance
rates. But among participants offered
the higher-priced ticket, just 19% of
those interacting with a person wanted
to buy it, while 49% of those interacting
with a bot did. When the other customer
was said to have paid a higher price, the
opposite happened: The human agent
elicited greater willingness to purchase
(89%) than the bot did (76%).
Subsequent studies explored why that
pattern occurred. People inferred that
the bots were both less selfish than their
human counterparts (when offers were
disappointing) and less benevolent
(when offers were better than expected),
and that influenced their willingness
to accept the offers. Those effects were
attenuated, however, when the bot was
anthropomorphized.
“Our results reveal that an AI ‘bad
cop’/human ‘good cop’ approach to
managing discrepant expectations
should have beneficial effects” when
communicating with consumers
about unforeseen delays or expedited
deliveries, stockouts or upgrades, and
so on, the researchers write. They add
a caveat: Although customers often
hold unrealistic expectations about a
product or service and would be well
served by inducements to accept it, “in
those instances where the worse-than-
expected offer is objectively detrimental
to consumers, the use of this approach…
raise[s] ethical concerns.”
ABOUT THE RESEARCH “Bad News?
Send an AI. Good News? Send a
Human,” by Aaron M. Garvey, TaeWoo Kim,
and Adam Duhachek (Journal of Marketing,
2022)

A RT I F I C I A L I N T E L L I G E N C E

Let a Robot Be
the “Bad Cop”
Research has shown that consumers
often would rather interact with human
agents than with automated ones. But
a new study paints a more nuanced
picture: When communicating a disap-
pointing offer, a bot gets better results.
In one experiment, participants
imagined that they were thinking of
buying a concert ticket from an online
resale agency. Some were told that a

Harvard Business Review

January–February 23