surprise, he and a colleague found that
the hiatus heuristic predicted a range
of customers’ buying practices for three
large businesses at least as well as, and
sometimes better than, two complex
models did.
In a 2008 paper, the two investigators
called the hiatus heuristic’s unexpected
power “a devastating result” for purveyors of complicated forecasting methods.
“It’s important to understand when
speedy, simple heuristics are the right
way to go,” von Wangenheim says.
Gut decisions
Situations infused with uncertainty that
require original thinking bring out the
best in heuristics, Berg suggests. Dallas
business entrepreneurs offer a prime
example. In a pressure-packed milieu
where millions of dollars ride on determinations of where to build a high-rise
office building or a new grocery store,
the entrepreneurs hitch their wagons
to heuristics.
When searching for locations suitable
for development or expansion, major
business players could scan all affordable properties in Dallas, or even the
world. Standard economic theory calls
for assessing potential costs and benefits
of every possible choice, meaning that
low property prices or enticing ameni-ties could compensate for erecting a ne w
mall somewhere other than Dallas.
In interviews with 49 Dallas business
owners and senior managers in charge of
locating new building spots, Berg heard
nothing about exhaustive property
searches. Most business people considered no more than three possible sites,
and often only one, in the Dallas vicinity
before signing off on a project. Locations
were usually discovered by chance, Berg
reports in an upcoming Journal of Business Research.
In one case, a real-estate developer
driving to a suburban golf course noticed
an undeveloped tract of land that struck
him as promising. He took a detour and
drove around to get a feel for the area.
His entrepreneurial radar went off.
Over the next few days, the developer
determined that a building project on
Looking ahead complex models (such as
multiple regression) are good at “fitting” to past
performance, but better predictions of future
success often come from simpler rules of
thumb. tallying merely counts good characteristics, and take-the-best relies on the first cue
that distinguishes between two options.
Performance of decision-making approaches
Accuracy
Fitting Prediction
75% 71%
73% 69%
77% 68%
source: G. GiGereNzer aNd h. briGhtoN/ TOPICS IN COGNITIVE
SCIENCE 1 2009
Investment
approach
take-the-best
tallying
multiple regression
the vacant land would produce at least a
20 percent annual return on initial cost
within two or three years. That financial
prospect sealed the deal.
Nearly all business owners chose
expansion sites using a version of the
simple real estate formula: “If I can get at
least x percentage of return on my initial
expense within y years, then I’ll do it.”
New ventures that resulted from this
limited search-and-analysis strategy
tended to be profitable, participants
said. A smaller number of failures pro-
vided valuable lessons for evaluating
new expansion sites.
Berg’s sample of entrepreneurs
scoffed when asked whether they
weighed costs and benefits of many
potential sites to calculate the best possible choice. No one has the time or data
to do that, the business people replied. In
a volatile business environment, there’s
no way to know whether such a determination would hold up in six months or a
year, they added.
Like Dallas business big shots, experts
in any endeavor tune out tons of informational noise and focus on a few
powerful but imperfect cues to get the job
done, Gigerenzer proposes. For them,
rationality consists of creative experimentation that leads to the discovery of
useful shortcuts, in his view.
Dropping an econ bomb
Gigerenzer and a colleague, Nobel lau-
reate economist Reinhard Selten of the
University of Bonn in Germany, took that
sacrilegious argument to a high cathedral
of financial orthodoxy last year. The two
were invited to speak, along with a pair
of business entrepreneurs, at a presti-
gious annual gathering of the economics
and business departments at Germany’s
Bielefeld University.
Explore more
s For Nathan berg’s website, including
access to his past publications, visit
www.utdallas.edu/~nberg/
June 4, 2011 | SCIENCE NEWS | 29
