Monday afternoon, I was sitting in my parents’ car, in the
middle of the usual New York City/New Jersey bound Thruway
traffic that comes at the end of a long weekend. We were wondering
about whether we’d made a mistake not to stop at the last
rest area, and though I was glad for a chance to chat with
my folks and save my employer a few bucks, I still found a
part of myself thinking fondly of the more predictable, comfortable
train ride I usually take.
Is it funny to call Amtrak predictable or reliable? To many
it would be. It’s no European or Chinese railway, surely.
Anyone who has taken it enough has a story of delays or breakdowns.
But it comes in for abuse way beyond what it’s due.
It’s become an example to me of one more way that humans are
not good at analyzing data on the fly. Along with being really
consistent about being much more scared of dramatic and extreme
but very very unlikely dangers as compared to pretty darn
bad and very common dangers, we also let first impressions
and reputation carry far more weight that they ought to. It
means that bad experiences with certain kinds of things—things
we’re uncertain about or that have mixed reputations, like
public transportation or small local businesses, for example,
loom so much larger than bad experiences with their alternatives.
Airlines, for example, have disastrous delays and other snafus—overbooking
and the like—on a near constant basis. But I’m astounded at
how infrequently people hold that against plane travel. True,
for a very long distance trip there are few alternatives,
but even for a mid distance trip to an urban center where
one could take a train instead, an accurate comparison of
the time and comfort of the two rarely happens. People usually
compare the actual time on the main vehicle, leaving out the
bit where airports are a pain the neck to get to, parking
takes time, you have to get there really early, there’s security
and then on the other end, you’ve arrived at an airport and
need to rent a car or get a taxi to go a pretty long way to
get where you are actually going, as opposed to pulling in
to the center of town. Plus delays and getting bumped and
all that. A train ride that is nominally longer may not really
be. But one late train five years ago and the whole idea gets
Can you imagine if people swore off driving the first time
they spun out on the ice or were late to something important
due to a traffic jam?
I worked on a report not long ago on buy-local policies implemented
by institutions that recognize their self-interest in supporting
their own local economies. The then-director of purchasing
at University of Pennsylvania, who was a leader in crafting
and implementing Penn’s trailblazing work in this area, refused
to pay premiums for local vendors. He made them adhere to
the same bidding rules as everyone else, and just mentored
them through the process. He told me that he frequently ran
into people on campus who didn’t want to buy things from a
local vendor because they once had a bad experience with one.
“How many times have you had a problem with an order from
IBM?” he would ask them. “Or other larger vendors?” Always
they had. “Then let’s not have a double standard.”
I think we’re probably all a little like those people he was
talking with. I’m sure that there are these few areas where
I’m sensitized to the double standard and resist it and many
more where I am blithely not. Something familiar, something
that has a reputation for being efficient, or a good value,
or worth it in some way, has many data points in our head.
In those cases, one bad experience is just one more data point.
It’s not a first impression. It’s not definitive. But for
anything new or different, something we distrust, or others
around us distrust, or is merely an unknown, one bad experience
There’s probably more of a role for marketing and branding
in here than I might usually think of. But in the meantime
the only clear response to this dynamic I can make is to check
myself for it at regular intervals and try to point it out
in a non obnoxious way when I run into it elsewhere. Point
it out to me too, OK?