HACK When “5 stars” does no one any service
From Amazon to Airbnb, the star-
rating. The opposite was also
rating system has become a standard
true: Lower scores for individual
metric, tapping the wisdom of the
dimensions pulled down the over-
crowd to give online service providers
all score, even if decor, for example,
instant feedback while consumers get
was less important to a customer
to see who the top-rated providers are.
primarily interested in the food.
But stars can be misleading, depending on the nature of the rating system and
This bias is relevant for
the dimensions being assessed.
platform providers. Although multidimensional ratings pro-
In seven experiments, participants
vide a fuller picture, they may
from Australia, Canada, England,
unduly push ratings in one direc-
Liechtenstein, New Zealand and the
tion or the other. So, instead of
United States were asked to evaluate
asking for dimensional and overall
restaurants, movies and universities.
ratings, platform providers might
Some ratings were single-dimensional,
be better off creating their own
meaning customers were just asked to
aggregates, for less biased results.
rate their experience overall; others
Also, as in the previous example,
were multidimensional, meaning each
providers should consider their
individual element was rated separate-
“choice architecture” carefully,
ly, so food, service, atmosphere, etc.
knowing which dimensions are better to ask (or not ask) their
Where individual dimensions were
customers to rate. Consumers,
rated highly, the overall rating tended to
too, will benefit from being able
be higher than on a platform that only
to evaluate the dimensions most
asked for a single, all-encompassing
helpful to them.
36 | IESE Business School Insight | no. 158
“When the stars shine too bright: the influence of multidimensional ratings on online consumer ratings” by IESE’s Christoph Schneider et al. is published in Management Science (2020).
