00:00:01
we're here with chunchun who is an
00:00:04
postdoc researcher with the Mac
00:00:06
Institute for innovation management he's
00:00:09
here to talk about his research looking
00:00:12
into bilateral rating systems in online
00:00:15
marketplaces welcome thank you for
00:00:19
inviting me it's a great honor to be
00:00:21
here to share my research with you can
00:00:23
you tell us first about your work what
00:00:26
are you trying to study
00:00:27
so our paper is essentially trying to
00:00:30
understand the impact of the rating
00:00:33
systems on those online marketplaces so
00:00:36
notice that a recent years the digital
00:00:38
innovations in especially the rapid
00:00:41
development of the mobile Internet of
00:00:43
servers as well as the smart forms they
00:00:46
have really changed people's everyday
00:00:47
life so one of the greatest example is
00:00:50
the sharing economies so the companies
00:00:52
like uber and airbnb is so these
00:00:55
companies works like a platforms that
00:00:58
connects millions of online individual
00:01:00
customers to millions of all fly
00:01:02
individual service providers three very
00:01:05
simple applications on their smartphones
00:01:07
so notice that on the demand side of
00:01:10
these platforms the count of the
00:01:13
customers now is able to make orders at
00:01:15
any time in any place and on the supply
00:01:18
side the individual service providers
00:01:21
now they can't decide when to work and
00:01:23
how long to work on the platform it is
00:01:26
because of these new technologies that
00:01:28
removes their physical obstacles between
00:01:31
two parties but on the other hand the
00:01:34
psychological obstacles namely that the
00:01:37
trust between the two parties is still
00:01:39
there
00:01:39
notice that the concept of trust is not
00:01:42
a new is not a new concept but it and
00:01:45
never play a important role as in the
00:01:48
current economy context and the reason
00:01:50
is because so the transactions that's
00:01:52
going to happen on these platforms is
00:01:54
between two or two parties of completely
00:01:57
strangers and notice that because of
00:02:00
these high liquidity of both parties
00:02:03
people the customers and the service
00:02:05
providers they are less likely to
00:02:07
encounter with each other for more than
00:02:10
once so what I mean is for example you
00:02:13
are less likely to stay in the
00:02:15
same accommodation listed on the Airbnb
00:02:17
website and the reason is because
00:02:20
probably you you want to change a
00:02:22
different place to visit in your next of
00:02:24
application right and even if your take
00:02:27
the ubers to go to office every day
00:02:29
still it is less likely that you meet
00:02:31
the same driver so exactly because of
00:02:34
this high liquidity of both side and
00:02:36
also almost negligible entering cost of
00:02:39
the supply side this create a big issues
00:02:42
in establishing the trust and notice
00:02:45
that the customers it is very difficult
00:02:47
for the customer to judge how serious
00:02:49
the service provides provider fears
00:02:51
about his own business activity and the
00:02:53
willingness to continue provide service
00:02:56
in the long run okay
00:02:57
and the rating system has been proved
00:03:00
that it is an effective tool to
00:03:02
establish the trust between two parties
00:03:04
it is natural that people want to have
00:03:07
reassurance from other users so that
00:03:10
they can make sure that they didn't
00:03:11
waste their money with the money on the
00:03:14
bad products or better service it is the
00:03:16
rating system that makes the waste of
00:03:19
crowds available that brings the people
00:03:21
who's completely stranger people's
00:03:23
together to conduct business activities
00:03:25
and our papers what we really are trying
00:03:28
to understand is how this impact of
00:03:31
these rating systems on the three
00:03:33
parties of the game which is the
00:03:35
platforms the service providers and also
00:03:37
the the customers so what are your
00:03:40
papers key takeaways and were there any
00:03:42
conclusions that surprised you so we're
00:03:45
still in the process of analyzing our
00:03:47
model but we have I think we have
00:03:50
already some interesting results to
00:03:51
share with you so we first in the first
00:03:53
stage we consider a platform without a
00:03:56
rating system versus a unilateral rating
00:03:59
system in a sense that only the
00:04:01
customers is able to submit their
00:04:03
ratings to the service providers and in
00:04:05
a next stage we're going to analyze the
00:04:07
bilateral rating system where both
00:04:09
parties they can read each other after
00:04:10
the transactions so currently what we
00:04:13
found is that the rating systems will
00:04:15
have different impact on the free party
00:04:17
of the game the the service providers
00:04:19
customers and the platform so from the
00:04:22
less will talk about this one by one so
00:04:25
first are from the perspective of the
00:04:27
plat platforms
00:04:28
we found that it is always their best
00:04:30
interest to implement this rating
00:04:32
systems and the underlying intuition
00:04:35
behind is because by allowing the
00:04:38
customers to read the service providers
00:04:40
it makes the service providers have more
00:04:43
motivations to exert more effort because
00:04:46
now their effort is more transparent and
00:04:49
observable through the ratings so these
00:04:52
kind of make the customers feel a little
00:04:53
bit happy because they get some extra
00:04:55
utilities and this kind of in turn makes
00:04:59
the other platforms more flexible in
00:05:02
redesign is pricing strategy so for
00:05:05
example right now the platform's is able
00:05:07
to increase the price a little bit and
00:05:09
also increase the fee charged by the a
00:05:11
charge of from via the service providers
00:05:14
a little bit so that's kind of both a
00:05:16
boost of the revenue of their platform
00:05:18
ok so that's for the platforms so then
00:05:22
when we turn to the service providers
00:05:24
one would attempt to think that these
00:05:27
should get worse after the
00:05:28
implementations of the rating systems
00:05:30
and this is also what we think initially
00:05:32
because simply because first they have
00:05:35
to work harder because their effort is
00:05:37
now observable and a second as the
00:05:40
reason we just mentioned the platforms
00:05:42
now going to increase the fee a little
00:05:44
bit so it sounds like that the service
00:05:47
provider they are in create double loss
00:05:48
and it is true that for the year revenue
00:05:51
per order they do get a decrease but
00:05:54
this is not the whole story this is only
00:05:56
part of the story and because we know
00:05:58
that on average now the service
00:06:00
providers there are making more effort
00:06:02
so he's gonna attract more customers to
00:06:05
use this platform to book service so
00:06:08
that's indeed what we found so under
00:06:10
some circumstances the transaction
00:06:12
volume gonna increase so great such that
00:06:14
is going to exceed the the loss of the
00:06:17
revenue promoter so in that case the
00:06:19
service provider is actually getting
00:06:21
better due to the fact that they are
00:06:23
being rated and the last part is for
00:06:26
their customers so it is natural to
00:06:28
think that after the implementation of
00:06:30
the rating systems the customers are it
00:06:33
should attract more customers to join
00:06:35
this platform and however we found that
00:06:37
it is actually not always the case and
00:06:39
the reason is really deep
00:06:41
on the pricing strategy of the platform
00:06:44
and notice that the platform care about
00:06:47
I care about two things one is the
00:06:49
revenue per order the other is a
00:06:50
transaction volume and you notice that
00:06:53
usually there's a tension between these
00:06:55
two factors so if you want to have a
00:06:58
higher profit margin per order usually
00:07:00
you cannot at the same time achieve the
00:07:02
goal of having a high transaction volume
00:07:04
right so okay so then what we found is
00:07:07
that if the customers valuation about
00:07:10
the service itself is at a pretty low
00:07:12
level and then in that case the
00:07:16
transaction volume plays a more
00:07:17
important role versus Avandia the
00:07:20
revenue per order to the platforms so in
00:07:23
the in that case what the platform's
00:07:25
gonna do is they're gonna they're gonna
00:07:27
adopt a less aggressive pricing strategy
00:07:30
such that it's gonna charge the fee from
00:07:33
the server's a relatively lower and then
00:07:36
it's gonna induce all the servers on the
00:07:38
platform to exert some F some level of
00:07:41
extra effort so that's got to make the
00:07:43
customer feel happy so more customers
00:07:45
gonna involve in these platforms and
00:07:47
increase the transaction volume while on
00:07:49
the other hand if the evaluation of the
00:07:53
customers to the service itself is
00:07:55
already pretty high so which means that
00:07:57
they are already very satisfied with the
00:07:59
service even if without any extra amount
00:08:02
of effort by the service provider in
00:08:05
this case we found that the revenue per
00:08:07
order plays a more dominant role than
00:08:09
the transaction volume so in this case
00:08:11
the platform will adopt a very
00:08:13
aggressive stress pricing strategy in a
00:08:16
sense that is going to increase the fee
00:08:17
charged from the service provider so
00:08:20
that's gonna make only part of the
00:08:22
service provider
00:08:23
exert some extra amount of effort but
00:08:26
the remaining service providers they
00:08:28
were just produced the usual the the
00:08:31
basic service without the actor effort
00:08:33
and these kind of this amount of the
00:08:35
service providers they're they just
00:08:37
barely make a living on the platform so
00:08:39
in this case the the customers the
00:08:42
amount of customers that involving these
00:08:44
platforms is actually lower than the
00:08:46
system are they on the platform without
00:08:48
a rating system that's that's our main
00:08:50
findings what are some practical
00:08:52
applications of your research okay
00:08:55
based on our analysis first we know that
00:08:57
okay it is it is always the best
00:08:59
interest for the platform to implement
00:09:02
such a rating systems and then a second
00:09:05
in terms of his pricing strategies first
00:09:08
we think that the platform should
00:09:09
conduct some marketing research to
00:09:11
gathers the informations about how the
00:09:14
customers feel about the service itself
00:09:16
so for example they need to know that
00:09:18
what is the woodenness to pay off the
00:09:19
customers to acquire this service and
00:09:21
after they know this information they
00:09:24
can identify which region of the the
00:09:26
customers valuation falls in and then
00:09:28
according to that information they know
00:09:30
either they should adopt a aggressive a
00:09:33
pricing strategy or a mild one okay so
00:09:36
what sets your research apart from work
00:09:39
in this area from prior work okay so for
00:09:43
the previous or works and they can be
00:09:45
described into two stream of
00:09:47
literature's so the first string is they
00:09:51
start at a perspective of the platforms
00:09:54
they try to understand that the optimal
00:09:56
pricing strategy of the platform but
00:09:58
under the assumption that they they they
00:10:01
think that the rating distribution or
00:10:03
the effort distributions across the
00:10:05
service they are pre pre given so they
00:10:08
are not changed so this is the first
00:10:09
string and a second string of the
00:10:11
literature is they start from the
00:10:13
precepts perspective of the service
00:10:15
provider they try to understand the
00:10:17
optimal decision of the service provider
00:10:19
and under the assumption that the prices
00:10:22
structure in the system is given and as
00:10:25
you may have noticed that actually the
00:10:27
the price decision of the platform and
00:10:30
the decision of the service provider
00:10:31
they are interrelated so they're going
00:10:34
to affect each other and eventually
00:10:35
they're going to reach an equilibrium so
00:10:38
that that is our research we endure
00:10:40
endure tonight both the a decision of
00:10:43
the platform and also the decision of
00:10:45
the service provider and also in the
00:10:47
next stage were also endogenous the
00:10:49
decision of the customers together so
00:10:51
and solve for the equilibrium that's
00:10:52
gonna give us a more complete picture
00:10:54
about the system so how will you follow
00:10:58
up your research okay yeah as we said
00:11:00
that we're we're still are this is still
00:11:02
a ongoing research so in the next stage
00:11:04
we're gonna keep focused on
00:11:07
analyzing the bilateral rating systems
00:11:09
we're also the servers they can read
00:11:11
back the customers and notice that lease
00:11:14
is a very important feature in the
00:11:16
platforms such that the customers there
00:11:18
they are not just a purchase a physical
00:11:21
product they actually purchase a service
00:11:23
so in this case both the customers and
00:11:25
the service providers they're going to
00:11:27
interact with each other for some period
00:11:29
of time it could be as short as a ride
00:11:32
like uber or as long as couple of months
00:11:35
like Airbnb so in this case that the
00:11:38
service provider is also under the
00:11:40
potential risks from the customer side
00:11:41
because the customers misbehavior may
00:11:44
may damage the the asset of the service
00:11:47
provider like the service cars or house
00:11:50
right so um and that's going to affect
00:11:53
the server's future ability to provide a
00:11:55
service so for the platforms like this
00:11:58
where the customers they're buying
00:11:59
service rather than a physical product a
00:12:01
bilateral Reading System is needed and
00:12:03
this is what we're going to focus in the
00:12:06
next stage well thank you so much for
00:12:08
joining us today okay thank you very
00:12:09
much for inviting me
00:12:26
you
00:12:30
[Music]
