00:00:01
i'm here with Wharton management
00:00:03
professor Tyler rye and doctoral student
00:00:06
Adam caster who are here to talk about
00:00:08
their paper cognitive neighborhoods in
00:00:10
the valuation of innovation across
00:00:12
national analysis welcome thank you
00:00:16
thanks so tell us about your paper yeah
00:00:19
so in a nutshell this paper is trying to
00:00:22
understand exactly how classification
00:00:24
systems which differ across countries
00:00:27
how that affects outcomes animation in
00:00:29
particular trying to understand how the
00:00:31
classification system affects how patent
00:00:33
examiner's search for prior art and how
00:00:36
they find sort of what might be relevant
00:00:38
or not relevant so what are some of the
00:00:40
key takeaways of your paper so the key
00:00:42
takeaways are that it's really important
00:00:44
to understand how the classification
00:00:47
system affects the way that an
00:00:49
innovators innovation is going to be is
00:00:51
going to be viewed and seen so if it's a
00:00:54
sign to a certain class it's a class
00:00:56
that is much more central and much more
00:01:00
related to other technology classes it's
00:01:01
going to be much more likely to be
00:01:02
picked up by others when they're doing
00:01:05
future examination search by not only
00:01:07
patent examiner's but also inventors so
00:01:09
it's going to much more likely we cited
00:01:10
and probably much more likely to be more
00:01:12
highly valued so that's an important key
00:01:14
takeaway yeah absolutely you know
00:01:17
something that's interesting to keep in
00:01:18
mind when we're looking at patents and
00:01:20
classification and the valuation of
00:01:22
intellectual property is there's an
00:01:24
assumption that the number of citations
00:01:26
that a patent gets in future innovation
00:01:30
is a reflection of its value and what
00:01:33
Adams done really nicely in this paper
00:01:35
is show that it's not just a reflection
00:01:38
of the quality of the patent that's
00:01:40
driving its valuation but it really is
00:01:43
where it gets slotted into the
00:01:44
classification system so examiner's like
00:01:47
all of us use heuristics and shortcuts
00:01:49
to think about what's related in what
00:01:50
isn't and what we've done in the paper
00:01:53
is actually modeled this out and shown
00:01:56
the structural characteristics of the
00:01:58
different patent classes in terms of how
00:01:59
they relate to each other based on how
00:02:02
examiners are searching and the big
00:02:04
takeaway is that the same piece of
00:02:06
intellectual property if it's in a class
00:02:09
a patent class that is in the
00:02:14
neighborhood of many
00:02:16
examiners are more likely to then cite
00:02:18
it in future patent searches and this is
00:02:21
going to affect its valuation so
00:02:23
irrespective of quality its examiner's
00:02:25
it's their cognition and it's the
00:02:27
properties of the categorization system
00:02:28
that really seemed to be driving
00:02:30
valuation in an important way as well as
00:02:33
perceptions of what is a breakthrough
00:02:34
technology versus what isn't and in to
00:02:38
add on to that another important aspect
00:02:39
that we find is that we're looking at
00:02:41
cross different patent systems so we're
00:02:43
looking at the US Japan and Germany and
00:02:45
you can see that there is structural
00:02:47
differences in the classification
00:02:48
systems across these different patenting
00:02:49
jurisdictions and that this itself you
00:02:52
know just even looking at the exact same
00:02:54
innovation being applied for in these
00:02:56
different systems kind of very different
00:02:57
outcomes purely based off of or at least
00:03:00
in a strong part due to the fact that
00:03:02
the way that the technology is
00:03:03
classified is different and so this
00:03:05
isn't really important for for
00:03:07
practitioners to sort of be aware of and
00:03:09
also just in terms of better
00:03:11
understanding exactly not only
00:03:13
evaluation itself but is how I was
00:03:15
mentioning what might be breakthrough
00:03:17
what might not be breakthrough what's
00:03:18
the best way to ensure that something
00:03:20
might be a breakthrough is going to be
00:03:21
seen by the most people going forward so
00:03:24
what conclusions have any surprised you
00:03:26
so I think for me personally I expected
00:03:31
that for technology classes that were
00:03:33
very central so we're very related other
00:03:36
technology classes I thought there'd be
00:03:38
a downside in the sense that if it's
00:03:40
highly related it might distract it
00:03:42
might detract attention away from that
00:03:43
focal class so if you think about a
00:03:46
patent examiner inventor who has a
00:03:48
technology and they're looking back if
00:03:51
it's if the technology class itself that
00:03:53
they're originally searching it is very
00:03:55
broad they might leave as much time to
00:03:57
search directly in that in that primary
00:03:58
class so I was really expecting there to
00:04:00
be sort of a negative effect as well at
00:04:02
least in terms of saying that while it's
00:04:06
going to it's going to say it's going to
00:04:07
basically push attention further
00:04:12
outwards and I thought there might be a
00:04:13
negative effect what was rising to us is
00:04:15
that we didn't find that at all in fact
00:04:17
what we do find that most of the you
00:04:19
know the difference in terms of what
00:04:21
technologies get cited more often than
00:04:23
not comes from citations from outside
00:04:25
the class there was actually no negative
00:04:27
effect within the class itself which I
00:04:28
think for me was was
00:04:29
surprising I thought there'd be some
00:04:30
some negative associated with with sort
00:04:32
of being more central yeah and for me my
00:04:35
my big takeaway is uh you know a little
00:04:38
less complicated in your uh which I
00:04:41
agree absolutely everything you've just
00:04:42
said but it surprised me that the same
00:04:46
piece of intellectual property applied
00:04:47
for in different systems is valued so
00:04:50
differently you know you'd expect that
00:04:52
there is some sort of underlying quality
00:04:54
that's reflected in the valuation of
00:04:56
intellectual property and the degree to
00:04:58
which it's picked up and built on and
00:04:59
what we found is that I mean this is
00:05:01
true to a certain extent but depending
00:05:04
on the properties of the classification
00:05:06
system this changes in really quite
00:05:08
dramatic ways and it you know it really
00:05:10
speaks to the cultural creation of
00:05:13
innovation and innovation outcomes what
00:05:17
are some of the practical implications
00:05:18
of your findings yeah so I think that
00:05:22
the first important thing to note
00:05:23
especially is is sort of practitioners
00:05:26
to understand the differences across
00:05:27
these systems now I think admittedly
00:05:30
some the research might be one step away
00:05:32
from you know really being able to
00:05:34
explain exactly how should patent
00:05:37
applicants really proceed in terms of
00:05:38
trying to maximize value so first I
00:05:42
think a first step in an area for future
00:05:43
researchers to understand exactly how
00:05:45
much agency they have in terms of what
00:05:47
classes are assigned certainly there is
00:05:49
at least indicators that the USPTO for
00:05:52
instance has a very strong hand and sort
00:05:54
of affecting what the eventual primary
00:05:55
classification is going to be which is
00:05:57
what we're looking at but also when
00:06:00
applicants file for a patent application
00:06:02
there they're required to include a
00:06:05
class when in their first application so
00:06:08
i think it's it's clear that the you
00:06:10
know the class is important what the
00:06:12
next step is to understand exactly how
00:06:13
much agency and what can Apple can't do
00:06:17
in order to ensure that their
00:06:18
application their applications are going
00:06:20
to be funneled into the classes that
00:06:21
they want it and be seen by the the
00:06:24
proper examiner's so what sets your
00:06:26
research apart I think you kind of
00:06:28
touched on that could you explain a
00:06:30
little bit more yeah absolutely so
00:06:31
theoretically there is a large
00:06:32
difference in terms of the way that
00:06:34
classification understanding its use as
00:06:36
a heuristic has been used and so this
00:06:38
context is very different because we're
00:06:40
trying to understand search and
00:06:42
particularly sir
00:06:43
for innovation so generally generally
00:06:46
the research is focused on consumers
00:06:49
looking for specific products which is a
00:06:51
slightly different context one where
00:06:52
they already have in their mind a
00:06:54
certain type of product they're going to
00:06:55
look for and so the dynamics are very
00:06:57
different when you're thinking about
00:06:58
technology on the other hand especially
00:07:01
in this paper we're looking at
00:07:03
nanotechnology which for a large part of
00:07:06
the period that we're studying is a very
00:07:07
nascent area is to understand exactly
00:07:09
how do people make sense of these new
00:07:11
technologies and in particular when
00:07:13
they're when they're searching trying to
00:07:14
understand that it's related where where
00:07:16
might this new you know piece of this
00:07:19
new innovation or adventure and fit into
00:07:21
the sort of broader sense and where my
00:07:23
in have might one frame it in order to
00:07:25
best take advantage of of the you know
00:07:27
the institution in terms of its
00:07:29
classification or aspects or become
00:07:31
really important I think this is
00:07:32
something that's that sort of missed and
00:07:34
more broadly we have a much more fluid
00:07:37
view of the way that these
00:07:38
classification systems work so
00:07:41
theoretically most of the research is
00:07:43
sort of focused on how these
00:07:45
classifications are sort of fixed in
00:07:47
time and all individuals really have
00:07:51
this sort of the same premonition or the
00:07:53
same sort of cognitive schema of what
00:07:55
these classification systems look like
00:07:56
whereas whereas we're taking much more
00:07:58
fluid look and trying to look at how a
00:08:00
new technology class know technology is
00:08:02
starting to become you know sort of more
00:08:04
mainstream and sort of how that affects
00:08:06
the innovations over time yeah and I the
00:08:08
only thing I'd add to that is I think
00:08:10
that one of the unique things about this
00:08:11
work that we're doing is instead of
00:08:13
looking at individual categories or
00:08:15
individual classifications which is the
00:08:17
traditional approach in the academic
00:08:19
literature we're really highlighting the
00:08:21
influence of the properties of the
00:08:23
classification system and so we're
00:08:24
taking a level of analysis from here and
00:08:26
expanding it out to the broader
00:08:28
topography of categories and how they
00:08:30
relate to each other now this leads to
00:08:32
some different questions and this is
00:08:34
part of you know the basis for you know
00:08:37
want to undertake this research in the
00:08:38
first place because the traditional
00:08:40
prediction is that if you want to be
00:08:42
understood and you want to be valued you
00:08:45
have to fit neatly within a particular
00:08:46
category and so this applies very
00:08:49
generally to technologies to people to
00:08:52
organizations and there's a large
00:08:54
literature that's looked at this
00:08:56
but what hasn't really been looked at is
00:08:58
how the embeddedness of the category
00:09:00
that you're in in a broader system of
00:09:02
classification effects outcomes for you
00:09:05
as a technology a person and
00:09:06
organization whatever it is and that's
00:09:08
really the piece that we're bringing to
00:09:10
bear in this research showing that yes
00:09:12
it's important what category you're in
00:09:14
but really what's driving the effects
00:09:16
especially when valuation is based on
00:09:19
being on the radar of lots of other
00:09:20
things is where this class sits in the
00:09:23
broader system what lessons are in your
00:09:28
research applicable to venture
00:09:30
capitalists inventors and businesses are
00:09:33
there any yeah I think so just as Tyler
00:09:36
was mentioning oftentimes we think of it
00:09:39
being necessary to sort of fit into a
00:09:41
single category have a very very salient
00:09:43
identity that fits into a single
00:09:45
category and that that is important to
00:09:46
degree but especially for I think in
00:09:48
particular new innovations new
00:09:50
technologies is in order to get garner
00:09:52
widespread attention you actually want
00:09:56
to sort of sometimes appear more broad
00:09:57
so in particular it might not be a bad
00:09:59
thing I think we're showing that's not a
00:10:01
bad thing to actually fall into multiple
00:10:03
different classes because that's going
00:10:04
to garner more intention so I think
00:10:05
fundamentally speaking it's important of
00:10:08
a sealing identity at the same time I
00:10:09
think our research is really going the
00:10:11
direction of showing that actually
00:10:12
there's a lot of benefits to sort of
00:10:14
spanning multiple different classes
00:10:16
product headers etc because that might
00:10:18
garner more attention which ultimately
00:10:19
my benefit producers themselves and
00:10:23
venture capitalist in terms of thinking
00:10:24
about who they should invest in and the
00:10:26
other thing that I'd add to that is to
00:10:28
the extent that venture capitalists or
00:10:29
other audiences are looking at patent
00:10:32
citations as a way to value startups IP
00:10:35
or even an established company's
00:10:36
intellectual property you know they
00:10:39
might be picking up some misleading
00:10:40
stuff here so doing close to diligence
00:10:43
is probably still very important
00:10:45
especially when you're thinking about
00:10:47
something like nanotechnology that
00:10:48
really is global in its scope and
00:10:50
revolutionary and its potential unless
00:10:52
you're actually getting under the hood
00:10:54
and understanding the nuts and bolts of
00:10:56
what this technology is doing what its
00:10:59
potential is in the real world you might
00:11:01
be missing stuff if you're just looking
00:11:03
at things like you know the traditional
00:11:05
measures of forward citations it's
00:11:07
picked up by a lot of other
00:11:08
innovations this must be important or
00:11:10
it's picked up by lots of diverse
00:11:12
innovations so this must be a
00:11:13
breakthrough so what we're showing is
00:11:15
you know there's a piece to that that's
00:11:17
true but if you're relying just on that
00:11:19
you're probably missing a big chunk of
00:11:21
what's actually driving quality in this
00:11:23
field
00:11:40
you
