00:00:00
Stefano Puntoni: But in many situations,
we do have lots and lots of
00:00:02
data. The problem is sometimes
those data are not the data you
00:00:06
need, or maybe those data are
not being thought of the right
00:00:09
way in terms of supporting the
decisions that are important to make.
00:00:12
Dan Loney: But I guess
00:00:13
there is an importance that we
need to kind of reinstall, in
00:00:17
many cases, on the human
component of this entire
00:00:20
process.
00:00:21
- Yeah, the human judgment is
crucial. So actually, what I
00:00:25
argue always is that as computers
and algorithms become smarter
00:00:29
and smarter, and the data
becomes better and better, we
00:00:32
should think harder, not less.
And so in a way, the key message
00:00:36
of the book almost is to say --
Dan Loney: And welcome to
00:00:39
a special edition of Ripple
Effect: Meet the Authors. I'm
00:00:42
your host, Dan Loney. In each
episode this month, the podcast
00:00:46
will feature Wharton faculty
authors in lively, fast moving
00:00:49
conversations about their latest
books and research. We're going
00:00:53
to be covering a diverse range
of topics, bringing you the
00:00:56
latest insights and knowledge
that you can apply to your life
00:00:59
and to work. And hello, and
welcome to the Ripple Effect:
00:01:03
Meet the Author series. I'm your
host the Dan Loney. Today we're
00:01:06
talking with Stefano Puntoni
about his book, <i>Decision-Driven</i>
00:01:10
<i>Analytics</i>. Stefano's book is
about leveraging human
00:01:14
intelligence to unlock the power
of data. And the book offers a
00:01:17
new approach about making good
decisions with data. Stefano,
00:01:22
great to see you again. Thanks
very much for your time today.
00:01:24
And I guess let's start out with
just explaining how you view, or
00:01:28
what your definition of decision-
driven analytics is. - Thanks,
00:01:32
Dan, for having me. And I'm
excited to be here and tell you
00:01:35
a little bit about our new book.
And the book is motivated by the
00:01:41
problems that we see in many
companies where analytics are
00:01:45
complex. Companies have invested
heavily. We have lots of data.
00:01:49
And yet companies sometimes seem
to struggle to extract value
00:01:52
from that data. And what we
argue in the book, basically, is
00:01:56
that the gold standard that many
companies have set themselves on,
00:02:01
which is data-driven
decision making, may to some
00:02:05
extent, actually get the point
wrong, meaning perhaps we argue
00:02:11
it shouldn't be the decision
making that is data driven.
00:02:13
Instead, we argue it should be
the data analytics that is
00:02:17
decision driven, meaning rather
than focusing on the data, and
00:02:21
then trying to figure out how do
we extract value to improve
00:02:24
our decision making, it might be
much better, in fact in many
00:02:27
situations, to do the opposite.
And say, let's start from the
00:02:30
decisions that we need to make
and figure out what data we need
00:02:33
to make those decisions.
00:02:35
- What kind of potential impact
then have we maybe
00:02:38
you've been looking at because --
that maybe the process hasn't
00:02:43
been exactly the way that it
should be and we're maybe not
00:02:46
extracting the value that we
really should? - The
00:02:50
idea of data-driven decision
making is not -- nobody argues that
00:02:53
using data to make decisions is a
bad idea. We need data to make
00:02:56
good decisions. What we argue in
the book is that the emphasis is
00:03:00
wrong. And that, you know, the
complexity and investments
00:03:04
required to get the data systems
in place, and all that work that
00:03:08
has to be done, complex,
technical, have attracted a lot
00:03:13
of attention. And at some point,
people end up in a situation
00:03:15
where looking at data and trying
to find a purpose for it, rather
00:03:19
than looking at the questions
that they need to answer and
00:03:24
find data for that purpose.
00:03:27
- Are there common mistakes that
end up popping up as you've
00:03:31
kind of investigated this by not
going more of a data-driven
00:03:36
perspective?
00:03:37
- Yes, so the mistakes that we see
companies do are basically of
00:03:42
a variety of types. But the two
that are very common are that
00:03:46
companies sometimes end up
answering the wrong questions,
00:03:48
because they haven't connected
the analytics to the decisions
00:03:51
that they need to make very
well. And because of that, they
00:03:54
explore stuff that is actually
not the stuff they should be
00:03:56
exploring. And the second is
that by focusing so much on the
00:04:00
data, you almost end up putting
this data on a pedestal and
00:04:03
glorifying them to the extent
where you might actually become
00:04:05
uncritical of them, and not
understanding what are the
00:04:08
pitfalls that you might run into
if you make decisions based on
00:04:11
this data. And so the basic
issue is actually that of a
00:04:17
disconnect between decision
making and data analytics,
00:04:20
where there is this large
gap between the people, the
00:04:24
leaders who make the decisions,
and the systems that are meant
00:04:27
to support them.
00:04:29
- I guess it's important to ask
these questions right now,
00:04:31
because it feels like we have
more data than ever. And in many
00:04:35
cases, I would assume the
decision makers are probably
00:04:39
inundated with this data, and
maybe don't know how to parse
00:04:43
through it and make the right
call.
00:04:44
- Yeah, it's like a strange irony in
that for a long time companies
00:04:47
were complaining they don't have
data to make decisions. Now
00:04:51
they're still complaining about
data, but now they complain,
00:04:53
oftentimes, you know, they have
too much data, meaning that
00:04:56
there's so many things that one
could look at and it's a bit --
00:04:59
it's easy to get lost.
Of course, we still have many
00:05:03
situations where we don't have
good data. And companies still
00:05:06
are struggling to figure out
ways in which they can achieve,
00:05:10
you know, kind of evidence-based
decision making in those
00:05:13
situations. And I think we need
to keep looking for great data
00:05:16
there. But in many situations,
we do have lots and lots of
00:05:20
data. The problem is that
sometimes those data are not the
00:05:23
data you need, or maybe those
data are not being thought of
00:05:26
the right way in terms of
supporting the decisions that
00:05:29
are important to make. - But
00:05:30
I guess there is an importance
that we need to kind of
00:05:34
reinstall, in many cases, on the
human component of this entire
00:05:38
process.
00:05:38
- Yeah, the human judgment is
crucial. So actually, what I
00:05:42
argue always is that as
computers and algorithms become
00:05:46
smarter and smarter, and the
data becomes better and better,
00:05:49
we should think harder, not
less. And so in a way, the key
00:05:53
message of the book almost is to
say that the secret to making
00:05:58
good decisions with data is that
before you even start looking at
00:06:02
the data, you need to do a lot
of thinking. So it takes a lot
00:06:05
of thinking without data to make
good decisions with data. - So
00:06:09
we're now in a time where
artificial intelligence is
00:06:12
playing a bigger and bigger role
in a lot of things that we do.
00:06:17
How potentially does AI kind of
impact the decision process and
00:06:22
either make it better or worse,
depending on all of this data
00:06:26
that we have now in the mix?
- Yeah,
00:06:27
AI is the biggest thing
happening today. And I'm very
00:06:31
excited about all the work that
the Wharton School is doing in
00:06:34
the area of AI. In fact, our
Dean, Erika James, just yesterday
00:06:37
announced a major new initiative
called the Wharton Initiative for
00:06:41
AI and Analytics. And we're
basically really pushing hard on
00:06:47
what we know and what we can do
to support business decision
00:06:50
making. But AI here, as I think
both, you know, a risk and an
00:06:57
enormous upside potential, in
that the risk is as we use more
00:07:02
and more complex algorithms and
technology, that gap that I
00:07:05
mentioned earlier between
decision making and the analyst
00:07:09
is only risking to grow even
larger, because now basically
00:07:11
the decision makers have even a
looser grasp of what these
00:07:16
techniques can do and what they
are about. And the people who
00:07:18
are technical become so
technical, they now lack, you
00:07:21
know, a link to the business and
maybe domain expertise that can
00:07:24
help them be good partners for
the leaders who are making
00:07:28
decisions. So that's the risk
part. But the upside potential
00:07:32
is enormous. I think we are
going to benefit in a million
00:07:36
different ways from AI to
improve our decision making, by
00:07:40
automating decisions in a much
smarter way, by providing tools
00:07:44
that can help support human
expert's decisions, and making
00:07:49
sure that the scarce expertise of
our human professionals is used
00:07:55
for the cases where that
expertise can really move the
00:07:58
needle, and leave the rest to AI.
- And
00:08:01
the goal is still the same from
the company's perspective.
00:08:04
They're looking for the best
outcome they can, you know,
00:08:07
create, but you're going about
it in a much more advanced way
00:08:11
with all of these different
components kind of in the mix
00:08:14
now.
00:08:14
- Yeah, for sure. AI is not going
to tell you what you want, and
00:08:17
what you should want. AI is
going to just help you get
00:08:19
there. So we still need the
leadership, we still need human
00:08:23
expertise and creativity,
imagination to think about what
00:08:26
the future might look like.
- Does that
00:08:27
then kind of take us into the
discussion a little bit about
00:08:31
what AI is going to mean for
companies moving forward?
00:08:35
Obviously, there's a lot of
discussion about AI replacing
00:08:38
humans. But it seems like more
and more, a lot of people are
00:08:41
coming around to the fact that
AI will be complementary to
00:08:44
humans.
00:08:44
- It will be both them. But I
think that a lot of the
00:08:47
conversations around AI have
this tone of human replacement,
00:08:51
human or AI. Can we develop
AI systems that emulate human
00:08:55
capabilities? And the assumption
of doing that, essentially, is
00:08:58
to say, well, once we have the
AI system that can do what the
00:09:01
human does, maybe we don't need
a human anymore. And I argue
00:09:04
that business schools can play
an important role to complement
00:09:08
that perspective with a
different one, which is what
00:09:12
kinds of AI systems do we need
to complement and augment human
00:09:16
experience and human expertise?
So instead of thinking about
00:09:19
human or AI, can we think about
human and AI? For example, we
00:09:23
just started a course here.
I'm teaching a course at Wharton
00:09:26
called "AI In Our Lives." And the
whole point is exactly to do
00:09:30
that and think about, what can AI
do, and how do we benefit from
00:09:34
that? And then what is it that
humans do incredibly well, and
00:09:38
how do we design an AI system to
support that?
00:09:40
- I would imagine that when you
look at the component of
00:09:46
the decision process that will
go on, you're talking about
00:09:49
things that can impact top to
bottom, up and down the corporate
00:09:53
ladder in terms of all kinds of
different decisions that may
00:09:56
come into play.
00:09:57
- Yeah, of course. In different
functions, at different levels
00:10:00
in the organization. And, you
know, one point that we make in
00:10:04
the book is that every decision
making -- maker needs to consider
00:10:08
what decisions are mine to make,
and not worry about those are
00:10:12
not yours to make. And of
course, you can lobby, you can
00:10:14
try to influence and bring, you
know, issues to the table for
00:10:19
those who make those decisions.
But you should primarily focus
00:10:22
on the decision you need to make
and then find out how technology
00:10:26
and data can help you making
those decisions. - One of
00:10:29
the other areas you talk about
in the book are the
00:10:32
decisions that end up having an
impact either in the area of
00:10:36
costs or the area of benefits.
And I guess there's a dynamic of
00:10:41
how that decision process will
end up impacting the thought of
00:10:46
the investment that a company
may make, but obviously, the
00:10:48
longer term benefit along the
way as well.
00:10:50
- Right, and that stems from the
vision of the leadership, of
00:10:53
course, what it is that we're
trying to pursue. I think AI is
00:10:57
one of those technologies, and
analytics more broadly, that
00:11:00
enables companies to think in
different ways about the benefit
00:11:04
that it can reap. Some of it might
be efficiency. So it could be
00:11:07
mostly a matter of making things
smoother, better, reducing risks,
00:11:11
and reducing costs, and managing
things that way. But there is
00:11:15
also a different way of thinking
of it, and say how do we make
00:11:18
things better and more
effective? And perhaps not just
00:11:21
say, how do we do what we do now
better, but to think bigger and
00:11:26
think about what is it that we
are not currently doing that we
00:11:28
could be doing. And that I
think is the real potential of
00:11:32
of data and AI.
00:11:33
- For people who read the book,
what do you hope is the message
00:11:36
that's delivered to them?
00:11:38
- We end with this message, and
I think it is really the key, key
00:11:42
thing, I believe, which is
ultimately data and analytics, they
00:11:46
attract a lot of attention.
They are sexy, shiny objects,
00:11:49
and they are, you know,
interesting problems. They can also
00:11:53
be scary and complicated, and
all of that. So it's not
00:11:56
surprising that so much
attention is dedicated to them.
00:11:59
But we never need -- we should
never forget that data and
00:12:02
algorithms are a means to an
end. What matters in the end is
00:12:06
the decisions that we make.
00:12:08
- This is the book, <i>Decision-
Driven Analytics. </i>Where can
00:12:11
people pick it up right now?
00:12:13
- On Amazon or any other platform
where you buy your books, or at the
00:12:16
Penn store.
00:12:18
- Stefano, great to talk to you.
Thanks very much for your time.
00:12:21
Stefano Puntoni from here at
the Wharton School. And again,
00:12:23
the book, <i>Decision-Driven
Analytics.</i> - Thank you for
00:12:27
listening to the Ripple Effect.
We hope you found this episode
00:12:30
informative and engaging. Don't
forget to subscribe and leave us
00:12:33
a review so that we can continue
to bring you the best insight
00:12:37
from the Wharton School.
