i3 - January/February 2017 - 47

economy. You're going to see businesses get invented that never
existed before and you're going to
see businesses that have existed for
a long time, go away. You're already
watching many examples of this.

How do you approach
machine learning?
Machine learning is applied to
the media buying process. We ask
computers what is the most efficient
way to buy media. It can be applied
to messaging. What is the most
efficient message to put in front
of somebody?
Machine learning can be applied
to the people side of things. What's
the right context? Who is the right
consumer? What is the right offer?
What is the right commerce experience that they want to have?
Machine learning can be applied
to the consumer's post purchase
phase. How do we get somebody, if
they have a good experience, to post
about it? If they had bad experience,
how do we rectify it?
Machine learning is starting to
change the nature of computing
because it is starting to get so much
more anticipatory and insightful.

How does machine learning
look at Omnicom specifically?
Basic machine learning is really at
the core of what we do at our data
and marketing sciences group,
Annalect, and our media buying
businesses OMD, PHD and Hearts
& Science. It is becoming an increasingly important component of our
shopper marketing, healthcare and
PR businesses. Machine learning is
in its early phase; however, it's
touching almost all corners of
Omnicom at this point.

CREDIT

Are we finally at the
inflexion point where ecommerce achieves a dominance people have predicted?
I guess I take a contrary point
of view, which is it's "only" five or

C TA . t e c h / i 3

"We have basically unlimited
computing power."
10 percent of total sales, but that's
up from zero. E-commerce is huge!
Go look at the back of a UPS truck
every time you see the door open.
It's filled with Amazon boxes.
Think of what that's doing to the
economy and how it is shifting an
entire class of consumer behavior.
Whether you're booking an airline
ticket, buying a book, consuming music
or entertainment - we've seen a massive
restructuring of not only how we use our
free time but how we spend our money.
Of course we all go to the grocery
store to buy food, at least occasionally.
But I've never quite understood when
people say it's "only" a small amount.
It's hundreds of billions of dollars, and
it's a global phenomenon. Most places
that are heavily penetrated with digital
have some form of e-commerce.

What are some of the leading
edge things that you're seeing
in e-commerce?
There's a lot less friction involved
in e-commerce. It's much faster. Things
like Amazon Prime have
really sped up the whole process of
acquiring things.
There's also machine learning. Think
about it in terms of e-commerce. It
anticipates what you want. It stackranks things in the same way that
Google does with Page Rank, but
applied to goods and services.
This makes buying things that much
more efficient - sorting through many
similar products to find the exact one
that you want, and then getting it
delivered to your home or office in two
days with free shipping.

What about a brand that is not
stack-ranked in a favorable
way? If I'm not a top brand,
how do I insert myself into
the conversation?

That's a very interesting question
because it's not just a vector around
who sells the most. One of the best
features inside of Amazon is the ranks
around consumer reviews. As a brand
you can really start to affect that by
responding to people that didn't have
a favorable experience.
The second thing that you can do
is buy ads inside of Amazon. If you
want to be up near the top you can
buy that media placement.
Smart brands are realizing that
it's not a one-size thing. It's not just
about price. Or it's not just about
brand. It's a combination of features,
brand, price, quality and ratings.
It's a combination of paid and
unpaid media - in other words,
reviews versus paid. That balance
is really how you build a defensible
position.

What is something that
you believe about the future?
I think about the amount of
automation and machine learning,
based on data that we are going to
see in the next five to 10 years. We
have basically unlimited computing
power. We have the cloud and we
have reams of data coming out of
every transaction.
At Omnicom, we are trying
to bring that all together in order
to be more insightful. We're already
doing this at Annalect and at our
newest agency network that emerged
from Annalect, Hearts & Science.
Hearts & Science launched earlier
this year after being awarded P&G's
business and recently went on to
win AT&T's media buying account.
Of course we're going to see things
like virtual reality and enhanced
mobile and things like that. The
bigger idea here is insights led by
data and machine learning. n

JANUARY/FEBRUARY 2017

47
http://cta.tech/i3
i3 - January/February 2017

Table of Contents for the Digital Edition of i3 - January/February 2017
