Nicole Wong at the FTC Town Hall on Behavioral Advertising


Thank you for the invitation
to participate today. My name’s Nicole Wong. I’m the deputy general counsel
for Google and one of my responsibilities is privacy
at our company. I’m going to cover today our
approach to privacy generally, as well the types of data we
collect for purposes of serving relevant advertising
to our users. At Google, we spend a lot of
time studying our own business because the entire industry
is changing so rapidly. In the online advertising
world, we found two things to be true. First, advertising
is a critical component of the web ecosystem. When we do our job well, we
connect consumers with information they want at
the time they want it. Online advertisers and
publishers, including small businesses and bloggers are
flourishing because of the ability to reach their consumers
in an effective and efficient way. And billions of dollars of
services and information are provided today for free– or nearly free– because of online advertising. The second thing we found to
be true is that our users’ trust and their privacy are
critical to our business. Because we support open
platforms, as Tim Armstrong was describing to you this
morning, our users are free to pick up and leave. And because
of that we have to work every day on every product
to earn their trust and their business. In advertising, we’ve created
a very robust ad platform without having to use much
information about a user it all in order to effectively
target the ads. And I’ll describe those systems
in more detail. Our business depends on getting
this balance right. And we’re committed to
continuing to provide the benefits of online advertising
in a way that protects user privacy. Let me spend just one minute
talking about our team. I am enormously fortunate to
work at a company where privacy isn’t just the lawyer’s
problem, instead it’s a value that’s affirmed
throughout the company, from our engineers through
our executives. And for that reason, our
approach to privacy is not to solve a privacy problem
by having a well-worded privacy policy. Although, I’ll tell you that
we spend a lot of time on those policies, and on things
like our recently released videos about what a log file
is or what a cookie is. And I hope you will check out
our new Google privacy channel on YouTube. But in addition to that, we work
really hard at designing privacy protections into
the product itself. The team that drives the process
looks like this– with a lot of experienced
leadership at the top, and importantly, attorneys who are
embedded with the products to make sure that the products
are designed with privacy in mind. We also have, of course, support
teams and security teams who are experts
at what they do. So let me turn to our
advertising offerings. They are basically two. We try to connect with consumers
when they search, known as our AdWords product,
and consumers when they visit websites, where we also
display ads, or in our AdSense product. This is a simplification, but
it’s necessary given the time constraints. Both of these offerings are
contextual advertising. So they get results based mostly
on the current context of the user, not on the user’s
past behavior or profile. So first, to look at AdWords. We’re connecting with consumers
when they search. And a Google AdWords advertiser
will purchase text ads, which you can see up here,
the mutual funds ad, that are provided in response
to a search query that’s entered into our
search engine. So the advertiser will design
that text ad, choose a keyword that triggers the ad– in this
case it’s mutual funds. The advertiser picks a language
and the geography it wants to target. And then the advertiser decides
how much it wants to pay when the user clicks
on that ad. With this inventory of ads,
Google will then match the ad to the chosen keyword. We check for the language
preference of the user. We check for the IP address in
order to get the geolocation, and then we algorithmically rank
the ads for relevancy to the users based on the quality
score that has to do it whether the landing page is of
quality, whether there’s a lot of clickthrough rate
on that ad. And then finally, we
run the auction. The advertiser then pays when
a user clicks on their ad. Importantly, we are targeting
here based on the user’s search term, not on a profile. So in our experience, ads are
more useful, and thus, more effective, when we can correctly
identify what the user is looking for
in that moment. This is in contrast to
behavioral targeting that’s based on a profile built on
past activity in order to target an ad. Our AdSense service works very
much the same, except instead of using keywords to target we
use the content of the page. So this is a page by Seat Guru
which is used to tell you what the best seats on the
airplane are. And we take terms based on that
in order to target the ads, much in the way
that key words use. Again, importantly here, we’re
matching on very limited information, and to
be really specific about what we collect. When a user comes to our site,
they never have to register to use Google. You can go up to any internet
kiosk, any computer and type in a search without registering
with us. And at that time, the only thing
we collect is standard log information– URL, IP address, basic
information about your computer and a cookie ID. The same is true when you view
an ad on one of our AdSense network partners– IP address, URL, time and
date, and the ad viewed. By the way, descriptions of the
types of stuff that’s in a log file is actually in our
privacy policy, and also in a recently released video that
describes what it is. So let me finally end with how
we protect user privacy. As I was saying, we deliver
timely, relevant ads with very little user information,
instead we use contextual targeting. We were also the first major
search engines to announce a finite logs policy of 18
months, after which we anonymize the IP addresses
and cookies. And our cookies expire
after two years. We limit the disclosure
of data. We don’t transfer P2 data
to advertisers. We have a team that’s dedicated
to reviewing all requests for user information
from the government or any other third party. And we have strong expert teams
for network security, software engineering, physical
security, to protect all of these systems. Leaving you with
a final word– this is a very important
discussion for us to be having across the industry. This is a very complex
business, with many stakeholders. And the third thing we found to
be true is that the online advertising industry
is evolving. So it’s appropriate for us to be
reviewing our practices in light of those changes, with an
eye to the continued health of the web ecosystem and
to the trust and privacy of our users. Thank you very much.

2 thoughts on “Nicole Wong at the FTC Town Hall on Behavioral Advertising

  1. the audio is slightly ahead, it makes it look like she is lip syncing. It makes it slightly annoying.

    I like your description of how advertising works. It helped me understand exactly how the advertising works! ;P

Leave a Reply

Your email address will not be published. Required fields are marked *