Transcript Diapositiva 1

Mapping Online Advertising:
From Anxiety to Method
Fernando Bermejo
Fellow, Berkman Center for Internet &
Society, Harvard University
January 12th, 2010
From Anxiety…
“Half the money I spend on advertising is
wasted. The trouble is, I don’t know which
half”. John Wanamaker (1838-1922)
…to Method
“The time has come when advertising has in
some hands reached the status of a science. It
is based on fixed principles and is reasonably
exact. The causes and effects have been
analyzed until they are well understood. The
correct methods of procedure have been proven
and established”. Scientific Advertising (1923), Claude
C. Hopkins
From Anxiety…
“47% of campaigns failed—how to make sure
yours sticks. 83% of ad spending was seriously
suboptimal—how do you get more impact without
spending a dime more?” What Sticks: Why most
advertising fails and how to guarantee yours succeeds
(2006)
…to Method
“In terms of efficiency, if not size, the
advertising industry is only now starting to grow
out of its century-long infancy, which might be
called ‘the Wanamaker era’”. The Economist
(07/06/2006)
The source of anxiety
Gift Economy/
Peer-Production
Property / Pay Model
Advertising…
Advertisers
Media
Audiences
ADVERTISERS
MEDIA: Ads
+ Content
PRODUCT &
SERVICES
AUDIENCES /
USERS
The Method
Collecting information
 Metrics
 Targeting
 Having the right context
 Pricing
 Funneling

Collecting information

Previous media: Audience and Market Research,
Surveys and Panels.

Online: Every interaction can be automatically
recorded. Every page viewed, every ad viewed,
every song or video played, every search
performed, every content created…
Collecting information
From discrete events to profiles: 1st-party
cookies, 3rd-party cookies, web bugs, and flash
cookies (100KB of information, no expiry date, stored in separate
location, not controlled by the browsers…)
Berkeley study on flash cookies (Soltani et al, 2009):
– 54 of top 100 sites use flash cookies (only 4 mention it in their
privacy policy).
– Even with a NAI opt-out cookie, flash cookies are employed.
– 31 had at least an overlap between a HTTP and a Flash cookie
(90% of the overlaps from cookies served by ad networks)
– Respawing of HTTP cookies occurred on several sites (Hulu.com,
About.com) and took place accross domains (Clearspring: Aol.com
and Mapquest.com).
Collecting information
Extent of 10 Top Third-Party Families
Growth of the Google family
“Our results show increasing aggregation of userrelated data by a steadily decreasing number of
entities” (Kryshnamurthy & Willis 2009)
Metrics
Unduplicated Unique Users
Hits
Targeting
From audiences, to segments, to increasingly
granural criteria (individual profiles)
Demographics
 Content consumed
 Content produced

– Search Keywords

Behavior
Targeting
segments
Targeting
ADVERTISERS
MEDIA: Ads
+ Content
PRODUCT &
SERVICES
AUDIENCES /
USERS
Having the right context
Both in terms of the general environment and
in terms of specific content…
 In previous media:
– professionalism, organizational norms, selfcontrol

Online:
– Production of specific content directly for
advertising purposes: e.g. Demand Media,
etc.
– Need to control user-generated content: e.g.
the new Technorati, Youtube video demotion,
etc.
Pricing
Previous media: Exposure in terms of CPM
 Online:

– CPM still 38%
– Performance: 58%
 Cost per Click (CPC)
 Cost per Action: lead … purchase
Pricing
ADVERTISERS
MEDIA: Ads
+ Content
PRODUCT &
SERVICES
AUDIENCES /
USERS
Funneling
The most uncontrolable part:
 Branding
 Tracking and Retargeting
Yahoo! UK: “We know where users have been and we can
tell where they're going”
AOL: “For example, LeadBack tracks users who have visited
your website but have not made a purchase. When
those same users visit other sites on our network,
LeadBack delivers your ads to them, driving them back
to your site and increasing the likelihood of a sale.”

Placing more customer activities in
controllable environments: The mobile
internet
Consequences
Collecting Information: A requires
increasing information from individuals.
 Metrics: A is not comfortable with
phenomena that cannot be easily
controlled, measured and reduced to
numbers.
 Targeting: A tries to move to the form of a
transaction between advertisers and
individuals, leaving third parties as
accesorial.

Consequences
Generating the right context: A prefers
certain environments and shapes content.
 Pricing: A tries to reduce its risk by asking
others to assume it.
 Funneling: A tries to be more intensive
and more extended.

Consequences
Privacy: Playing cat and mouse (?)
Content: De-intermediation (?)
Structure: Advertising as steroids (?)
The online content layer…
Gift
Pay
Advertising
Mapping Online Advertising:
From Anxiety to Method
Fernando Bermejo
Fellow, Berkman Center for Internet &
Society, Harvard University
January 12th, 2010