Sie sind auf Seite 1von 4


Creative people
arent the only ones
needed in radicalinnovation groups

Hard-to-read type
improves readers

Technology adoption
affects national

Erskine Bowles on
making sacrifices to
invest in Americas


The New Science of Viral Ads

Five techniques can help companies make commercials
that people will watch and share by Thales Teixeira


ts the holy grail of digital marketing:

the viral ad, a pitch that large numbers
of viewers decide to share with family
and friends.
Several techniques derived from new
technology can help advertisers attain
this. In our research, two colleagues and I
use infrared eye-tracking scanners to determine exactly what people are looking
at when they watch video ads. We also use
a system that analyzes facial expressions
to reveal what viewers are feeling. These
technologies make it possible to isolate elements that cause people to stop watching

and to find ones that keep them engaged.

In addition, they make it possible to determine what kinds of ads are most likely
to be shared and what types of people are
most likely to share them. Here are five big
problems online advertisers face, along
with solutions that have emerged from our

Prominent Branding Puts

Off Viewers

When people watch ads, they focus on a

few things, such as the actors mouths and

eyes. They also focus on logos. This isnt

the boon it might seem: The more prominent or intrusive the logo, the more likely
viewers are to stop watching even if they
know and like the brand. Why? People
seem to have an unconscious aversion to
being persuaded, so when they see a logo,
they resist.

The solution: Utilize brand pulsing. Smart advertisers unobtrusively

weave the brand image throughout the ad.

Experiments have shown that this can increase viewership by as much as 20%. One
of the best examples of the technique is
March 2012Harvard Business Review25


Coca-Colas animated Happiness Factory

ad. (Like all the other videos referenced in
this article, its available on YouTube.) It
depicts a fantasy version of what happens
inside a Coke machine when someone inserts money. A Coke bottle is shown repeatedly, but each appearance is quick; you can
almost imagine that the story would work
without the bottle. In fact, a good question to ask when conceiving an ad is: If I
removed the brand image, would the content still be intrinsically interesting? If the
answer is yes, viewers are more likely to
keep watching.

People Get Bored Right Away

After recording viewers expressions with

video cameras, we use automated technology that measures the distances between
various parts of the face to identify smiles,
frowns, and other expressions that correlate closely with emotions. (Previous research relied on human coders; automating
the process improves accuracy and allows
for a much larger sampling.) After analyzing thousands of reactions to many ads,
second by second, and tracking exactly
when people stop watching, we found that
keeping viewers involved depends in large
part on two emotions: joy and surprise. To
maximize viewership, its important to
generate at least one of these responses
early on. Traditionally, though, advertisers
have constructed narratives that escalate
toward a dramatic climax or a surprise ending. Such commercials may have worked
on TV decades ago, but todays online
viewers need to be hooked in the opening

The solution: Create joy or surprise

right away. Two videos stand out for elic-

iting these emotions at the start. In one, the

familiar Apple spokesman is joined by Mr.
Bean, who dances crazily for the remainder of the spot. (The video, it turns out, is
not an official Apple ad but a well-crafted
parody of Apples Get a Mac series.)
Bud Lights Swear Jar ad opens with a
surprise: When an office sets up a jar that
workers must pay into as a penalty for pro26Harvard Business ReviewMarch 2012

fanity, one employee immediately curses

because he knows the money will be used
to buy Bud Light. Both videos hook people


People Like an Ad but

Wont Share It

Getting time-crunched viewers to watch a

60-second ad is no small feat, but it wont
necessarily make the ad go viral. Experiments I conducted on my own demonPeople Watch for a While
strate that even though people may enjoy
But Then Stop
Although the Mr. Bean video initially suc- an ad themselves, they wont always send
it to others. In particular, I found that alceeds in attracting viewers, it doesnt keep
though shock may get people to watch an
them watching. Thats because the joy
ad privately, it often works against their dethe video creates is delivered at a fairly
sire to share the spot.
constant level. Weve found that ads that
Bud Lights Clothing Drive ad uses the
produce stable emotional states generally
arent effective at engaging viewers for same cast, setting, and general structure as
Swear Jar. Here, an office worker tries to
very long.
The solution: Build an emotional create enthusiasm for a charity drive by ofroller coaster. Viewers are most likely to fering a Bud Light for every article of used
continue watching a video ad if they ex- clothing donated. The characters respond
by removing clothes theyre wearing, and
perience emotional ups and downs. This
the scenes that ensue contain increasfits with psychological-research findings
ing degrees of nudity (private parts are
about human adaptability. When we come

Videos that deliver constant levels of

joy or surprise dont engage viewers
for very long. Advertisers need to
build an emotional roller coaster.
into a warm home on a cold winter day, or obscured by black bars). Like Swear Jar,
when we receive a pay raise, we experience Clothing Drive garnered high viewership.
But unlike Swear Jar, it was not widely
pleasure, but the feeling is transitory; the
shared. The nudity was too shocking.
novelty soon wears off. So advertisers need
to briefly terminate viewers feelings of joy
The solution: Surprise but dont
or surprise and then quickly restore them, shock. Consider Evians Roller Babies
creating an emotional roller coastermuch
ad, which features computer-generated
the way a movie generates suspense by al- infants roller dancing to a hip-hop song. It
ternating tension and relief.
uses all three strategies suggested above.
The Swear Jar video makes skillful
The brand is relatively unobtrusive but apuse of the roller-coaster technique. The
pears frequently throughout the 60-second
opening scene, which sets up the ads con- spot. Within seven seconds of the opening,
ceit, lasts just 15 seconds. The remainder viewers see an infant on roller skates movof the 60-second spot consists of seven
ing his head rhythmically, like a rappera
scenes with bleeped-out profanities, each
scene thats sufficiently surprising to hook
conveying its own surprise and humor. By
them. The rest of the spot consists of 11
delivering a fresh dose of these elements
different scenes of infants executing deevery six seconds or so, the ad holds on to
lightful dance moves. Unlike the Mr. Bean
its viewers.
video, in which the dancing is continuous,



this ad cuts from scene to scene, modulating the viewers joy and offering repeated
surprises. Roller Babies has been viewed
more than 50 million times on YouTube. In
the world of viral ads, thats a home run.

ANALYSIS: Instead of putting
the logo front and center, this ad
weaves it unobtrusively throughouta tactic that can increase
viewership by 20%.
YOUTUBE VIEWS: 5.4 million


People Still Wont Share the Ad

Even when an ad has been perfectly tailored to go viral, only a subset of those who
watch it will share it. In fact, my research
shows that whether or not an ad is shared
depends as much on the personality types
of viewers as on the ad itself.

The solution: Target the viewers

who will. Ive identified two attributes of

people who frequently share ads: Extroversion and egocentricity. The first is hardly
surprising, but the second is, at least on the
face of it. Why would egocentric people be
inclined to sharean act thats usually associated with helping others? I believe that
in many cases its because they are looking to increase their social status. Their
primary aim in posting or e-mailing an
ad link isnt to make others joyful; its to
display their own taste, media savvy, and
Its hard to target viewers on the basis of
personality type, but thats apt to change as
social media evolve. For instance, companies are already placing ads on the pages of
Facebook users who frequently post links
and are reaching out to Twitterers who
have large followings. The ability to find
these archetypal sharers will become just
as important as the ability to reach certain
demographic groups has traditionally been.
AS VIEWERS gain increasing control over
which ads they sit through, advertisers will
have to become more consumer-centric.
Theyll need to think harder about the
value a video offers to the viewer, instead
of considering primarily how well the
video serves the brand. The result will be
ads that are both more effective and more
HBR Reprint F1203A


ANALYSIS: Unlike traditional TV
ads that build to an emotional
climax, this video elicits joy in its
opening seconds, immediately
hooking viewers.
YOUTUBE VIEWS: 5.5 million

ANALYSIS: Eight scenes, each
with its own punch line, deliver
the interrupted doses of joy and
surprise needed to sustain viewers interest.
YOUTUBE VIEWS: 5.5 million

BUD LIGHTClothing Drive
ANALYSIS: The relatively low
number of views suggests that
people felt inhibited about
sharing this ad with family and
YOUTUBE VIEWS: 1.9 million


EVIANRoller Babies
ANALYSIS: This ad has everything: It opens on a joyful note,
cuts between scenes for an emotional roller coaster, and isnt too
edgy for people to share.
YOUTUBE VIEWS: 51 million

Thales Teixeira is an assistant professor at

Harvard Business School.
March 2012Harvard Business Review27

Harvard Business Review Notice of Use Restrictions, May 2009

Harvard Business Review and Harvard Business Publishing Newsletter content on EBSCOhost is licensed for
the private individual use of authorized EBSCOhost users. It is not intended for use as assigned course material
in academic institutions nor as corporate learning or training materials in businesses. Academic licensees may
not use this content in electronic reserves, electronic course packs, persistent linking from syllabi or by any
other means of incorporating the content into course resources. Business licensees may not host this content on
learning management systems or use persistent linking or other means to incorporate the content into learning
management systems. Harvard Business Publishing will be pleased to grant permission to make this content
available through such means. For rates and permission, contact