Safety First: How to Approach and Execute Safe Media Buying Online

SAFETY FIRST: HOW TO
APPROACH AND EXECUTE
SAFE MEDIA BUYING ONLINE
2 0 1 8
+
© 2018 by Fullscreen. Privileged and conﬁdential.

The evolution of brand safety can be attributed to
the technological advancements in media
placement. Here’s what we’ll cover:
WHAT IS BRAND
SAFETY AND WHY
DOES IT MATTER?
‣ The History: From Print To Programmatic
‣ What’s Changed: Alignment, Context, and Semantics
‣ What’s Available: Controls, Risks, and Rewards
‣ The Future: Why Transparency Is Key

SUB-SECTION TITLE
CAN GO ON TWO LINES
THE HISTORY

BRAND SAFETY CHALLENGES
EXISTED LONG BEFORE THE
INTERNET.
“
“

HISTORY OF BRAND SAFETY
HOW WE GOT HERE
Ads started to appear in
newspapers. Human
judgment was used to
determine ad placement.
FIRST AD APPEARS  
IN A NEWSPAPER
1704 18941891 1941 1994
With ad agencies starting to
open, the responsibility of
brand safety and proper
curation at scale had a
consolidated place to live.
FIRST AD  
AGENCY OPENS
Brands now started to
have more characteristics
that made ad placement
complicated due to brand
alignment needs. This
made brand safety more
challenging to manage.
BRAND IDENTITIES  
START TO FORM
Ads started appearing in
family living rooms in-
between shows and
sometimes featured in
shows or broadcast news
segments.
FIRST TV  
COMMERCIAL APPEARS
This ﬁrst banner ad from
AT&T showed you all the
places you could go with
the internet. This opened
the door for the endless
possibilities that have led
to the complicated brand
safety world we live in
today.
FIRST AD ON THE
INTERNET

HISTORY OF BRAND SAFETY
According to CNNMoney,
pre-roll ads from
companies like Procter &
Gamble (Swiffer),
Anheuser-Busch (Bud
Light), and Johnson &
Johnson (Aveeno) were
placed before ISIS
propaganda videos and
jihad-themed videos.
ADS BEFORE ISIS VIDEOS
STARTED SHOWING UP
MAR
2015
Havas Media, one of the largest European advertising and
marketing firms, decided to pull all YouTube UK spending
after their ads were spotted on ISIS recruitment videos. A
week later some of the biggest US advertisers followed,
pausing all non-search Google advertising, including AT&T,
Dish, FX, Networks, General Motors, J&J, McDonalds,
PepsiCo, Starbucks, Verizon, Walmart, and more.
ADS SHOW UP AGAIN ON EXTREMIST
VIDEOS OF MAJOR BRANDS LEADING
TO YOUTUBE BOYCOTT
YouTube’s first attempt to
limit monetization on
unsafe content for kids.
However, this backfired
with LGBTQ influencers
now seeing their content
hidden as the platform
quickly deemed channels
“unsafe”.
YOUTUBE UPDATED
POLICIES FOR CREATORS
On May 25, 2018 the General Data Protection Regulation
went into place in the EU. Although this is not directly tied to
Brand Safety it is the first major step in government agencies
getting involved in data regulation online. Sunsetting the
ability of demographic and behavioral targeting makes
brand safety more challenging. This is forcing advertisers and
marketers to be more innovative than ever.
GENERAL DATA 
PROTECTION REGULATION
Although most brands returned to running
ads on YouTube and elsewhere online,
many major players are still boycotting.
This lack of confidence in brand safety is
due to the fact that no one solution can
guarantee the unknowns that come with
safety concerns. Marketing and
technology companies must come up with
solutions that allow a scale for human
judgment.
BRANDS STILL BOYCOTTING
MAR
2017
NOV
2017
MAY
2018
TODAY
WHERE WE ARE TODAY

BRAND SAFETY IN TERMS OF BRAND
ALIGNMENT
BRAND SAFETY IN TERMS OF BOT TRAFFIC
AND FAKE FOLLOWERS
TWO MAIN CATEGORIES

Changes in technology have allowed us to target
very speciﬁc audiences and content. 
 
These same technological advancements have
created new concerns for brands as our
understanding of where an ad is placed and what
that environment might look like has become much
more dynamic and complex.
BRAND ALIGNMENT
EXAMPLE

THE BRAND SAFETY BASICS
Aligning your brand with popular content
can help to increase visibility and targeted
recall. If executed improperly, this can also
lead to catastrophic failure. Alignment is
primarily about location.
‣ Insurance and Disasters
‣ Luxury Brands and Luxury Events
‣ Sports and Alcohol
ALIGNMENT
Contextual targeting takes brand
alignments one step further by thinking
about the tone and opportunity to adjust
or integrate your brand story for a specific
time and place.
‣ Memorial
‣ Celebratory
‣ Curated
CONTEXT
As we move from basic alignments to
more advanced contextual targeting and
messaging, we run the risk of
misunderstanding the context. Semantics
reference the variations of meaning
behind contextual touch points.
‣ Jaguar Car vs. Jaguar Cat
‣ Stock Market Crash vs. Car Crash
‣ Woman’s Purse vs. Racing Purse
SEMANTICS

SUB-SECTION TITLE
CAN GO ON TWO LINES
WHAT’S CHANGED

THE WORLD
HAS CHANGED
400 hours of video
are uploaded to
YouTube every
MINUTE
Daily Active YouTube
users
VIDEOS WATCHED
A DAY
5B
3 0 M I L L I O N
[Source : https://www.omnicoreagency.com/youtube-statistics/]© 2018 by Fullscreen. Privileged and conﬁdential.

TRANSPARENCY IS EVERYTHING.
IF WE CAN’T SEE IT, MEASURE
IT, OR REPORT IT, WE CANNOT
CONTROL IT.
“
“

AS THE MEDIA
MARKETPLACE EVOLVES,  
WE SEE BOTH INCREASED
RISK AND REWARD, BUT
ALSO A DECREASE IN
TRANSPARENCY

THESE CHANGES IN OUR BUYING METHODOLOGIES
EFFECT THE AVAILABLE CONTROLS AND POTENTIAL RISK
‣ Publisher/Paper
‣ Section of Paper
‣ Location In Section
‣ Context and Competition
PRINT
‣ Publisher/Website
‣ Section of Site
‣ Location In Section
‣ Context and Competition
PUBLISHER DIRECT
DIGITAL MEDIA
‣ Demographics
‣ Psychographics
‣ Rate
‣ Desired Outcome
‣ Whitelist/Blacklist
‣ Negative and Positive
Targeting
PROGRAMMATIC
DIGITAL MEDIA ‣ Demographics
‣ Psychographics
‣ Rate
‣ Desired Outcome
‣ Negative and Positive
Targeting
SOCIAL MEDIA

Social media platforms have changed the game. 
With user-generated content and a limited
understanding of each unique user’s experience on
site/application, we often don’t even know what
content ads are aligned with or who is actually
engaging with them. 
These are just a few of many unknown unknowns.
HAVE WE FULLY
RECOGNIZED THE
RISK?

SUB-SECTION TITLE
CAN GO ON TWO LINES
WHAT’S AVAILABLE

An increased focus for brands, a reduced appetite
for risk, a marketplace full of new controls.
‣ Available Tools and Services
‣ Limitations and Unknowns
WHAT IS THE STATE  
OF BRAND SAFETY
TODAY?

BRAND SAFETY TODAY
Brand Safety is a fast-moving segment of
the media industry and some solutions
have become commonplace.
AVAILABLE TOOLS AND SERVICES
While many advanced tools are now
available, they all still have various
limitations. The more complex the content,
the more difficult this becomes.
LIMITATIONS
When entering more dynamic landscapes
such as social media or video content,
whereby advertisers often cannot control
user engagement, we don’t always know
where things may go wrong.
UNKNOWNS

THERE IS MORE THAN ONE PATH TO BRAND SAFETY

SUB-SECTION TITLE
CAN GO ON TWO LINES
THE FUTURE

What we know is coming, and what we expect to see.
‣ New Tools
‣ Manual Processes
‣ Fullscreen Best Practices
WHAT WILL BRAND
SAFETY MEAN AND LOOK
LIKE GOING FORWARD?

THE CHALLENGE…THE VOLUME OF VIDEO CONTENT IS
TOO LARGE FOR COMPREHENSIVE MANUAL REVIEW.
… [AND] EASY WAYS TO SCAN… ARE RARE.
THE SOLUTION TO THIS CHALLENGE REQUIRES
TECHNOLOGY THAT CAN DETECT CONTENT THAT IS
UNSAFE (SUCH AS THOSE TOPICS RELATED TO DRUGS,
TERRORISM, OR HATE SPEECH) THROUGH A FULL-PAGE
ANALYSIS USING PAGE TEXT AND METADATA AND, IN THE
CASE OF VIDEO PLATFORMS, VISUAL IDENTIFICATION.
– JOHN MONTGOMERY
Global EVP, Brand Safety GroupM
[Source: Moat: Deﬁning Brand Safety-White Paper, September 13, 2017]
“
“

RECENT BRAND SAFETY
DEVELOPMENTS
‣ Eligibility requirements for content monetization on YouTube have grown significantly,
increasing the quality threshold required to be included in the YouTube Partner
Program (YPP). Now, only channels with at least 1,000 subscribers and 4,000 watch
hours over the last 12 months will be included.
‣ Google has made a large resource commitment to protect brand safety, as they’ve
promised a huge hiring wave – ultimately, they will have over 10,000 people by the
end of the year.
‣ These hires will be responsible for one of the most impactful changes that Google
has implemented: moving forward, Google Preferred lineup ads will only run on
videos that have been human-verified.
‣ Google is also revising their content filtration settings when targeting ads, allowing for
“conservative”, “standard”, and “edgy” settings, while making the reach implications
clear.
‣ Furthermore, they are promising full transparency by pushing forward integrations
with Integral Ad Science, DoubleVerify, and other independent brand safety tools.
[Source: https://www.iprospect.com/en/us/our-blog/google-youtube-tackle-brand-safety/]© 2018 by Fullscreen. Privileged and confidential.

MODERN CHALLENGES OF BRAND SAFETY
Fully-automated or fully-
manual vetting is not
scalable across entire
content libraries to properly
protect brands.
Brand subjectivity to what is
safe and what isn’t depends
on the brand itself and the
audience.
More transparency is
needed in the industry and
on a proactive basis from
platforms and partners.

RECENCYRELEVANCY
RELEVANCE AND RECENCY
CAMPAIGN SPECIFIC CONTENT THAT PROVIDES RELEVANCE AND RECENCY YIELDS OPTIMAL BRAND PERFORMANCE. HAVING A WHITELIST OF
MILLIONS OF VIDEOS DOES NOT PROVIDE FOR RELEVANCE AND RECENCY.
OPTIMAL PLAN
DEVELOPMENT
PRIOR TO BRAND
SAFETY MEASURES
‣ POPULARITY
‣ SCALE
‣ VELOCITY
CONTEXTUAL AND
AUDIENCE TARGETING
PERFORMS 2.5X BETTER

BRAND SAFETY IS NOT UNIVERSAL
BRAND-SPECIFIC
Brands must define according to
their values and audience.
EXAMPLES
‣ Competitor content
‣ News events
‣ Violent entertainment
UNIVERSAL
Appearing next to unambiguously
unsafe content.
EXAMPLES
‣ Hate speech
‣ Terrorism
‣ Violence and death
WHAT IS IMPORTANT TO YOUR BRAND?
[Source: Moat: Deﬁning Brand Safety-White Paper, September 13, 2017]© 2018 by Fullscreen. Privileged and conﬁdential.

Manual 
Review
‣ Shield is the first-ever BRAND SAFETY TOOL that identifies and flags
potentially unsafe videos prior to launching media campaigns by assigning
video ratings that trigger human reviews - ensuring advertisements are
aligned only with safe environments.
‣ Utilizes IMAGE and AUDIO recognition in addition to analyzing signals from
meta data, tags, titles, comments, and ratings.
‣ Key differentiator is Shield’s ability to review every 3 seconds of visual
identification (not just the thumbnail) and all audio.
‣ The data obtained from these signals is audited against our corpus of known
unsafe terms while its MACHINE LEARNING capabilities, constantly informed
by a human layer, discover new unsafe words in real-time in an evolving
landscape.
‣ An easy-to-understand score is generated based on weight of signal triggers
and results are bucketed into rating groups.
TV-PG
PARENTAL GUIDANCE SUGGESTED
TV-14
PARENTS STRONGLY CAUTIONED
TV-G
GENERAL  
AUDIENCE
TV-MA
MATURE AUDIENCE  
ONLY
N/A
TRIGGERS HUMAN REVIEW
PROCESS
BLACKLIST

BENEFITS OF FULLSCREEN SHIELD
Machines + Humans =
Most thorough review process
Constantly evolving and
learning algorithm
Relevancy and Recency are
key - Massive whitelists are
not scalable and do not
allow for recency
Proprietary access to
brand-safe video inventory

‣ Brand safety cannot be solved with a pure automation
or pure manual approach — it needs both.
‣ Brand safety is subjective — conversations must
happen prior to launch and throughout ad campaigns.
‣ There is too much UGC content posted daily to
manually vet each potential placement.
‣ Whitelisting does not allow for scale or recency of
placements.
‣ To truly understand the context of a video or image —
it’s necessary to look “inside” the video, the audio and
all the meta data available.
KEY TAKEAWAYS

WHAT IS BRAND SAFETY
WHAT WE ARE NOT TALKING ABOUT, BUT HOW IS IT RELATED?
Fake news has become highly
politicized the last several years.
Although the battle in mitigating
fake news is not the same as
ensuring brand safety, it is related.
If ads show up on or next to
stories that are revealed to be
fake news, or are even accused
of being fake news, it can be
harmful for brands. This adds a
huge layer of complexity to
ensuring brands are being safe.
FAKE NEWS
The GDPR was implemented as a
result of brands overreaching and
platforms oversharing. It directly
impacted the businesses that
relied on certain demographic
and behavioral targeting. This
inability to target in certain ways,
although somewhat helpful in the
long run, makes brand safety
assurance even more difficult as
brands try to find work-around
targeting tactics.
GDPR
Many influencers and publishers
saw less monetization on their
videos in the last year. This is due
to platform updates that are
working towards rewarding
people that produce successful,
brand-safe content. Publishers
and influencers reliant on this
revenue that weren’t deemed
brand-safe took a hit. Although
this does not fix the issue of
brand safety, it changes the
environment and encourages the
creation of brand-safe content.
DEMONETIZATION
Many platforms and 3rd party
tools have added updates to
whitelist/blacklists channels,
pages or videos directly. These
updates are the first steps to aide
in brand safety but are not the
long term solution as the nuances
are specific to each brand.
PLATFORM CHANGES

THANK YOU
FOR MORE INFORMATION, PLEASE CONTACT:
MARK WILLIAMS - MARK.WILLIAMS@FULLSCREEN.COM

Safety First: How to Approach and Execute Safe Media Buying Online

More Related Content

What's hot (20)

Similar to Safety First: How to Approach and Execute Safe Media Buying Online (20)

More from Association of National Advertisers (20)

Recently uploaded (20)

Safety First: How to Approach and Execute Safe Media Buying Online