one year later
In late 2017 and then a year later in 2018, we asked brands about the overall state of brand safety in the world of digital marketing. Overall, marketers say that some of the industry’s recent brand safety treatments have been effective — but they’re worried some of them may be overly restrictive.
0%
48%
insufficient viewability
41%
contextual irrelevance
42%
lack of brand awareness
Only 10% use it specifically for ad targeting.
Download the guide
360-degree video
Virtual reality
Augmented
reality
Wearables
Computer
Vision
SEAWORLD
CRISIS
The brand safety
Time for a check-up:
Download our full guide to brand safety here.
Then and now
Here’s how many digital media pros said brand safety was a serious problem for their marketing efforts in late 2017:
90%
60%
But at the end of last year, it was:
The platforms took preventative measures
Major social media companies have made big strides, enabling more third party measurement and booting toxic accounts.
Social platforms have certainly taken brand safety seriously.
- Joe Barone, managing partner of brand safety, GroupM
When it comes to improving on brand safety efforts over the past 12 months, which of the following platforms’ efforts would you rate as fair or better?
Twitter sanitized its platform, and marketers approved
In 2017, digital media pros ranked Twitter dead last in brand safety:
Which of the following platforms is the most brand safe?
99%
99%
98%
98%
93%
93%
97%
Twitter
Facebook
Snapchat
Instagram
LinkedIn
Publisher sites
YouTube
45%
26%
14%
10%
3%
1%
LinkedIn
Facebook
Publisher sites
YouTube
Search
Twitter
Between November 2017 and November 2018, which of the following platforms was the most brand-safe?
But after a year of booting unsavory and fraudulent accounts from its platform, Twitter surged into the lead:
Search
Twitter
Publisher sites
Facebook
Instagram
LinkedIn
45%
5%
4%
2%
25%
17%
In came the
specialists
Over the past year, many brands have hired their own brand safety specialists. Their value may be somewhat limited.
Has the introduction of a brand safety specialist or specialists reduced your company’s exposure to brand-unsafe content?
The threat vectors are different
What types of brand-unsafe material have you most commonly encountered adjacent to your content?
Brands are still exposed to unsafe content. But between 2017 and 2018, things changed.
17%
17%
17%
Competitor's branding
Bad news
Vulgar language
Division politics
Fake news
Hate speech
45%
38%
25%
17%
42%
36%
32%
45%
30%
2017
2018
Treatments are
changing too
What have been your top three most effective brand safety solutions?
Between 2017 and 2018, digital media pros started gravitating toward different brand safety treatments.
Direct relationships
with publishers
35%
31%
47%
2017
2018
62%
39%
37%
Blacklists
Natural language
context detection
Blacklists
Keyword
detection
Image
recognition
The surge in marketers citing blacklists may reflect an increased fear of publishers.
I think the market has resulted in the need for these blunt force instruments.
- Ryan Pauley, chief revenue officer at Vox Media
But brands are
inoculating themselves
against favorable audiences
Here’s how many marketers said that the use of various brand safety tools, including blacklisting, created an inability to reach specific audiences in late 2017:
A year later
it was:
30%
69%
There are innovative
new therapies
COMPUTER VISION
“Also known as image recognition, computer vision is an AI-driven technology that uses neural networks sort and identify imagery. Among other uses, it can detect brand-unsafe content in real-time and automatically keep marketers out of unsafe environments."
13%
31%
In 2017, here’s how many marketers said natural language context detection was an effective brand safety treatment
A year later, the numbers have surged for both:
18%
And, computer vision
Computer vision
35%
Natural language context detection
Computer vision and natural language context detection may just be the key to inoculating brands against negative adjacencies without keeping them away from favorable audiences.
Download the full guide to learn more.
Yes, a little: 64 %
Yes, moderately so: 31 %
Yes, a lot: 5 %
Click me!
Click me!
