The advert censors of ASA have published a five year strategy, with a focus on more censorship of online advertising including exploring the use of machine learning in regulation.
The strategy will be officially launched at an ASA conference in Manchester, entitled The Future of Ad Regulation.
ASA explains the highlights of its strategy:
We will prioritise the protection of vulnerable people and appropriately limiting children and young people's exposure to age-restricted ads in sectors like food, gambling and alcohol We will listen in new ways, including research, data-driven
intelligence gathering and machine learning 203 our own or that of others - to find out which other advertising-related issues are the most important to tackle We will develop our thought-leadership in online ad regulation, including on
advertising content and targeting issues relating to areas like voice, facial recognition, machine-generated personalised content and biometrics We will explore lighter-touch ways for people to flag concerns We will explore whether our
decision-making processes and governance always allow us to act nimbly, in line with people's expectations of regulating an increasingly online advertising world We will explore new technological solutions, including machine learning, to improve
Online trends are reflected in the balance of our workload - 88% of the 7,099 ads amended or withdrawn in 2017 following our action were online ads, either in whole or in part. Meanwhile, two-thirds of the 19,000 cases we resolved last year were
about online ads.
Our guiding principle is that people should benefit from the same level of protection against irresponsible online ads as they do offline. The ad rules apply just as strongly online as they do to ads in more traditional media.
Our recent rebalancing towards more proactive regulation has had a positive impact, evidenced by steep rises in the number of ads withdrawn or changed (7,009 last year, up 47% on 2016) and the number of pieces of advice and training delivered to
businesses (on course to exceed 400,000 this year). This emphasis on proactive regulation -- intervening before people need to complain about problematic ads -- will continue under the new strategy.
The launch event - The Future of Ad Regulation conference - will take place at Manchester Central Convention Complex on 1 November. Speakers will include Professor Tanya Byron, Reg Bailey, BBC Breakfast's Tina Daheley, Marketing Week's Russell
Parsons, ASA Chief Executive Guy Parker and ASA Chairman David Currie.
Online ASA Chief Executive, Guy Parker said:
We're a much more proactive regulator as a result of the work we've done in the last five years. In the next five, we want to have even more impact regulating online advertising. Online is already well over half of our regulation, but we've more
work to do to take further steps towards our ambition of making every UK ad a responsible ad.
Lord Currie, Chairman of the ASA said:
The new strategy will ensure that protecting consumers remains at the heart of what we do but that our system is also fit for purpose when regulating newer forms of advertising. This also means harnessing new technology to improve our ways of
working in identifying problem ads.
A pre-roll ad seen on YouTube in June 2018 for Spotify featured a number of scenes in quick succession and tense sound effects that imitated the style of a horror film. The ad opened with a shot of three characters having breakfast. One
character said, Can you play the wakeup playlist? and they played a particular song from their phone. That was followed by a shot of another character rousing himself and saying, Turn that up. As the music was turned up, a shot showed a horror
film style doll in a dilapidated old room raising its head and tense music was played to accompany the song. Several shots followed of the doll ambushing the characters in the ad whenever they played the song and implicitly attacking them. The
final shots showed one character attempting to convince the other not to play the song. The ad showed the character taking hold of the other character's hand to stop him playing it but then the doll's hand reached out to press play. The final
shots of the ad showed the doll's face alongside text which stated, Killer songs you can't resist.
The ad was seen during a video on the YouTube channel for DanTDM, a gaming channel.
The complainant, who was a parent said their children saw the ad and found it distressing, and objected that the ad was:
unduly distressing; and
irresponsibly targeted, because it was seen during videos that were of appeal to children.
Spotify said that the ad was intended for an adult audience and was particularly targeted towards adults aged 18 to 34. They understood that the tools provided to them by YouTube to target ads towards a particular age group and demographic used a
combination of self-identification by YouTube users and probabilistic data based on the user's behaviour across the internet. Their agency had applied relevant content exclusions including ensuring that the ad was not shown alongside shocking or
graphic content. Additionally they applied a function so that users could skip the ad after five seconds. They noted that the first encounter with the doll in the ad occurred after 12 seconds and that between 7 and 12 seconds the ad introduced
cues as to the tone of the ad so they considered that viewers would have had the opportunity to skip the ad at any point if they considered the content to be distressing.
Spotify provided information from YouTube which listed the demographic data of viewers of logged-in viewers of the YouTube channel on which the ad was seen by the complainant. They explained that the data showed that 89% of viewers of the channel
were aged 18 or over and that most (73%) were aged between 18 and 44. Only 11% of viewers were aged between 13 and 17. Spotify said that the ad had appeared prior to a video about a video game that was marketed as a stealth and horror game.
ASA Assessment: Complaints 1 & 2 upheld in part
The ASA considered that although violence was not explicitly shown in the ad, it was implied. The ad contained several scenes that were suggestive of a horror film, including tense music and scenes of characters looking scared or in distress. In
two scenes in particular, actors were shown playing the song in bed and in the shower when they were ambushed by the doll. We considered that those scenes would be seen by viewers as reminiscent of famous scenes from horror films.
We first considered whether the ad was likely to cause undue distress to adults who saw it. The ad featured shots reminiscent of a horror film. However, we considered a number of scenes, including the doll nodding its head to the rhythm of the
song and the doll's hand pressing the play button on a device that had the Spotify app open, would be seen by viewers as humorous. We considered that although some might find the ad mildly scary, most adult viewers would find the ad overall to be
humorous rather than frightening and it was unlikely to cause distress to them.
However, we did consider that the nature of the ad meant it was not suitable to be seen by children because it was likely to be distressing to them. In particular, the ad contained scenes that had tense sound effects and imagery similar to a
horror film including the implied threat of violence. The fact the ad was set inside the home, including a bedtime setting, and featured a doll, meant it was particularly likely to cause distress to children who saw it. We did not consider that
the context of the ad justified the distress. In addition, the nature of the ad as emulating a horror trailer was deliberately not made clear from the start of the ad and children were likely to be exposed to some of the potentially frightening
scenes before they, or parents viewing with them, realised that was the case. We considered the ad therefore should have been appropriately targeted to avoid the risk of children seeing it.
We considered that the ad may have been appropriate to show before content on YouTube that was unlikely to be of particular interest to children. However, when seen by the complainant the ad was juxtaposed against unrelated content for the video
game Hello Neighbour . Although the video game was marketed as a stealth horror game, it included colourful cartoonish images and was rated by the ESRB as suitable for players aged 10+ and by PEGI as suitable for players aged seven or
older. We therefore considered that it was reasonable to expect that content about Hello Neighbour was more likely to appeal to children.
The figures provided by Spotify showed that 11% of viewers of the DanTDM were between the ages of 13 and 17, based on viewer demographics relating to logged-in users. However, the channel made use of cartoonish imagery and included videos of
video games popular with children and media including Fortnite and The Incredibles. We noted videos on the channel were presented in an enthusiastic manner by a youthful presenter who had won an award from a children's television network. Taken
altogether, we considered that from the content of the videos and presentational style, the channel would have particular appeal to children. For those reasons we concluded that the ads had appeared before videos that were likely to be of appeal
or interest to children.
We concluded that the ad was unlikely to cause distress to adults, but that it was likely to cause undue distress to children. Therefore, because the ad had appeared before videos of appeal to children, we concluded that it had been
We told Spotify to ensure that future ads did not cause distress to children without justifiable reason, and to ensure ads that were unsuitable for viewing by children were appropriately targeted.
Three TV ads for an online retailer for womenswear, Nasty Girl, seen in June 2018:
a. The first ad featured a model posing in various outfits including swimwear, a dress and a tank top with a skirt. The model was also shown playing tennis and golf.
b. The second ad featured the same scenes shown in ad (a), with large on-screen text regarding next day delivery.
c. The third ad was an abridged version of ads (b) and (c) with large on-screen text regarding next day delivery.
22 complainants who believed the model looked unhealthily thin challenged whether the ads were socially irresponsible.
Nasty Gal Ltd stated that the model featured in the ad was a UK size eight and that her body mass index (BMI) was within the healthy range for an adult woman.
Clearcast stated that the model weighed 134lbs and was 178cm tall with a BMI of 18.8, which sat well within the healthy weight and BMI range in accordance with NHS guidelines. They said that some viewers may subjectively view the model to be too
slender, whilst others would recognise her to be of a healthy appearance, which was supported by the NHS guidelines.
ASA Assessment: Complainsts upheld
The ASA considered that while the female model in the ads generally appeared to be in proportion, there were specific scenes, which because of her poses, drew attention to her slimness. For instance, the ads showed the model lying on a sun
lounger stretching her arms, which emphasised their slimness and length. Furthermore, towards the end of the ads were scenes showing the model spraying mist on herself, which placed focus on her chest where her rib cage was visible and appeared
We considered that the model appeared unhealthily underweight in those scenes and concluded that the ads were therefore irresponsible.
The ads must not be broadcast again in their current form. We told Nasty Gal Ltd to ensure that the content in their ads were prepared responsibly.
Online adverts placed by Scottish companies are to be trawled by automated bots to proactively seek out commercials which break censorship rules.
The automated technology is part of a new strategy to be unveiled next month by the Advertising Standards Authority, which will use the software to identify adverts and social media posts which could potentially be in breach of official
standards. They will then be assessed by humans and a decision made as to whether action should be taken.
ASA chief executive Guy Parker told Scotland on Sunday that Scottish companies and organisations were likely to be specifically targeted under the new, UK-wide strategy. Parker regurgitated the old trope that the innocent have nothing to fear
I don't think responsible Scottish companies have anything to fear -- on the contrary, they will welcome better online regulation.
We want to make more adverts responsible online than we have at the moment. We are looking at how we can responsibly automate something that would flag up things that we would then want humans to review. We want to be in a position by 2023 where
we are an organisation that is using this technology in a way that makes adverts more responsible.
It seems that Scotland was chosen as the Guinea-pig for the new system as ASA says that Scots historically don't complain much about adverts, although there was an upturn last year. Parker notes that the most complaints UK-wide come from
"better off, middle class people in London and the southeast of England".