Until recently, the concept of brand safety hardly registered in Japan, with advertisers equating an uncommonly orderly society with an orderly online ecosystem. Attitudes have changed rapidly in a short space of time, helped by exposés on fraud and placement issues by media outlets such as NHK and Toyokeizai, as well as more vocal public statements by multinationals such as P&G condemning the state of online advertising globally.
Takahiro Doi, chief executive of SPI Interactive a consultancy that advises on best practice in online advertising, says there is still “a big difference in the degree of sense of crisis among advertisers” depending on their level of knowledge and whether they are following top-down initiatives, which are most evident among global brands. For those that are concerned, he says safe placement of ads, rather than deliberate fraudulent activity, is the focus. Among the most recent services to capitalise on this demand is Cheq, which launched late last year and claims to anticipate and prevent unsafe ad placements.
Making the case for its service, Cheq last year commissioned a global study with IPG in partnership with BMW and Hulu to determine the effect of a bad placement on brand image. In some cases, respondents saw the brand as exploiting shock value, while perceptions of quality, reputation, trustworthiness and care for customers declined between 18% and 22%.
The Israeli company’s entry to Japan, which has its share of potentially damaging content via radical personal political sites among other platforms, is indicative of an appetite for preventive ad verification services. Founded by Guy Tytunovich, a former Israeli defence intelligence official, Cheq operates in partnership with Dentsu-owned Cyber Communications Inc (CCI). It competes with the likes of Integral Ad Science (IAS), which offers features such as firewalls and pre-bidding.
Like White Ops in the US, these services differ from those such as MOAT and Momentum, which focus on analysis and reporting but don’t necessarily help prevent advertisers from appearing next to a piece of hate speech or sexually explicit content.
The most common means of blocking bad placements is to rely on past data via a DSP, but the gold standard is real-time adjustment, says Shigehiro Ando, who works with Cheq at CCI. As well as Cheq, players that also claim to block risky placements in real-time include Doubleverify, which is looking to expand in the region.
Tytunovich criticises pre-bid verification, saying it is limited in scope and fails to address the possibility that the data used to make decisions could itself be fraudulent. He says Cheq looks at “over 700 different parameters” to ensure accuracy, which includes identifying content for what it really is. Blanket prevention measures risk blocking placements that are in fact perfectly safe.
“Accuracy is the make or break of the game,” he says. “A lot of what people are blocking should not be blocked. This is damaging because everyone loses scale. Publishers lose users and the advertiser loses reach. So being probabilistic rather than deterministic is a very bad thing.”
Campaign asked IAS for a response to Tytunovich’s assertion. Manasa Denning, senior director of solutions engineering for Asia-Pacific and North America, says the company has “a team of data scientists all working to ensure our customers understand the true value of their impressions and assessing hundreds of metrics, from viewability and fraud to geo-compliance and ad clutter.”
Denning says IAS ensures accuracy and viewability “by using next generation, cutting-edge media analytics and measurement solutions”.
Tytunovich claims to use NLP to decipher content “in a similar way to how the human brain works” as opposed to following a strict set of rules that could flag a statement such as ‘Le Bron James killed it last night’ as unsafe. He says another priority is to help publishers continue to operate by ensuring they maximise the monetisation potential of content that is suitable for advertisers, while steering advertisers away from content that, while legitimate, carries negative associations.
“We need to understand good content is fueled by advertising dollars and the content needs to remain empowered,” he says. “Not every page is worthy of monetisation. If it’s a story about a terror attack, the publisher has to carry it, but they need to be remunerated elsewhere.”
He gives Goo as an example of a publisher that Cheq helped “get back into media plans” after an advertiser blacklisted it by showing that around 80% of its content is safe to appear against.
Cheq is still in its infancy in Japan, but has so far worked with major brands in the automotive, construction and FMCG sectors. Despite rising interest in these services, there are challenges to growth. One that Doi highlights a reluctance to pay. “Most of the time, the cost is based on CPM, so many advertisers hate the cost increase and hesitate,” he says.
Japanese clients typically take high quality inventory for granted, agrees Ando. But Doi urges them to consider that the KPIs they rely on without the help of a verification service likely contain “a mix of illegitimate impressions, clicks or ads that damage the brand”.
Another issue is that each tool vendor has its own standards in terms of judging safety and fraud. “Therefore, unless the tool is 100% reliable, you cannot rely on the tool alone,” Doi says. “It is a mistake to think that all will be solved if a tool is introduced.”
Doi thinks all stakeholders—advertisers, agencies, platforms and publishers, must assume responsibility for brand safety. For advertisers, he says the approach should vary depending on what sort of company they are.
“For advertisers whose brand image is the key consideration, whitelisting ads would be the best bet,” he says. “For these brands, it is much safer to preselect websites rather than advertising on an unspecified number due to low price.”
More important than selecting a verification tool is selecting the right platform for programmatic buying, he says. “All purchases and delivery of ad inventory will depend on the capabilities of the platform.”
He says advertisers should be absolutely clear about their criteria for placements. Even if it’s difficult to set a risk tolerance level, they should be able to indicate content to avoid. An ad verification tool should then be employed “to make up for the parts that cannot be dealt with manually”.
Ultimately, advanced detection and prevention of problems online “is impossible without these tools”, he notes. But however sophisticated, they are to be used precisely as tools, not as a substitute for first-hand understanding.