A recent survey of media buyers conducted by semantic advertising vendor Peer39 revealed – as you might expect – an intense interest among that audience in page quality and quality controls on their online campaigns. Only five percent of respondents said page quality doesn’t matter, and only eight percent said they don’t currently use quality controls. For 87 percent of them, about 50 percent or more of campaigns require quality controls.
The top quality attributes for campaigns, they say, are content-rich environments (52 percent), home pages (51 percent), and user-generated content (55 percent).
UGC is a tricky problem in the online advertising space, because it adds more risk – site owners do what they can to ensure that comments don’t transgress boundaries but moderation only goes so far, or is otherwise subject to time-, resource- or cost-constraints. Not to mention that user comments that some advertisers may find inappropriate aren’t necessarily something that would be flagged as problematic by human moderators or automated systems.
“UGC is valuable content but it contains pieces that can be problematic,” says Dr. Jonathan Schler, Peer39′s CTO and General Manager of its Israel research and development facility. “When you look at comments it depends on the threshold of how sensitive the advertiser is.” For instance, a user comment on an article about summer eating that says it’s a bad idea to make soda a part of barbecue plans because it makes kids fat isn’t likely to raise an alarm for anyone but a fizzy beverage company, and even some of them might be okay with it. Semantic technology that can help score how strong or relevant such comments are to the page and match those results to the advertiser’s particular comfort level can help avoid problematic placements. “Tying this together with knowing the main topics and categories [the page is about] gives you insight into what types of UGC you can or cannot consume,” he says.
Overall, quality attributes are subjective to each campaign, he notes, “but it’s important information to provide to customers. When you are working in a real-time advertising environment you need to make decisions in milliseconds, and to make the right decisions you have to have that information.” Semantic advertising technology that can go very deep to analyze a page to narrow down its focus to three or four main topics makes it easier to target ads against the page. “With lots of keywords or renormalized information, those systems don’t work efficiently,” Schler says. He says that internal tests there have shown that there is a 2x-times average lift in the performance of campaigns that used semantic conversions.
Peer39 says that by month’s end it expects over 7 billion impressions a day with close to 100,000 requests per second with a minimal latency of 5 ms or lower. The company in May also totaled up $27 million in VC investment with its May raising of $5.2 million from Canaan Partners and other firms.
Other semantic advertising companies also have been enjoying investor interest. Admantx, for instance, received $2.8 million in first round funding this month. And Taykey this month formally launched as, a full-service advertising platform for social media that uses semantic algorithms to monitor trends and help advertisers target to users based on what their online behavior is likely to be in the near future. It announced $9 million funding round led by Sequoia Capital at the same time. Kontera, whose In-Text Semantic Analysis Technology identifies story level topics and then selects the most appropriate keywords and keyword phrases to highlight within the page, also made a June announcement that it is partnering with AppNexus (also a partner of Peer39’s) so that its technology can be accessed by the real-time ad platform’s global ad network, DSP and other online ad company customers.