By now it’s a cliché to say that fraud follows the money – but it’s true.
And with consumer packaged goods (CPG) companies slated to represent $4.2 billion in digital ad spend this year – a number eMarketer predicts will hit more than $7 billion by 2018 – bots, and their human creators, are starting to add CPG brands to the top of their shopping list.
Ari Jacoby, CEO of digital ad company Solve Media, noted Q2 2014 had significantly higher amounts of bot activity among CPG brand advertising largely due to US mobile-based bot traffic, which was 40% higher than what Solve has seen in the past. Global mobile bot traffic was up a little more than 20% – a phenomenon corroborated by Andy Fan, founder and CEO of Shanghai-based fraud detection solution RTB Asia.
“CPG vertical campaigns have a higher percentage – possibly the highest – of bot-related fraud over travel, gaming, auto or financial services,” Fan told AdExchanger. “In China, the goal of most CPG vertical campaigns is branding, not direct response or ecommerce, and this usually leads to loose goal measurement, which makes CPG advertisers less likely to notice fraud.”
Spend on search and desktop display is up among CPGs – there’s been a roughly 11% year-over-year increase, according to Kantar Media – but it’s mobile that’s going to be the future bugbear, especially as CPG brands start to spend more on mobile programmatic.
Low CPMs are one issue – but high CPM inventory is no picnic, either. As CPGs start to turn their attention to digital video, bad actors will no doubt start to turn their attention to CPGs. As AOL’s former SVP of global video ad sales, Charles Gabriel, commented back in May, “CPG will be the most aggressive in putting budget dollars through programmatic in video.”
But at the moment, video is a fraudster’s playground.
“Video is the perfect storm of ad fraud,” said Forensiq CEO David Sendroff. “It has high CPMs and there’s a lot of demand, but not so much supply. It’s not surprising to see CPG players targeted.”
So-called “contextual” ad fraud is also an issue for CPGs, said Sendroff, who observed that “relevant content can easily be spoofed or replicated.” In a strange twist, the quest for brand safety and the practice of targeting ads against surrounding content could be creating certain vulnerabilities.
By creating an exact duplicate of a legitimate site, a bad actor can encourage contextually targeted ad placements. From there, it’s a classic PPC fraud situation.
AdExchanger Daily
Get our editors’ roundup delivered to your inbox every weekday.
Daily Roundup
“Criminals understand that there is demand for contextual advertising from the buy side and are sophisticated enough to provide the appearance of proper context in the absence of a human audience,” Jacoby said. “Contextually targeted ads are generally more expensive for marketers and therefore very attractive to fraudsters.”
Bad actors also take advantage of audience targeting mechanisms to do their dirty work, said Jalal Nasir, CEO and co-founder of ad analytics shop Pixalate, which added fraud detection and protection to its platform back in March.
“Females, moms specifically, are a key audience type that vertical tries to go after, [and] most agencies and networks running CPG campaigns end up targeting more blog sites, which account for a log of bogus/fraudulent websites,” Nasir said.
But in the end, it all comes down to the almighty dollar.
“CPG is one of the biggest spenders in the ad industry,” said Nasir. “P&G, for example, is one of the biggest advertisers in the world.”
(And even the biggest advertisers aren’t immune. In a recent experiment conducted by Forensiq, the company purposely infected a virtual computer with a particularly unpleasant bot to show what happens after a machine in compromised by malware. By the time 24 hours had elapsed, the lone bot had visited thousands of websites and generated about 10,000 false impressions for some blue-chip brands, including Aleve, Tide and P&G. And that’s just one bot, let alone a botnet comprised of multiple nasty little zombie robots. Watch a video about the Forensiq experiment here.)
