Connected TV is set to become more popular this year amid the explosion of streaming services being launched to rival Netflix and Amazon Prime.
While the established players are ad-free, there are also ad-supported services such as ITV Hub, All 4 in the UK and US/international offerings such as Disney Plus, Apple TV Plus, HBO Max and NBC's Peacock. Even before these services were due to launch, Ofcom revealed last summer that nearly one in two UK households had a streaming subscription, while eMarketer estimates a slightly higher take-up in the US at 55%.
However, with greater consumer demand and more inventory on connected TV, there is the greater danger of ad fraud, warns Roy Rosenfeld, SVP product management at DoubleVerify.
While ad fraud on connected TV is relatively small in comparison to online display (web) and ads within mobile apps, it is growing very quickly from this lower base. DoubleVerify recently tracked a 120% year-on-year increase in fraudulent CTV and mobile apps, and flagged more than 200 fraudulent CTV apps in the first half of 2019 alone.
"CTV fraud is a big threat, and we expect this trend to continue into 2020. Already today we identify over half a million CTV devices a day as being infected or generating invalid traffic as it relates to ads," Rosenfeld says.
"It’s growing very rapidly and when we think about a media area that grows so quickly, you factor in future growth and the strategic value of it – this is clear to everyone with CTV to reach young folks in premium environment, full-screen, no multitasking – you can’t reach young people in linear TV any more."
DoubleVerify does not give a general number for how much fraud there is across connected TV inventory, because "it depends on how advertisers are buying it", such as on what devices and how they are trying to target audiences. However, in general, Rosenfeld says: "We’ve seen [connected TV ad fraud] everywhere across the board – some [platforms] are 1%, some are 50-60%."
The most common type of connected TV ad fraud targets the technique of server-side ad insertion –SSAI. This is where ads are stitched into the content stream at the server end, before it is beamed out to a user’s device, in order to deliver a seamless viewing experience. This is unlike online display ads that are traded programmatically through real-time bidding, in which buying auctions take place within nanoseconds to decide which ads are served on a web page (and hence why page loads can be slow on sites with a heavy load of display ads and tracking spyware).
Rosenfeld explains that SSAI fraud is so dangerous because there is nothing stopping fraudsters from writing software that runs on a server, runs requests for ads and claims ads have been completed.
"When someone wants to measure ad delivery, it could be DoubleVerify or an ad vendor, we want to know what type of device was it, was the video ad completed? There’s no actual communication with the device, it’s with the server… we have to trust that the server is ‘telling the truth’.
"As it turns out, that’s a design problem that fraudsters can abuse and there’s nothing to differentiate between legitimate SSAI and someone just making up inventory out of thin air."
The distinctive trait of online ad fraud is that it doesn’t appear to come from a particular region of the world, as, for example, cyber fraud has emanated from Eastern Europe, Rosenfeld says.
"There isn’t a ton of sophisticated malware that is driving the SSAI fraud. Malware exists, but to set up a fraudulent SSAI system you don’t need to be a genius cybersecurity hacker to come up with a system like that. You just need to be familiar with the ad ecosystem and generate fraudulent inventory."
However, media agencies that have spoken to Campaign have insisted that CTV ad fraud is not a major concern among brand advertisers, partly because the main players are established broadcasters and YouTube, owned by Google.
One source, who holds a senior trading role at a major media agency network, insists that brands should approach measurement vendors’ warnings about fraud with a pinch of salt because they have a vested interest.
Nevertheless, the same source acknowledges there is a risk with new players entering the CTV space and this risk could increase as the market grows.
"I do think there is a risk with some newer players circumventing the platforms to give fake views. The reality is today the market is quite small beyond the main players.
"There are not the products in the UK to combat this yet. Integral Ad Solutions is testing solutions in the US but that’s not even available to test in the UK for a while. There are also some integrations with key players that are owned by the media owner. This dilutes the independent, third-party measurement needs of agencies and advertisers."
For example, one challenge facing over-the-top platforms is the inability to tag content on CTV with specific tags that measure for fraud.
For now, the broadcasters have full control over who they sell their inventory to and the pipes that connect supply to their demand. This control, in theory, leads to less fraud, compared with the online display ad market, where inventory is available via multiple demand-side platforms and multiple resellers.
But part of the reason why Google and Facebook have sucked up so much of the online adspend that has driven the ad industry’s growth in the past decade has come from small and medium-sized business buying directly, often programmatically. Clearly there is an opportunity for connected TV platforms to offer programmatic and personalised advertising traded through exchanges rather than their own infrastructure, and with this comes the higher risk of fraud.
During a Campaign event about the opportunities of connected TV last week, ITV’s director of advertising data Lara Izlan pointed to the "second mover advantage" that the sector has, implying that the mistakes of online display are less likely to be repeated.
"It’s OK that we’re not there yet but let’s do it right and do it for the long haul. We have an amazing opportunity to shape those pipes and those frameworks for CTV," Izlan said. "We will have missed a trick if we don’t learn from programmatic and automated and try to build something that is more robust, data privacy compliant, more sustainable, including defining what metrics look like."
Dave Castell, general manager of inventory and partnerships EMEA at The Trade Desk, was even more bullish and warned against creating new walled gardens as programmatic advertising comes to television. "Let’s honour the idea of premium content and allowing people to buy fluidly through different platforms," Castell said. "The solutions are here now. Let’s experiment together. Collectively, we can test it and refine it to make sure the industry is ready."
And yet it remains to be how seen how willing major brands, in particular, will be to use connected TV as a field to experiment, particularly in light of recent brand safety scandals on video-sharing behemoth YouTube, where ads are regularly traded programmatically and without human oversight.
Paul Wright, managing director for UK, France, Middle East, and Turkey at mobile attribution platform AppsFlyer, warns that $4.6bn of mobile media spend was exposed to ad fraud last year, with entertainment services accounting for nearly 5%.
Wright tells Campaign: "Consumer media consumption behaviours are converging across smartphones, connected TVs, laptops, tablets and other IoT devices, presenting often highly sophisticated fraudsters with opportunities to make money.
"It's no surprise that there is a growing trend in connected TV in particular, as this is a less mature ecosystem when compared with other connected devices that already have robust anti-fraud solutions."