Last week, eMarketer unveiled its inaugural forecast on connected TV advertising in the U.S., projecting CTV ad spend to grow 37.6% this year to reach $6.94 billion and surpass $10 billion in 2020.
While the growth estimate is certainly strong, eMarketer analyst Eric Haggstrom pointed out that some critical factors are holding CTV spend back from growing even faster.
“Measurement is a huge problem that is holding back linear TV advertisers from advertising on CTV,” he wrote. “There is no single, commonly accepted measurement across platforms like there is in TV. Also, CTV targeting, attribution and programmatic capabilities are significantly behind those of other leading digital ad platforms.”
I totally agree with Haggstrom that linear TV advertisers are holding back budgets from CTV and that measurement is a big part of the issue. However, I’m not sure that much advertiser hesitancy is because connected TV ad providers lack the capabilities of other digital platforms. Rather, I believe what’s holding back CTV is its platforms’ lack of much of what TV advertisers demand of linear TV.
Thus, the question for CTV advertising isn’t how fast it can become like other digital channels. Instead, it is how fast it can integrate into the TV ad ecosystem. That means integrated, cross-platform measurement. That means cross-platform reach and frequency reporting. That means solving its fraud problem, which is already estimated to swallow more than 20% of CTV ads bought on programmatic platforms.
What will this require? Here are some of my thoughts:
Recommitment to measurement panels. We can’t solve cross-platform measurement and fraud without large, vetted, transparent audience panels. Nielsen’s current panel may be smaller than many of us would like, but we must have panel measurement. Full stop. “Walled garden” reporting across connected TV without panel measurement won’t move budgets..
Get our definitions straight. I like how eMarketer defines CTV advertising: “digital advertising that appears on connected TV (CTV) devices. Examples include display ads that appear on home screens and in-stream video ads that appear on CTVs from platforms like Hulu, Roku and YouTube; excludes network-sold inventory from traditional linear TV and addressable TV advertising.”
Full transparency. Already, much of what is delivered across CTV lacks transparency on channels and apps that are not fully disclosed. Yes, as we have seen in mobile advertising, connected TV has plenty of its own version of “flashlight apps” that seem to run a lot of ads, even if they are running in the background as pop-unders.
Nipping fraud in the bud. Pixalate reports that 22% of CTV ads bought programmatically are fraudulent.
We’ve seen this movie before. Please, let’s stop it now before this market gets any bigger.
What do you think? Is CTV ready to be integrated in the TV ad ecosystem?
I agree, Dave---especially about the need to measure audiences in a way that is compatible with how the vast bulk of TV ad "impressions" are measured--which is not to say that we have perfect panel measurements---we don't. But one must have common grounds in place for all players before improvements are considered. I would add---again---that TV advertisers and sellers need to make significant changes in the way they do business as well. The upfront, corporate buying system badly needs modification---my proposal for two upfronts--one for the lowball CPM tonnage buyers; the other for better targeted, brand by brand buyers---is one of the ways forward. And sellers need to think about giving up some control to allow this degree of flexibility to wend its way into the buying and selling system.
Having just worked on a connected Tv project, most of the major players sell their inventory direct, not RTB. 90% is sold that way. So if we have fraud its 22% of 10% of the inventory.
In other words a small amount. Quotes in articles like this are misleading.
Gary, point taken. However, given the CTV numbers reported by The Trade Desk, as well as the volume of YouTube in CTV, it's almost certain that programmatic is more than 10% of total CTV already. Even if it is that low, it would mean that CTV fraud is already well over a $100 million in the US on it's way to $200 million. If programmatic is more like 30% of the total market then the total fraud number is several multiples of that. I would hope that we're not so jaded in our industry that we think that it is okay to soend that much shareholder money that way. That is why I quoted it like I did. I don't think that it is okay.
Fraud isnt acceptable in any manner, what concerns me more is the lack of education and knowledge in this space to ensure the right questions are asked brand to seller.
Totally agree Gary. What's needed most to combat fraud is better education, particularly about what buyers need to ask and expect from sellers and platforms here.