Marketecture: Get Smart. Fast.
00:14:49 · Ari Paparo, Alex Cone
[00:14:49] S0: Right. Because that's matters ultimately. Right? You know. [00:14:52] S2: I... You know, that... That's been our theory of success is if if it's not driving marketer outcomes, like, it... We're probably not gonna see incentives change. So, yes, I would say that that is that is mission critical. In fact, publishers told us as we talked to them in the lead up to this and got feedback, like, we like this idea, but, like, we'll believe it when we see it in terms of, like, getting rewarded for better experiences. So anyway, yeah, Jounce publishes, published a report on Monday that effectively says all four... There's four metrics. We haven't named them all yet. Happy to. But all four metrics that are measuring different dimensions of ad load have correlation with their, what they call, conversion score metric, which I believe, as I understand it, is effectively, like, what... Like, there's, like, a baseline of what is expected as a conversion from a site and then, like, how much you diverge either above that expected or below. So that's what he overlaid and he saw all four metrics are correlated but the that the primary metric is this average ad count in viewport. Like you can kinda pin the other metrics to this one. [00:16:04] S0: Yeah. They're all correlated with each other basically.
00:16:04 · Ari Paparo, Alex Cone
[00:16:04] S0: Yeah. They're all correlated with each other basically. [00:16:06] S2: They're all correlated with each other and like that is actually fairly intuitive if you think about it. So what average ad count and viewport measures is as a user's... Real users moving around like in a like scrolling around in a in a site, what is actually in view in terms of ads, as a count and it's to, like, two decimal point precision. So somewhere around, like, 1.2, 1.3 ads in view, it it any get, like, on average, you do start to see, like, where the performance starts to drop on various, like, and others have shared very similar data with us privately. Chris is the first one to share publicly. And that makes sense because right now you've got another ad vying for your attention as another ad comes into view. So I think that... That's one of those where I'm like, I think marketers and agencies, like, are likely to... Like, not only just because of Chris's data, but, like, because it makes this intuitive sense, like, try to see if that drives performance for them. [00:17:11] S0: Yeah. Yeah. So the interesting... Well, another interesting thing here is that the graphs that were shown and we were... This... I don't have the graphs to show on screen, but we'll put them in the notes. It's not really a linear relationship between ad density and count. It's more like a... It's a curve where it's pretty flat for a while, and then the... When you really get to excessive ad load that it... The performance drops off radically.
00:17:11 · Ari Paparo, Alex Cone
[00:17:11] S0: Yeah. Yeah. So the interesting... Well, another interesting thing here is that the graphs that were shown and we were... This... I don't have the graphs to show on screen, but we'll put them in the notes. It's not really a linear relationship between ad density and count. It's more like a... It's a curve where it's pretty flat for a while, and then the... When you really get to excessive ad load that it... The performance drops off radically. [00:17:34] S2: Yeah. One of the things Chris pointed out in, which this also makes a lot of intuitive sense. Right? So like, you don't... Like, while... He was like, it's always better to have a lower ad count, which I think up to a certain point, and then there's no ads and then, like, it can't be doing anything. But, [00:17:51] S0: like Exactly. [00:17:51] S2: Yeah. Not always better to have lower density. And his explanation for that is the intuitive part, which is like, you could have, like, a little postage stamp ad that is different than a bigger ad. The way I've been thinking about it is two sites with 20% ad density on average. One of them has an average of one ad in view. The other has three, four, five. That is a different... Like, so comparing apples to apples on just density, I think, is part of why we decided, like, hey, there there's more dimensions you could look at here. But I think if you combine those, actually, like, could give you more insight both from an optimization standpoint, but also, you know, in targeting and Yeah. Reporting insights.
00:17:51 · Alex Cone, Ari Paparo
[00:17:51] S2: Yeah. Not always better to have lower density. And his explanation for that is the intuitive part, which is like, you could have, like, a little postage stamp ad that is different than a bigger ad. The way I've been thinking about it is two sites with 20% ad density on average. One of them has an average of one ad in view. The other has three, four, five. That is a different... Like, so comparing apples to apples on just density, I think, is part of why we decided, like, hey, there there's more dimensions you could look at here. But I think if you combine those, actually, like, could give you more insight both from an optimization standpoint, but also, you know, in targeting and Yeah. Reporting insights. [00:18:37] S0: So the obvious thing here is to, use this data to score URLs that you're bidding on, you know, negatively score higher ad loads. What about viewability? Like, I think, you know, folks pay a lot of money for predicted viewability scores across web pages. It feels like you're giving away for free, and they could avoid the tax.
00:18:37 · Ari Paparo, Alex Cone
[00:18:37] S0: So the obvious thing here is to, use this data to score URLs that you're bidding on, you know, negatively score higher ad loads. What about viewability? Like, I think, you know, folks pay a lot of money for predicted viewability scores across web pages. It feels like you're giving away for free, and they could avoid the tax. [00:19:02] S2: I wouldn't necessarily call these viewability because at at the end, this is page level data. It's not, placement level data. So what you're really getting is, like, what what on average... I guess, you could back into, like, well, if there's one ad on average in view... In the viewport, like, that's potentially helpful to a viewability vendor. The way that we've thought about these metrics is, obviously, you can grab them directly, but we think that they will probably take off more if they get baked into products. So I I I I don't think there's a a viewability solved directly in this. That said, the browsers had this thing called intersection observer for a long time that actually makes viewability pretty straightforward, but that's a whole another topic. [00:19:52] S0: Yeah. I've been hearing that the viewability vendors are gonna go out of business, like, five years because of exactly what said. The browsers can do it. So why are paying anybody? But I'll leave that to another episode also. So is Google using these signals for advertising?
00:19:52 · Ari Paparo, Alex Cone
[00:19:52] S0: Yeah. I've been hearing that the viewability vendors are gonna go out of business, like, five years because of exactly what said. The browsers can do it. So why are paying anybody? But I'll leave that to another episode also. So is Google using these signals for advertising? [00:20:06] S2: Google is exploring the use of these signals. Our friends at DV three sixty had provided a supportive quote like some other folks that aren't in Google for the initial launch. I'm, like, I'm trying to get those guys to use it just the same way that I'm I'm out here in London. Right? Like, trying to get more folks to use it. It it will be great when they do, and I'm glad that they're exploring it. It's like, they're kind of at the same stage as everybody else [00:20:32] S0: on Right. It's on the road map. How many meetings do Let... [00:20:36] S2: Let's get this on the road map. [00:20:37] S0: How many meetings... Internal meetings did it take to get that supportive quote? [00:20:43] S2: It it it helps it helps that a lot of the people that work on ads are people that, like... Because I've worked in ads for so long, like, at various companies, like, happen to also now work here. So it wasn't it wasn't so bad. I... The... But, yeah, you you worked here before. You know it takes... You have to bring everybody along. [00:21:02] S0: Yeah. You need you need to get a... Create a deck, and then the most important thing is you have to get to the meeting first so you could plug in your deck and someone else doesn't present their deck. It's like deck priority.