Why CTV is Important for CPG

It’s estimated that investments in Connected TV (“CTV”) grew 40.6% year over year in 2020. Advertisers and marketers have their eyes on this as a hot possibility for connecting with their target audience and rightfully so as more people have moved away from broadcast television subscriptions. 

But the new nature of this vertical and the innate challenges with data capture and analysis will prove problematic for advertisers, according to Mobilewalla’s CEO Anindya Datta and their VP of Sales-CPG, Jim Mahoney. While the movement of eyeballs to CTV has driven budgets to follow that movement, it’s not an equal transition of accurate data. There are many more complex factors to think about when advertising through CTV. 

Since it’s not comparing apples to apples, marketers have to consider unique practices and influencing factors when evaluating ad spend and performance. While it’s likely that new methods of capture and measurement will emerge, but for now there are more questions than answers. 

During this episode, those challenges and possibilities are explored and discussed covering topics like scarcity, fragmentation, privacy, diversity, and measurement. 

Transcript

Mobilewalla - Data Point of View - Anindya Datta, Jim Mahoney - Transcript

[00:00:00] Jim Mahoney: Okay. 

[00:00:03] Laurie Hood: Thank you for listening. I'm Laurie Hood, CMO at Mobilewalla and this is Data Point of View. Today joining me are my colleagues Anindya Datta, CEO and founder of Mobilewalla, and Jim Mahoney, Sales leader for the Mobilewalla CPG practice. Today's podcast is why CTV is important for CPG. And COVID-19 accelerated a lot of different viewing behaviors. With consumer spending more time at home watching, streaming entertainment, and many of the major studios, either launching or acquiring streaming apps, like Disney+, Hulu, Peacock, Tubi, HBO Max. I mean, the list goes on and these acquisitions were made, or these introductions were made as a hedge against in-person theatrical viewing shutdowns. As many of us know, the theaters closed, Broadway [00:01:00] closed, et cetera. But what this did was this trend resulted in massive viewer growth, as well as, corresponding significant ad spend. So, we're going to jump right in and get started today. And Jim, why don't you start with some of the challenges that CPG companies are facing and why CTV should be super important to them? 

[00:01:24] Jim Mahoney: Thanks, Laurie. And it's great to be talking with you and Anindya. Thanks. It's great to chat about the CTV, and all the growth that's happening there. And when we talk about challenges, really you have to start with the growth and what the challenges are being caused there. And so, as Laurie talked about the pandemic, a lot of viewership, the proliferation of devices, et cetera. And I think we're showing some statistics from both the IAB and Nielsen on some of that growth. But, I think one of the big challenges is that because there's been this movement of not only the eyeballs, but, you know, [00:02:00] budgets follow that movement. And so, you've got almost three quarters of clients shifted money from broadcast in the CTV last year. And another third of them expect to spend more in CTV. And so the budgets are really going there, but the other challenge is a scarcity. And from that, we mean, this year the, it depends on who you believe, if it's GroupM or IAB. The market for CTV spend is going to be anywhere between 9 billion and $13 billion. And we just had a $4,5 billion upfront, in late spring. And so, depending on the numbers you're looking at, that's either a third to half of that inventory is already committed through a upfront spend. And that's a lot because, you know, the networks that have the apps, Peacock for NBC, Tubi for Fox, et cetera. So, and then Disney+ for ABC. So, so what that's done is created a, a [00:03:00] scarcity. The other thing we know is, I think this through the IAB about 30 to 40% of the inventory for CTV is being purchased programmatically. Well, that means that 60% of it is not, and it's direct. And a lot of that was through that upfront obviously, but what that creates is issues around frequency capping and really a fragmentation of these buys. So it's much harder. I mean, programmatic, spend can be measured, and you can do a much better job of understanding your targeting and your sales lift or attribution if you have the right data associated with it. But, in a lot of cases like with YouTube, YouTube has a huge audience that are probably the largest connected TV, with about 120 million a month, but they basically can tell you reach. And they can tell you a little bit whether it's on-target reach, but you really don't get a sense for, "Hey, did it work?" Like you can't compare it to anything. And so, that's a [00:04:00] big challenge, the scarcity and, and the, and the fragmentation. And Anindya, I think it would be interesting for people to understand better, you know, sort of the nature of CTV data itself because it compares differently to things like, you know, other digital type of data, like either cookies or maids or what have you. And, and I think that gets also to part of the problem that we can solve with the use of CTV data.

[00:04:31] Anindya Datta: Thanks Jim for asking the question and thanks Laurie for moderating this podcast. So that's, that's, that's, you know, wonderful introduction, Laurie and Jim, and in particular Jim with regard to the fragmentation. So your question was sort of the, some of the problems you identify. I mean, how does that data available from CTVs? How does it either sort of, sort of help address some of those challenges and how, [00:05:00] how, how is it, sort of different from, from let's say other digital data. Right. So, she knows, so, so, so I'll answer this in two ways. The f, the first easy answer is how is the data different? Well, the closes that CTV data, or the type of data that CTV data resembles the most closely is data that comes out of, in the mobile context, right. I mean, in, in, cookie-based targeting, you actually end up targeting a cookie and not the screen. But in mobile, because screens are attached to IDs, and in, in CTV screens are attached to IDs. So, so CTV data, data that is coming from CTV viewing is the most similar to data that comes from mobile usage. Right. So that's at a high level, but there a three and, you know, at a high level, the structure is [00:06:00] roughly the same. I mean, what's coming out is the idea of the device. You know, the, the connectivity of the device, for instance, what, what, you know, IP the device is connected to, and, and you know, what to each app viewing is happening type of devices so on. But there are sort of three I'll, I'll mark three sort of very important differences. The first is with respect to ID in mobile. By and large. I mean, what I'm about to say is not a 100% accurate because people, individual consumers can go and, and, and turn this off. But in mobile, in a vast majority of mobile data, you get to see a unique ID, and unique addressable ID for the device that the consumer is using. And it's known by various names. You know, the generic name is MAIDs, right, mobile advertiser ID. If [00:07:00] it's, if it's an Android device, it's an Android ID. If it's an Apple device, it's IDFA, or an Apple ID. But that ID is by and large a singular ID for the device. So from a device, if you, if you use 10 different apps, when data comes out of each of those 10 apps, they referred to the 

[00:07:25] device using the same ID. Right. So, and, and so that's, the first major difference is that in CTV this is no longer the case. Right. So CTVs also have addressable IDs known as CTV IDs, in fact, if you were to look at sort of CTV data versus mobile data, the IDs kind of the structure, the seam similar. But the one huge difference is that the CTV IDs themselves have heavy fragmentation. So if you are, if you are viewing content [00:08:00] on a specific TV unit from two different apps, chances are very likely the two apps refer to the TV using two different IDs. So that's, that's the first major difference ID fragmentation. again, just recapping. In mobile, the mobile device is usually referenced by the same addressable digital identifier by the various apps on the mobile device whereas in connected TV that's not the case. if you're using, you know, if, if you're watching content on Hulu, as opposed to Tubi, the TV will be referenced with different apps. I'll go into the implication of this later, but just the different, this is one major difference. The second, the other two major differences relate to the key characteristics that allow [00:09:00] these signals to be, to be, to be used meaningfully by advertisers. First is apps, right. In, in mobile, or even, even in, in, in the, in the desktop world, the application through which you consume content has been a very significant identifier, significant indicator of who you are. Right. I mean, if you are using a fitness app, I mean, you, you, you are presumed to have, if you're using is female-centric app, you're more likely to be female. And in, in sort of the desktop, and especially in the, in the mobile world, app proliferation is huge. Right. Right now, I mean, I couldn't even tell you, I would say that the Apple app store probably has over 3 million apps now. Right. And, and one, and again, I'll get into the [00:10:00] implications later. One of the significant differences in connected TV is number of apps are far, far, far less, and multiple orders of magnitude less. Right. We'll get into what this means, but this is a huge different between the two data types. And the third major difference is, is sort of goes the other way. Right. One of the ways that advertisers, marketers have tried to isolate devices, have tried to identify devices uniquely, especially when in the cases when the device identifier is not available, is by the type of device. So, so, you know, iPhone, Android, right. So the same is true in the CTV world. I mean, you have Samsung, and you have LG, and Vizio and all bunch of TVs, but the specific makes and models of TV [00:11:00] devices far outnumber the makes and models. Like for instance, you know, iPhone, if you, if you look across all iPhones, it's iPhone 7 to iPhone 12, they're probably across all the iPhone. They're probably a hundred different models. I mean, I'm, I'm making this up, but I don't think I'm too far off in, in scale. In CTV is Samsung specific TV model really is a thousand. Right. So, so, so it's, and it's kind of the reverse of apps. In apps there's significantly more fragmentation of apps in mobile, far fewer in CTV. In models, There are far more models in CTV and far fewer in, in, in the mobile world. Right. So these are again, recapping the three major differences. First, ID fragmentation. [00:12:00] CTV IDs are, are, are different, the same, same TV has different IDs. And second you know, I, I, I can see you're excited with my...

[00:12:12] Jim Mahoney: I was so excited about that, that my, my light fell off. I think so what we're going to do is we're going to pick it up again from the, your recapping Those three. And let me see if I can do something here so that doesn't happen again, but that is really funny. I'm sorry. 

[00:12:39] Laurie Hood: Sometimes you can just prompt it up. 

[00:12:41] Jim Mahoney: Yeah, I'll just, if it's down low like this, is, is that okay? 

[00:12:49] Laurie Hood: works that works. 

[00:12:50] Jim Mahoney: Is that better or

[00:12:51] Anindya Datta: Even better than before. 

[00:12:53] Jim Mahoney: Just has to be sufficient. It's not, I don't think it's going anywhere. Let me take a big swig of some, you know, something that looks like [00:13:00] coffee was probably really scotch.

[00:13:03] Anindya Datta: All right. So I'll go again, guys.

[00:13:06] Jim Mahoney: Yup.

[00:13:08] Anindya Datta: Recapping the three, three major differences. First is ID fragmentation. The same TV is usually referred to with different IDs by different apps, Whereas the same device, same mobile device for instance, that's not true, different apps up on it, we'll, we'll refer to using the same ID, that's number one. Number two, far more apps in mobile than on CTV. And number three, far more device models in CTV then in a mobile. Right. So, so having, so these are differences. So let me just spend a few minutes on what the implications of these differences. The, the ID fragmentation, really the implications are, are quite, quite heavy. Right. And you, you, you are talking about frequency [00:14:00] capping, Jim. So I'll bring that into this as well. So what, what this ID fragmentation makes, or the are, or what gets hard, what gets difficult due to ID fragmentation is to for advertisers and marketers to understand which unique TV is it that I'm talking about. Right. So for instance, if, if, you know, using the same sort of portrait building techniques that have long existed in the digital world, you say that, "Ok . I believe that, the, the, the, the, the person I'm seeking to, to reach has, you know, is likely a user of app A, app B, and app C." In the mobile world, you know, you, you simply pick domains that are associated with all three apps that breaks down in CTV. right? Because, because each of those three apps might be referring to the TV with different IDs. So [00:15:00] that's, and, and as a result, the result is not just, you know, difficulty of building portraits. Once you are, once you are using good IDs to send content, deliver content under those TVs, unknowingly to you, all your content could be going to the same TV actually. Right. Because you thought that, "Hey I'm". and there's no way. And frequency capping of course are things as you want to distribute your, your, your, your, your content, and those become very different. The, the issue that the other two cause are, are also similar, but they just make the problem harder. For instance, fragmentation of apps in mobile has actually caused the portraiture of who consumers to be easier. Right. I mean, in, in the desktop or in, for instance, you know, it turns out that there are [00:16:00] many more apps than actually websites. Right. So, so, so the, the truism is that in mobile, you can do much better portrait than in desktop. That problem really becomes much, much, much harder in the CTV world. I mean, because in CTV right now, you know, again, I don't know exactly how many apps there are, but the number of apps on a, on a, you know, that people want to use is like 10. Right? And so, so, so, and, and, and that, if you, if you take two mobile devices randomly and see, "Hey, let me see which common apps they use." You'll see a bunch of unique apps that each uses. If you take two TVs and say, "Let me look at what, what apps are unique in this and common." You'll find that most of the apps are basically in common. Right. So it's hard to distinguish. Basically app fragmentation makes it hard to distinguish one CTV user from another. So that's, and, [00:17:00] and, the, the, the fragmentation, the much higher fragmentation in model IDs, in a types of TVs is actually good for, for CTVs. I mean, given that, you know, that eventually, like in most cases, you will end up wanting to understand, let's say the household the TV is in, or, or, or the nature of the users of the TV. Given that it is easy, you know, even in a household, two TVs are very unlikely to be the same. I mean, they should be easily distinguishable, and that allows you to have much more granularity in let's say, Who viewed content. Right. So you might find that in a particular household There was one particular TV which was on, and certain people that were not in the household. You can distinguish which TV that was. You can, and you can make a much, much more granular different than who's watching what TV. So these [00:18:00] are roughly, I hope this makes sense, but these are some of the more, more, more sort of, big differences between data and some of the implications. You know there are clearly other implications. But these are some of the high level implications.

[00:18:13] Jim Mahoney: Well, and what, and what, and what that does is kind of reminds you of the types of data that people are able to get today, which is very different from what you're talking about. Right. So, so if I'm Peacock and NBC, I know 'cause I have a logged in users. I know what, and even if they're, you know, privacy compliant and anonymous households, what have you, I know what people are watching or what households are watching. And by that, the, you know, the, the programs, and there's a lot of depth of data there, but that's only usable when I'm selling and advertising into Peacock properties. And the same thing goes for Tubi, and what have you. The other thing is, is I might be using a video ad server. So it's, I'm [00:19:00] using a video ad server like Innovid, I know like where I placed those ads. I know what apps they went into. But what I don't see is in a household, or in a, let's say a cohort of households, which should be privacy compliant. I don't know what the app usage is. Like, you know, I, and so, my understanding of data that we have available is, you know, you can get into things for targeting, such as, you know, the types of apps that people are are using, because if it's Disney+ maybe it's more family-oriented target versus if it's Hulu, you know, maybe it's a slightly more adult-oriented, and, and certainly sports obviously for reaching certain individuals. But the other thing that's really helpful with this data would be just light, medium, and heavy CTV use. I mean, so people that are either cord cutters or that they're less likely to be reached in broadcast television. [00:20:00] And I know there's a lot of CPG clients that are very interested in reaching those folks. So, I don't know if you have some further perspective on that. 

[00:20:10] Laurie Hood: Jim, I'm going to jump in before we dig into that and kind of move into the use cases, and ask, let's talk a little bit first about privacy and consent, and what does it mean to get and use this data, you know, in the current climate. Because Anindya, you talked a lot about the aspects of the data, and Jim, you were starting to get into all the cool things you can do, but, but what is, what are the privacy and consent implications?

[00:20:42] Anindya Datta: Okay. Am I, am I? 

[00:20:44] Laurie Hood: Yes. 

[00:20:45] Anindya Datta: Okay, so this is something, was there a script for this?

[00:20:51] Laurie Hood: No, you were just supposed to answer it, but would you like me to answer it? 

[00:20:56] Anindya Datta: yes, because I can, but I won't, I will not be [00:21:00] very good.

[00:21:02] Laurie Hood: Okay. Then I will. We'll start out with, start here. Well, I'll restart here. 

[00:21:08] Anindya Datta: But I have a, I have a key point to make in rejoinder to Jim's point. So let's, let's get.

[00:21:13] Laurie Hood: Well, that's what, we're going to move into, into that. So, let me, let me just restart. Pause. 

[00:21:21] Anindya Datta: So you should like this about me. When I don't know something, I just say, "I don't know."

[00:21:27] Jim Mahoney: It happens so infrequently though. 

[00:21:30] Laurie Hood: Well, if you've worked in the office. So Jim, before, I want to interrupt you before we go a lot deeper into use cases. I think one of the things that's important to talk about are the privacy and consent implications. And where's this data coming from, and, and how has that privacy, how's that consent being obtained, and ensuring that you're partnering with a [00:22:00] vendor who is honoring the different opt-outs. And so, it's important as you start exploring. Getting access to CTV data that the vendors are adhering to all of the, you know, local, national, international regulatory standards. That they are overtly obtaining consent, that you can get kind of proof of that, and that you can also really, really deeply understand how that vendor's dealing with opt outs and any other kind of notifications. So I think as you're looking at this data, it's very important to understand how it fits in with your company's current privacy and consent strategies and what that vendor's doing to ensure that they have the rights to share that data. So, with that, let's transition into all of the really interesting and compelling use cases that, that these CPG [00:23:00] companies can, can work to implement as they sort of fill in and complete that picture of their, their, their buyers.

[00:23:11] Jim Mahoney: So Anindya, I think you had a comment or a question on one of the things that we covered earlier relative to the, you know, some of the targeting opportunities.

[00:23:21] Anindya Datta: That's right. So I was going to say Jim, that you, you, you sort of point, you outlined a couple of, sort of actually more than two, you outlined a few problems that let's say someone like Disney would have. Right. Under, under the CTV regime, trying to, you know, getting data out. I would like to stay a very, very, very basic problem. That, that should highlight the issue that CTV causes. So let's say you are Disney. right? And you have, so you have the Disney app in which you have a massive usage. And of course you have a number of CTV apps [00:24:00] as well trough which you have usage of. The, the, as, as, as sort of sort of people in this field know what, what the Disney. So it's very important for Disney to create portraits of their customers. Right. Create portraits of the users. In the app world Disney knows that here are the devices that are using my apps, and they can simply go out with that, and go to any number of data providers and tell them, "Tell me what", you know, and again, in a very, you know, privacy compliant way. "What, what, tell me what other apps these devices use?" And pulling that information in, and, and, and, you know, we could say, if it tells me the households these devices belong to it so on, they can create very nuanced portraits of the people that are using, using Disney [00:25:00] apps. This completely breaks down in the CTV world. That Disney cannot take, the simple operation that almost Every publisher does. Right. Or, you know, Disney cannot go to, take, "Hey, here are the CTV IDs to the Disney app. Can you tell me what else has done? Because those don't match anything else." Right. So this is a huge, this is a very significant problem that happens. And of course, this translates into, you know, Laurie was referring to, you know, and, and you are much more in touch with CPG companies than I am. So I'll let you elaborate. But. I would think that this, this, this problem then transfers over to large CPG companies as well. Who are of course interested in building in, in understanding holistically, who their likely prospect, their current customers, and prospective customers. What apps are they using? Which TVs are there on?. Same, [00:26:00] same problem sort of arises. So this is a very basic problem in the CTV space that, that, that, that occur, you know, arises in CTV, and is not there in other digital.

[00:26:12] Jim Mahoney: I mean things as simple as, you know, high occupancy, high occupancy households. I mean, the, like if there's a diversity of apps, CTV apps in a household where, you know, for me, for example, having, you know, two adult children that are home during the summer, And then, then they're gone in September. They're off to grad school or they're off to college, or what have you. the usage of the apps changes dramatically. And the, the shopping changes dramatically. And so, you know, not only the volume of things that you're buying, because you have a bigger, a larger group of people in the house, but also the stores that you're going to go to. Are you gonna really Go to the Walmart or Target, or is it just, is it going to be Kroger, Publix or grocery store? [00:27:00] Because you don't have to make the bigger trips, or as frequent trips and you don't and what have you. And so understanding that is critical for CPG clients. I mean, they, they love to be selling to the household that has, you know, teenage, teenagers, and you know, perhaps, a grandmother or grandfather living in the house with two parents. I mean, so 4, 5, 6 people in a household generates a lotta, a lot of laundry. Also, you know, if you're selling beverages, you know, you're selling juice, you're selling water, what have you. I mean, those are interesting to, to CPG companies because of just the, you know, they buy more of everything. And understanding the app usage is, is helpful. So, 'cause they, you know, there can be multiple shoppers in that, in that household.

[00:27:52] Anindya Datta: No. Look, no, completely, you know, to, to, to state that I live in state in a broader way. It's it's yes. So, [00:28:00] not only does it become harder to identify things like high-occupancy households, I'm going to postulate that this household occupancy itself is harder to determine in CTV. right? Because in the mobile world, for instance, right, there are many, many signals that come in, That help you design the great sort of occupants. For instance, one of the, you know, in both CTV, and in mobile, typically in the data that you get, you're getting location signals, right, where, where, where the users are allowing location signals to go out, you're getting location signals. But, and you know, you're getting as many locations signals in CTV as you're getting in mobile, which is why I did not bring it up in my differences. But one huge sort of semantic differences, in the mobile world those locations change much more than in the CTV world because people are not [00:29:00] carrying their TVs with them. Right. So, so what happens is that in a, in the mobile world, when, once you see a device and you see that it routinely goes from, you know, at night is, is one place in the morning is another place. If you see this pattern, you know, that belongs to a person and occupant. And, you know, he or she is going to work in the morning, and home at night. Such semantic signals become absent in CTV. Right. So just the CTV itself. I mean the CTV itself because of fewer apps. Excuse me, guys. Okay, I'm going to start over. So in the CTV world, [00:30:00] because of fewer apps, you have a harder time disambiguating people because of lack of variation in location data, that also creates a similar effect, right. I mean, you're not able to distinguish patterns. So overall, not only is it difficult to identify high-occupancy household Jim. I think in CTV, yes, identify occupants itself, which is a very significant sort of significantly important thing that marketers do in the context of other data that becomes much more difficult to do in CTV.

[00:30:39] Jim Mahoney: Well, and the other outlier you have there is, people that, I mean, there's a lot of, obviously a lot of viewing happening on mobile, not just people's phones, but also on their tablets. And their tablets are more mobile certainly than a smart TV would be. And so that's got to be Kind of accounted for in some of the [00:31:00] data evaluation and sort of the data aggregation that people do. There's two other quick things. One on, one more on targeting and then definitely want to just chat briefly about measurement. So on the targeting side, the other thing that if you know certain devices such as a gaming devices. That, that's, that's a critical one because if, you know, there's a lot of CTV that is accessible through X-Box and so forth, PlayStation. So, you've got, a lot of people that are watching content that way, but you know a lot about them because of that. And you are able to reach, I mean, you look at CPG clients that have whole brands centered around gaming. And, you know, sort of the, the fuel that people need for extended gaming. And I know that, they, they actually, a lot of brands are, are aligned with, the fuel you need for binge watching of, of shows and whole seasons of shows as well. So, but gaming is a really important one. And then I guess the other thing is, [00:32:00] on attribution. and I know from my Recent previous experience that, you have because of the fragmentation and a lot of panel-based transactional data, that attribution there's a lot of clients in the CPG world that just have given up on attribution and even other forms of measurement that will be maybe less, stringent, such as maybe sales lift or, or what have you. But. if you have the right data and some of the right transactional datasets, attribution, and at minimum, some kind of sales lift measurement is more possible because you have closer to, at least in my understanding, you know, closer to a census level of data available in, in the CTV dataset than you would otherwise when it was more fragmented among, you know, Charter and Comcast and, and, and others who, you know, and also that data will cost a fortune to do. So, [00:33:00] we interested in your thought on, you know, measurement how, how the measurement landscape changes because of, you know, the availability of some of this data?

[00:33:08] Anindya Datta: I, I agree with you. You know, so, so clearly because of, because of what you referred to as population scale, definitely higher scale observability under the same umbrella in CTV. You know, attribution for sales, for instance becomes easier. But one, so I just highlight, I mean, I want to repeat what you said, but I just highlight one specific attribution scenario that I think becomes harder in the case of CTV. So, as you know Jim, one of the, one of the most significant sort of, and there are a bunch of companies that have come up on this, one of the most significant attribution scenarios is visitation attribution. So, I show an ad for something, and, and that requires you to go, you know, I show natural museum, or I show an ad [00:34:00] for a DIY store, or I show an ad for circus that's coming to town, whatever. And, and, and, and success is measured, That was that, you know, did that actually drive the consumer to that, to that, that, whatever it was advertised. So that actually becomes harder, I believe in the CTV case. In the mobile case, you know, in the mobile sort of the whole, this, this type of attribution came up because people carry their mobile everywhere. Right. So if you saw, if the ad was seen on mobile and the mobile was seen at the circus, then it's assumed that you are the circus, of what that breaks down in the case of CTV. Right. So, so I think location visitation attribution becomes a very challenging element in CTV. And, and, you know, I know for instance, a couple of companies, and, and some of my, my entrepreneur friends who are trying to solve that problem by connecting sort of CTV data to other data. Right. But of course, [00:35:00] CTV data by itself doesn't allow that to happen. But if you can, if you can prove the coexistence of let's say a mobile device and a CTV, right, They belong to the same household. Then then, then through indirect means such attribution becomes possible. So I just wanted to point that out. 

[00:35:17] Laurie Hood: No y'all, this is great. Great conversation. Really nice overview of CTV data, some of the challenges, some really compelling use cases. But as we are running out of time, let's kind of wrap it up the quickfire challenge with sort of what are the three things, what are the takeaways? What are three things companies should be considering as they look at at obtaining and leveraging CPG data going forward? 

[00:35:47] Anindya Datta: Okay. So. so I would say the three, three key things that users of CTV data should be aware of is as follows. [00:36:00] The first one is kind of obvious, scale. The scale of CTV data, and by scale I mean the number of signals that are coming, and the number of unique devices that there exist is is much lower than from other, other digital sources. Right. 

[00:36:20] Laurie Hood: But do you think that'll, do you see that changing over time with, with the prevalence of CTV? 

[00:36:28] Anindya Datta: I don't see, I don't see the number of devices changing that much Laurie. At least in the United States because the CTV penetration is very high, but of course, in emerging markets, clearly that's going to change. And, and, and because large, you know, CPG companies, global CPG companies, there are many emerging markets that are major markets for them. 

[00:36:52] Laurie Hood: Yeah Makes sense. 

[00:36:53] Anindya Datta: But I do believe the number of signals will change. Right. The number of, you know, as, as, as [00:37:00] Jim said, you know, the advertising landscape in CTV is relatively new. So the advertising landscape itself, I believe will change. The volume of advertising will go up. Budgets will shift as, as more and more, and with that signals will go up. And with more signals, that ability to do things with the data will also go up. Right. 

[00:37:23] Laurie Hood: Yeah. That makes sense. 

[00:37:25] Anindya Datta: The two others, I believe, I'll give you two, you asked for three, I'll give you two others, which I believe are much more sort of profound if you will. One is the ability, regardless of the number of TVs or the number of signals, the ability to disambiguate users in the CTV world is much more challenging than in other digital landscapes. And unless that is addressed, the ability to target CTV marketing [00:38:00] is going to be constraint. Right. I mean, if you, if you can not know that, "Hey, this particular household where this TV exists has four individuals. And these are the nature of those individuals." Marketing is hard, right? And now, I mean, let alone nature of those individual just knowing there four individuals itself is a hard problem. So that's one, disambiguating users. And the final thing I'll say, is that, so because of the naturally scale constrained kind of nature of CTV data, and some of the endemic problems that are gonna, that's gonna make things like disambiguation hard. And, and, you know, I, at least I don't foresee how the need, the technology of CTV will change to make that happen. I think [00:39:00] the way to make that happen is by combining CTV data with other stuff. Right. So, so, so, so things like what household does the CTV belong to? What other digital channels are being used in the household? And can I, if I can do that, then I should slowly be able to combine CTV signals with mobile signals to create really powerful artifacts. Right. So the ability to connect TVs to other forms of content consumption also, I believe is, is, is a very important requirement in CTVs that needs to be addressed.

[00:39:36] Laurie Hood: Yeah, no, that totally makes sense. Jim, any, any final points to chime in with? 

[00:39:40] Jim Mahoney: Can I throw a CPG spin on that? So the VP of Digital for Nestle months ago talked about how, they went from a company that was really not collecting first-party data to one that really is quite a bit more now. And, and so they're, they're, they're exploring all kinds of new sources of data [00:40:00] because it's very fragmented and they want to understand their users better. And so, that connectivity to the ecosystem is really important because whoever's providing their supporting their identity graph needs to have the connectivity so that these things are actionable. You know, the second thing is, is that I think Forrester just did a study with experience around, the types of partners or the number of partners that people were using. And they said that there's, on average, eight different solutions companies are using today for connectivity and identity and what have you. And then also like six, eight solutions and six companies that are providing those solutions. And that was a survey of about 300 clients and 75% of them said that that was kind of their, their ecosystem. So, so, you, you need to educate yourself on, even, you know, companies of all different types that have this type of unique data that can help you. [00:41:00] And, and so that's kind of what we're trying to help people with in.

[00:41:03] Laurie Hood: Well, awesome. Well, gentlemen thank you so much for joining us today, and for the insight and the perspective in this really important area that, that's only going to grow and only going to increase. And the need to explore these types of data sources and the key use cases you can address with them is only going to become more apparent. For those of you listening, want to thank you for your time and please join us for another episode of Data Point of View. Thank you again.

[00:41:34] Anindya Datta: Thank you, Laurie. Thank you, Jim.

[00:41:36] Jim Mahoney: Thanks. 

[00:41:37] Laurie Hood: Thanks, guys.