Beyond Attribution: Joanna Drews on Measurement Truth and the Future of Advertising Effectiveness
- Aug 6
- 48 min read
Advertising has never had more data—or less certainty.
Marketers now have billions of signals, AI-powered dashboards, clean rooms, attribution models, MMM, and platform reporting. Yet one question remains harder than ever to answer: What actually worked?
In this episode of Signal & Noise, Brett House and Rio Longacre sit down with Joanna Drews, Co-Founder & CEO of HyphaMetrics, to explore why the future of advertising measurement may depend less on collecting more data and more on measuring it better.
Joanna explains why deterministic attribution is breaking down, how AI, representative panels, and person-level measurement can provide more trustworthy insights, and why the industry is shifting toward hybrid measurement models across TV, streaming, gaming, mobile, and digital media.
They also discuss:
Why traditional attribution is falling short
What MMM still misses
AI's role in media planning and measurement
The challenges of cross-platform measurement
Why trust may become advertising's most valuable metric
Whether you're a marketer, agency leader, publisher, or data scientist, this conversation offers a practical look at how measurement must evolve for the AI era.
Because better AI doesn't start with better algorithms. It starts with better measurement.
Watch the full episode and join the conversation.
🔑 What We Cover💡 Key Takeaways🎯 Why This Episode Matters
Read the full transcript below.
Brett (00:01.31)
Hey everybody, welcome back to Signal of Noise. I'm Brett House, joined by my co-host Rio Longacre. And we are thrilled to have Joanna Drews, a guest that we had on in Miami at Possible, even in Miami. And we've been trying to get you on a full time, a full episode. One time it kind of, I think we ended up having to reschedule.
Rio (00:06.923)
there.
Rio (00:21.742)
Well, we tried a can, but this worked out better because now we've got a full hour with you. Maybe more.
Brett (00:25.798)
Now we've got a Friday afternoon to go like long form with you, Joanna. So welcome to the show. Thrilled to have you calling in from Seattle, a city that I've never been to, but, but, and apparently I culturally, I won't, I won't adapt to very well according to Rio, but the food is good and the music is good. But yeah. And so for those
Rio (00:40.504)
Food is good.
People are not very friendly, but food is great. Yeah, music scene's great. Weather's great, three months a year it's great, the rest of year it's not.
Joanna Drews (00:45.778)
you
Brett (00:50.046)
Yep. Totally, totally. We'll go see the Soundgarden or Soundgarden got its name. I'd go see some some post grunge music. yes, Joanna is the co-founder and CEO of Hyphometrics. You know, we obviously have gotten to know each other a little bit. You're deeply knowledgeable around the measurement and analytics space. You work companies like iCrossing, Groupim, Comscore.
Joanna Drews (00:50.276)
Exactly. We'll help you out, Brett. Come visit when the sun's still out. We'll share. Yeah.
Brett (01:18.166)
before founding and this has been and co-founding and now the CEO of Hyphometrics and this has been a path for you. You've been doing this for what about six seven years at this point, right?
Joanna Drews (01:27.972)
Yes, Hypha Metrics was formed seven years ago. But to your point about it being a path, all those roles I had prior to Hypha very much led me to this point in time. I'd say the journey exceeds seven years. I wouldn't say, I won't tell you the exact number though.
Brett (01:44.939)
Yeah, you want to encapsulate your life in the last seven years, right? There's a lot of life experience, But you're originally from the East Coast, right? Is Philly, are you from Philly?
Joanna Drews (01:47.792)
Yeah, I'm sorry.
Joanna Drews (01:52.976)
Yeah. No, I'm actually from the DC area, from the Mid-South. And I went to town, feeling.
Brett (01:56.526)
the DC area, but you went to Temple, right? We always end up going that far back, regardless of how old we are.
Rio (02:03.893)
And then you moved out Seattle for this opportunity for this company or had you been out there previously?
Joanna Drews (02:04.539)
Good night.
Joanna Drews (02:10.073)
So I was on the West Coast. started the company in New York. We started right before COVID broke out actually.
score and WPP and otherwise those were all in New York and then I moved out west did a COVID move and
You know, we're very much a operational company. So forward-facing, we're a data company, and our product is data, and we're incredibly excited about the unmeasurables. But to create those on the back end, we're deep AI. We built our model seven years ago when we started the company, so we have a proprietary infrastructure to ourselves. We have hardware. We have different types of meters at this point, several dozens of them.
When you're at the West Coast, Seattle is a hub, right? You have right behind me, you see the water. So it's an important part of our global supply chain process. And it's very strategic. For the hardware, we also have the AI. We're agnostic to which cloud service that we work with. And there are great tech companies here in Seattle, and we're forming great relationships with them. So when...
Brett (03:04.474)
Yeah.
Brett (03:10.525)
for the hardware, which I, yep. Yep.
Joanna Drews (03:27.377)
When you're looking at the West Coast and you have a company like HIFAA, Seattle is a strategic position.
Rio (03:32.385)
Well, Microsoft is located just up the street a few miles and AWS is just up the hill or Amazon is just up the hill. So very convenient.
Brett (03:32.519)
Yeah.
Joanna Drews (03:38.07)
Exactly. And I visited both this past week, so there you go. It's great to be here.
Rio (03:43.254)
know, Redmond's not far. You know, and the Amazon, it's cool, like what they've done to, they've really transformed a lot of Seattle. I mean, that whole part. know some of the locals will complain about how it's driven up rents and stuff, but it's, that whole part of town is really amazing. For anyone who hasn't been there, it is worth checking out all the new buildings.
Brett (03:43.783)
Well.
Joanna Drews (03:57.381)
Yeah.
Brett (03:57.534)
Yeah, and you guys have defined yourself as sort of a media measurement technology company. seeing the data as a service, AI powered data as a service is sort of a current sort of positioning play, right? And the whole hardware, I mean, I think a la Nielsen, and we'll probably talk about that in some level of detail, because there's some interesting juicy tidbits of information around, yeah, I was part of the Nielsen mafia, but not on the measurement side. came from.
Rio (04:17.805)
Yeah, Brett's part of the Nielsen Mafia if he hasn't told you yet.
Brett (04:23.997)
I came from the MarTech data side through the X-Lite acquisition, but I certainly saw that world. But you guys have, I think it's what you call the puck, right? Which is the hardware that gets installed in the house through a panel sort of acquisition process. And then it actually powers the mesh network, right? Of the house.
Joanna Drews (04:43.452)
So we create a mesh network in a couple of different ways actually.
Brett (04:45.285)
You create a mesh network. Yeah. And then that allows each piece of content being consumed regardless of device to be seen and tracked without the use of ACR technology, as far as I understand it. And because you've acquired these panelists, Kind of think of like a Nielsen home, like the Simpsons Nielsen home. You know who they are and what devices they use. So you can actually match back.
whether they're gaming or user generated content, social videos, CTV, linear, you can track it back to the individual, right?
Joanna Drews (05:20.666)
Yeah, so I would just add a couple of things to that. is that we are not the only industry that uses panels. Panels, it's a scientific tool. It's a group of people that's representative of a larger group of people. Civil engineers use it in order to make their decisioning. The medical community uses it. We're all familiar.
Brett (05:23.613)
That's my best approximation.
Brett (05:32.508)
Yes.
Joanna Drews (05:46.483)
clinical studies. There are different panels of all sorts, political polling. That could be a panel, right?
Rio (05:50.061)
Would a focus group be a form of panel?
Brett (05:51.335)
Yeah.
Joanna Drews (05:54.675)
But you know, there are panels that service almost every single industry across the top Fortune 500 companies here in the United States and otherwise, right? Even the federal government relies heavily on panels in order to decide what their departments need to do and focus on. So when we build our panel and we are the only company with a chief panel officer in our space, Chuck Shuttles, he's absolutely fantastic. When we think about panels, we're not looking only at our
Brett (06:11.964)
Yeah.
Joanna Drews (06:24.688)
space. Chuck is looking at every single panel built globally and what is best in class. And that's what we've brought over within to our
Brett (06:33.607)
to model your own panel, right? And I think maybe your underlying point here is that, because I've never met a data analytics or data science professional, and I've known a lot through my career, that hasn't actually talked how critical is a source of truth, the sort of panels are from an accuracy perspective, considering all of the bad data and bad signal that we have, it does allow for correction and validation and things like that, right?
Joanna Drews (06:57.436)
Right, panels should always be used with larger data sources because a panel is directional for those larger data sources. It's the ground truth and the quality of the ground truth is based on the methodology and the technology of that panel, right? So in our circumstance, our panel and our methodology,
Brett (07:14.47)
Yeah.
Joanna Drews (07:20.924)
We look to other industries and we have a completely touch free and passive experience for our panelists. When we send the hardware to the homes, it's completely self-install. We don't visit those homes, we don't rip apart your TVs. We very much think about the consumer experience, the product experience, and try to make it as seamless as possible. And also just think about that from an accuracy standpoint, right? If you as a panelist have to
Get up and touch a button to validate your presence or
Brett (07:54.299)
Which is how Nielsen operated, right? You had a special TV controller with buttons for each person in, you know, I this goes back 30, 40, 50 years, right? With a button for each person in the family so that they knew who was technically, at least the individual controlling the remote was in the room.
Joanna Drews (08:02.32)
Yeah.
Joanna Drews (08:08.484)
Yeah, but to be fair to them, I won't single just them out, right? They have global competitors. And when you look at the tech stack, I'm not seeing much differentiation across all of them. of course, we're US based and we're looking at this here. But outside of the US, HIFAA does licenses, hardware and software to other markets. And we're active on that front. And the pains that we feel in the US marketplace as far as lack of measurement, that's felt globally as well.
So this is pretty much a global problem, and it's very much a technical problem, and I would say a panel modernization problem.
Brett (08:38.671)
Yeah.
Brett (08:43.473)
Yeah.
Joanna Drews (08:49.872)
So we're fortunate enough to start from scratch. We're fortunate enough. And I've built an audio ACR panel in the past for a company called AxWave. And the rest of the executive team has either built panels themselves using audio ACR or been customers of that data or otherwise. So we're intimately familiar with the ins and outs and back ends and what it takes. And again, we're fortunate enough to start from scratch. We have a proprietary technology that is patented.
Brett (08:58.374)
Mm-hmm.
Joanna Drews (09:19.806)
uniquely ours and we're bringing forward that consumer experience and that level of expectation into the panel experience. And then we're creating a panel science that is aligned with other industries in order to make that quality threshold as high as possible. Because we...
Rio (09:39.214)
So combining the concept of a panel, which is definitely something that other industries use. I've worked in healthcare a lot, especially in pharma, as you mentioned, not only for R &D, but also for when you're going to commercialize drugs, very important part of it. So certainly the approach of combining the concept of a panel with some of the bigger data initiatives, can certainly see that's the direction things are going. Part of the reason why we wanted to have you on this was not only hear about the company, but just talk about, like you look at the direction things are going.
Joanna Drews (09:45.392)
Yeah. Yeah.
Rio (10:08.109)
There more signals than ever. Signals, as you know, and I think we're going to discuss today, not always connected. We've talked a lot about on this pod about how they're not always reliable, right? And then because of that, attribution is under a lot of pressure. At the same time, MMM is maybe having its moment, right, Brett? We talked a ton about this on this. Yet the complaints are when you look at an attribution, the platforms, they graded their own homework, and as media gets more fragmented between...
Brett (10:18.767)
Yeah.
Rio (10:35.999)
some of these new channels, TTV being one, but you can also throw retail media and others in there as well. It's really, really tough in order to have that single source of truth and know and really know what's happening and what type of media is effective. And you throw AI into the mix, right? It's starting to make planning, optimization, some of the even creative decisioning more automated, in theory, more database, which would be a good thing, closing that loop between performance and actually decisioning.
Would really be a good thing. So maybe we start with a question about like looking at advertising measurement Why is it so hard like in and how is it maybe improving?
Joanna Drews (11:13.586)
Yeah, well I think it's not so hard, which is why I started IFA. I think it shouldn't be so hard. That's a better way putting it. I think we as a industry made it really.
Brett (11:19.459)
Hahaha
Rio (11:20.587)
Or shouldn't be so hard maybe, right?
Joanna Drews (11:30.706)
because of my career that we're using.
Joanna Drews (11:36.711)
is not to be clear we build everything from scratch and we will never build anything that already exists we believe that to license any technology or build a similar approach is just adding more
the situation. So it's very important for us always to bring something unique to the table. But I think that's what's made it hard that there is we've been thinking as an industry very singularly in regards to how things should be measured. And what I mean by that is go back 100 years ago.
Brett (11:55.216)
Yeah.
Joanna Drews (12:09.446)
when diaries started. When you look at the diary methodology, it's actually not different from watermarking or ACR. Someone wrote something down, and then they went backwards in time to check if it was accurate from an ad schedule and a content schedule standpoint to figure out where it was aired. That's exactly what watermarking and ACR does. So how does that transcend to streaming, AVOD, video gaming, mobile?
Brett (12:11.301)
Yep.
Brett (12:29.274)
Yeah.
Yeah.
Joanna Drews (12:38.222)
It does not.
Brett (12:38.332)
And for those that don't know what diaries are, even when I was at Nielsen a few years back, they were still using them for local TV ratings. And they're literally the local TV stations filling out more or less spreadsheets and sending them to kind of a central hub to be data crunched and analyzed. And there's obviously a huge delay in terms of how quickly this data is turned around and reported on.
And that's how they actually drove. They eventually phased that out because it was a super manual process, but they phased that out in like the mid 2000, 2017, 2018, like very late in the game. Right. So, yeah, that's, that's an interesting point. I think it's, go ahead.
Joanna Drews (13:21.966)
But regardless if it's you're using that old method or you're using the automated methods of ACR and watermarking, all of those methods are confined to scheduled data. So anything you watch that is unscheduled, that tech does not transcend. And that's exactly why I started HIFA, because our tech does not follow that daisy chain of events that we've been stuck on for 100 years. But instead, we've leveraged AI.
Brett (13:45.785)
Yep.
Joanna Drews (13:51.251)
and machine learning and created this unique approach that measures the experience exactly the way the consumer does it in real time.
Brett (14:00.156)
Yeah, and that could be on multiple devices. And just so people that don't know sort of automatic content recognition, which is ACR, it's used in Shazam. If you've ever Shazam'd a song, right? And what it's doing is it's sampling. It's taking a kind of a small, either a visual snapshot from the television glass, right? It could be as little as like every, it could be as frequent, I should say, as every 500 milliseconds or so, or a brief audio clip, right? And then it sort of downloads that data.
Compresses it and then pushes it back to the manufacturer's servers, right? Yeah, I've been setting up on the stuff. I knew a little bit about it, but I did do some further right? Yeah, and and so it
Rio (14:34.699)
Have you been studying up on this stuff, Brett?
I know you've worked in a space, I'm just giving you our time. So we did have Yevgeny Popov from Samba there in ACR company on the pod a couple of months ago.
Brett (14:43.876)
Yeah, so you send it back, you send it back to the servers and then it's matched against basically a catalog of content, right? And tell me if I'm wrong, Joanna, because this was, I mean, I knew the basics of it, but I wanted to see kind of fundamental. And then, know, so you're going through this whole match process where you're saying sample of content is recognized audio or visual. It's compared to it to basically, you know, metadata for the entire content catalog. And then eventually it's reported out that this much time was spent.
on this particular show at this particular time of day. Is that?
Joanna Drews (15:14.502)
Yeah, I'll add one thing. One is that the baseline of information that you're starting with, basically, let's use Shazam as your example. You have a library of music, but a new song is playing and you're taking your library of music and you're smashing it against that new song trying to find a match. If it's not in your library, it's not gonna match, so you're never gonna measure it.
Brett (15:33.105)
Yeah.
Brett (15:37.009)
Yeah.
Joanna Drews (15:37.809)
Right? So until it gets added to your library, then it will be measured. And that's why, for example, YouTube is so hard for every other company other than Hype but a Measure because there's so much new content. It's not in a schedule. It's not in a catalog. It's not already existing. So when you're using ACR as a solution, if it's not already in your catalog, you're not going to catch it. And secondarily, everyone should be questioning their ACR vendors. When are you putting it in your catalog?
because if it ever aired or if someone was ever exposed to it prior to that, it was not measured.
Brett (16:12.506)
Yeah, and with the growth of UGC content, you think there's such a proliferation of content types, content lengths that you probably overwhelm the system pretty quickly. And know Nielsen purchased Gracenote as sort of a meta content match play, right? And Gracenote would do the same thing. would match content to a catalog. It would show on your television screen or your computer screen what song are you playing from what band when that was relevant.
So that's interesting. how is that architected? I'm curious as to how are you recognizing, so somebody's playing a game, let's say it's a newly released game, or they're watching a YouTube video. I mean, how are you actually recognizing, categorizing, naming? Is it just pulling from existing metadata that's associated with that content at its source? Or how are you doing that?
Joanna Drews (17:01.158)
Yeah, so we have a partnership with TiVo. So we do use their metadata. However, we measure absolutely everything that's occurring on the TV screen on a second by second basis with no gaps. So throw navigation in there, ads during navigation, multi-screen viewing, video gaming. The permutations are endless. as I mentioned, we work with TiVo. We use their metadata library for the inventory that they do have. However,
What we like to say is we measure everything ACR measures and everything it does not.
And there isn't a single metadata library company I can call on in order to inform the unmeasurables. And we as a company are making those judgment calls ourselves. We learn from the TiVo data in order to create new rule sets and otherwise. But we work with our customers and trade organizations to make sure that those judgment calls that we're making today and in the future are aligned with the industry and its expectations as well.
Brett (17:43.355)
Yeah.
Rio (18:07.031)
So, the industry seems to have spent years chasing larger and larger data sets, more this, measure this, more, I mean, that's been the trend, right? We talked about that a lot in this.
Joanna Drews (18:15.185)
Yeah.
Joanna Drews (18:19.506)
Yeah.
Rio (18:19.821)
More data doesn't necessarily mean more insight, right? Like, and I think the scale is to a large extent created more of an illusion of accuracy when maybe things aren't as accurate as they could be. Like, I think your panel approach is very interesting. I'm wondering if you can comment on how a representative panel conceptually can sometimes reveal more signal than billions of billions of.
Joanna Drews (18:22.191)
Exactly.
Joanna Drews (18:31.591)
Yeah.
Rio (18:46.665)
Signals that are collected across all these other places can potentially do love your thoughts on that
Joanna Drews (18:52.016)
Yeah, so we're reflective of the US population, we always will be, to your point, a representative panel in three factors, geography, demography, and technography. But you can build a panel and be reflective of those three things. The biceps, exactly.
Brett (19:07.515)
Technography, I like that term. So that would be device-based, like what technology are you interacting with, right? Yeah.
Joanna Drews (19:14.034)
Yeah, but anybody could build a panel that's reflective of those three factors, right? The differentiators, the technology that you put into that panel, right? So our technology is proprietary. Our software layer is called UNI, the Unified Neuromedia Identification Engine. And that is where our special.
Brett (19:19.353)
Yeah.
Brett (19:34.135)
No connection to that, my favorite Japanese briny, the uni.
Joanna Drews (19:37.296)
Thank
I mean, it was like a special nod to it because it's also my favorite, but it's important. And that software is what's measuring everything. We've basically taken machine learning and trained computer vision to be the media eyes on the video stream. So whether if you're in a video gaming environment, for example, you're not going to change the channel often. If you're in a scheduled environment.
Brett (19:44.5)
Yeah!
Joanna Drews (20:06.992)
set top box, there's certain factors like spaces of time between commercials that are different between content and otherwise. And it's looking for all of that. And then we have a stack of AI algorithms that are source dependent. So if you're using your cable box versus your video game console versus your OS versus versus versus, the right AI algorithm.
is kicking in to decipher exactly what it is that you're watching.
Brett (20:36.141)
On and on what device specifically because you said yeah, and you guys have partnerships. I mean just not not to focus too much in high for metrics I know we're gonna we're gonna continue to talk about that but but just Do you guys have partnerships with other providers outside of TiVo that help with or do you need that when you start to talk about? iPads and iPhones and Android phones and other devices gaming devices PS4s
Joanna Drews (20:59.152)
Yeah, just to clarify, we measure every single individual's media exposure within the home. That means every single TV and how many people are sitting in front of it, every single phone, computer, personal device, IoT device.
Brett (21:07.471)
Yeah.
Joanna Drews (21:16.846)
IoT device in the home and who's exposed to that, right? And our library grows with us. So going back to when we were talking about ACR and that Shazam example, you can only measure what's in your library. The way our system works is we have a pre-existing library because the company is seven years old. We're in a thousand households today.
Brett (21:18.885)
Yep.
Joanna Drews (21:39.997)
Of course, we're using that library to train and our relationship with TiVo to train our models. However, every time we run into a new piece of content, we don't have the challenge of not being able to measure it. Instead, the machine learning computer vision in AI makes assumptions, and then it goes through a process to validate those assumptions. And then moving forward, it will always know that piece of content or ad or whatever it might be.
piece that it saw earlier. That's how we have no gaps in measurement.
Brett (22:13.615)
Yeah, like across devices specifically. Yeah. Got it.
Rio (22:14.049)
So Joanna, so looking at listening to the previous conversation you had with Brett as well as looking at information on your website, like my understanding, correct me if I'm wrong here, is your organization's, Hyphometrics is not looking to actually become a currency, measurement currency for those who don't know the space too well. And Nielsen would be the traditional measurement currency for linear television. There's a lot of alternate currencies now that have popped up in recent years and have gotten accreditation. Video amp, guess, would be one of those, right? There's a few other, yep.
Brett (22:39.255)
Yep. High spot. Yep.
Rio (22:43.149)
There's a few others as well that are gaining some traction. So my understanding is you don't want to be a currency and it seems like you're trying to provide signals to make the existing currencies better. Is that fair to say?
Joanna Drews (22:56.528)
Yes, we've never had the position of currency. We've never had the position of grading the data that we collect. We're purely here to measure the exact origin of exposure to media that an individual
And our position in the marketplace is to create a foundation of just that, media exposures across every single device, omni-channel and cross-device, and allow every company in our space to license that data. And they can create currencies. And we're an integral input to MMM models, for example. Because to your point about large data sets, you can have thousands of them. But how do you stitch them together?
Brett (23:31.951)
Yep. Yep.
Joanna Drews (23:38.995)
you need our data set that shows you exactly what's occurring from an omni-channel and cross-device perspective with no gaps to say, this is how, for example, Yellowstone performed on Peacock, Paramount.
on the set-top box, in prime, right? The permutations are endless. None of those large datasets will tell you that. We'll tell you all the permutations and how it was viewed and what people did before and afterwards and what else they like. And then you can take that knowledge and provide a stronger direction to your larger datasets for greater accuracy.
Brett (23:58.468)
Yeah. Yeah.
Brett (24:17.903)
So before we go deep into MMM, I wanted to ask you is the future a new currency? And part of me just coming from the MarTech, AdTech, DataTech space, I've always been like, yeah, you do need something to train a common set of metrics to trade against to kind of reduce friction from the buying and selling process. Yeah, that everybody can trust. Yeah, because the buyer is gonna say one thing, the seller is gonna see another, and then there's gonna be friction in the process and it's gonna break down.
Rio (24:34.839)
Well, that's how media gets bought and sold, right? You need something to evaluate like what you're buying.
Rio (24:45.293)
Yeah, and until Nielsen came out, people might not realize it was almost like there was not a huge market for actually like advertising and television until there's some way to actually like quantify what you're buying and selling. Nielsen did that. It might not be perfect, but it did change things.
Brett (24:49.081)
Yeah.
Brett (24:53.274)
Yeah.
Joanna Drews (24:56.946)
Exactly.
Brett (24:57.455)
Yeah, and s-
Yeah, and Scott McKinley talked about that, about how, I mean, you've got to them credit for literally creating what we see as the modern advertising ecosystem that started in the days of Mad Men, you know, on television, right? Because you had to have something to trade against, right? Do you think that that's the future or do you think that it's more of an intelligence layer, you know, kind of maybe extrapolating out to MMM and other, you know, metrics and attribution that lets marketers sort of triangulate in perfect signal?
Joanna Drews (25:29.042)
Yeah, it's always been our position since day one, seven years ago, that the market needs an equalized equal measurement of linear streaming and digital that has impressions on top of it. What the market does with that is pure prediction, right? We're just here to give you the God-honest truth and how...
Brett (25:42.841)
Yeah.
Brett (25:47.737)
Yeah.
Joanna Drews (25:53.159)
to interact on it is your decision. Personally, what I think is that we're in a really interesting time right now, You know, just this week we had some big news on the MRC front. Everybody's been debating whether the alternative currency war is over or not. But when I look at currencies right now in our space, I think of Nielsen, Comscore, iSpot, but I think of MMM.
Brett (26:02.169)
This is what we want. We want the personal, what you think, Joanna.
Brett (26:15.097)
Yeah.
Joanna Drews (26:23.258)
I think of retail media networks, right? Being built in-house. Large brands are moving away from their agencies entirely. We have agentic buying that's on the forefront. All of those scenarios need a data set that tells them how we're behaving. The entire purpose of our data set, but all measurement, is to act like a mirror.
Brett (26:25.306)
Yep.
Brett (26:40.88)
Yeah.
Joanna Drews (26:47.246)
and provide companies visibility into what we do as consumers. And right now, that mirror is incredibly murky. So that's end of our data.
Brett (26:55.897)
Yeah, and they want to have some predictive capacity of like if I spend money here, it will effectively reach this audience versus this incremental audience in this other place, right? They need to have a predictive capacity to say, yes, this is accurate to a certain degree of confidence because otherwise you're not going to go out and spend hundreds of millions of dollars on a channel that you have no evidence reaches. Yeah.
Rio (27:16.231)
You won't spend it and it might not be effective.
Joanna Drews (27:19.228)
Well, right, how much money are you wasting, right? You don't know. When your data set is riddled with holes, exactly, if your data set's riddled with holes, let's take user generated content, right? You can't equivalize it to linear content because linear content's measurement quality is so high, and then you have the user generated content, and you know as a CMO that...
Brett (27:25.488)
The wanna maker problem. Yeah.
Joanna Drews (27:44.88)
You want to invest more in it because you're spending more time in it. Your peers are. It's relevant to your brand. But how do you ensure that when you don't have an equivalized measurement to all of your other options, right?
Brett (27:48.432)
Yep.
Brett (27:57.945)
Yeah, how do you aggregate user generated content, is like I said, voluminous and multi-channel into kind of one larger view, which is what's useful because otherwise you're just dealing with kind of multiplicity, right? That's tough to do, right? And so you're saying you guys are helping to create that sort of layer that helps you understand in aggregate.
what that particular media channel is driving or the type of... So how do you do that with... I mean, are some of the challenges, let me reframe this as a question, are some of the challenges around, I know with any sort of panel growth, you need to get to some level of statistical significance, you've got to hit a... That kind of matches the entire US population. So you can extrapolate out from your initial panel accurately. How do you do that with...
Joanna Drews (28:22.308)
Exactly.
Rio (28:23.597)
Yeah.
Brett (28:48.956)
You said a thousand, I thought it was five thousand for some reason.
Joanna Drews (28:52.026)
No, we're in 1,000 households and we're building to 5,000. So we'll be in 5,000 later this year. It's an industry accepted standard to be at 5,000 households, but nobody's ever been able to measure absolutely everything within the home. And we all know there's a ton of fragmentation and it's our milestone to reach these 5,000 households, but we are...
Brett (28:55.821)
you're building 25,000. OK. Right.
Brett (29:04.217)
Yep.
Brett (29:11.289)
Yeah.
Joanna Drews (29:18.77)
for the massive amount of fragmentation that we will see. And then we have to decide as a company, but I'll say even as an industry, how much larger do we want that panel to be? Maybe 5,000 isn't sufficient anymore due to all the fragmentation, right? And the good news is because we're not using technology that's 30 or 50 years old, we actually come in at...
below 10 % of the cost of traditional technology. So there's nothing to stop HIFAA in upcoming years to reach much greater thresholds.
Rio (29:50.753)
Yeah, why not expand further? Yeah, yeah.
Brett (29:51.383)
Yeah, and for context, Nielsen's got 42,000 households, I think, right? 100,000 people. But there's a but, what's the but?
Joanna Drews (29:56.403)
Yes, but there's always a nice but. But those are tons of different panels. They have a panel that measures linear content. Another panel that measures advertising. That was their Arbitron purchase. Then they have a panel here and a panel there. And they pull it all together. And it's that black box that we all trust. And they call it 40,000 households. And then they have.
Brett (30:02.585)
Yeah.
Brett (30:06.724)
Yeah.
Brett (30:12.141)
Yeah.
Brett (30:20.876)
Okay, yeah, Arbitron was audio, right? That was their audio acquisition, right? Was Arbitron? Yeah. Yeah.
Joanna Drews (30:24.624)
Yes, Arbitron was built for radio and it's an audio solution.
Brett (30:29.473)
Nielsen Audio, yeah.
Rio (30:31.745)
Yeah, so Joanna, looking at, I mean, I know you mentioned agentic and Brett, I know you wanted to pivot into MMM a little bit at some point during this discussion, which is definitely germane to the discussion. It'll be really interesting. looking at as media planning and buying and optimization, and people are testing it now as he becomes more agentic, right? Meaning people are using AI to run things and do these processes for them. Like, do you see hyphometrics becoming like a source of truth for
Brett (30:37.476)
Yep.
Rio (30:58.933)
these agentic models or these agentic solutions with the agents themselves as they do this? And do you see your role helping distinguish what are actual human exposures from your panel relative to the signals people are getting from devices, model audiences, or other signals from other places? How do you see your role changing as things become more agentified, is I guess my question.
Joanna Drews (31:25.242)
Yeah, I'm excited for this question. Thank you. Yeah, exactly. Well, the way that marketplace is looking like these days, and it's not that different from what
Brett (31:28.57)
You just write it to a genetic media by a good measurement, which is that.
Joanna Drews (31:42.642)
is that it's entirely reliant on first party data that's coming from the brand. That's a self-fulfilling prophecy. You need competitive intelligence in order to make smarter, more optimized, and streamlined decisioning, right?
Brett (31:56.005)
Well, hold on, let me question that. So you said MMM's entirely reliant on first party data.
Joanna Drews (32:01.53)
No, no, no, not entirely. It is one of the largest components, right? Yeah, it is.
Brett (32:03.158)
Yeah, yeah, because that was like...
Yes, yes, because they tend to be very large models right behind MMM. They can have all sorts of third party inputs, right? You could have a seasonality data input. You can have all sorts of different things. Customer propensity, right? Yes, which would be third party or even second party data sources, right?
Joanna Drews (32:10.705)
Yes.
Joanna Drews (32:22.65)
Right, in that example within the MMM framework, what we introduced is zero party data and a view across every single permutation. And then again, going back to your data science, you need that ground truth to then make sense of the rest, right? And apply weights correctly that reflect the US population and otherwise. In the agentic world and where we're
seeing is that there's a big reliance on brands to own data in order to fuel that decisioning. And then separately, yes, there's access to other available data sets that are
diminishing in quality due to consumer data rate laws and wall gardens and otherwise. But what's missing, again, is what HIFAP provides, which is that holistic view where we spend our time. And that solves a couple of things. One, accuracy for the models themselves.
Brett (33:09.773)
Yeah.
Joanna Drews (33:23.376)
Two, for the agentic buying to be optimized in itself and include competitive intelligence. And then lastly, the tokens themselves, they need to be optimized or else the costs go through the roof. And it can't be optimized without a ground set truth like ours that shows where we're spending.
Brett (33:36.441)
Yeah.
Brett (33:44.141)
It should, yeah, it fundamentally shouldn't be repoling information it already has because that's where you're to get token usage through the roof for any, right? And so MMM, mean, I think, you know, I think in its media mix modeling or marketing mix modeling, depending on who you talk to, is certainly kind of making a resurgence, right? You know, and I think, you know, to your point around signal loss, cookie loss, you know, ATT with mobile.
Rio (34:06.497)
Yeah, partially because MTAs become so challenging with signal loss and the problems you're describing.
Brett (34:09.304)
Yeah, user level attributions kind of breaking to your point Rio, right? Right, know, privacy restrictions, all those sorts of things. And you see a lot of these companies, we talked to some of them at possible where we met you the first time. I think you were our last, no second to last interview there. MutantX, yeah, yeah, so we talked to MutantX, which Mike Finnerty, who's a friend of mine from New Star Days is now the US president of, Circana, well, we didn't talk to Circana, but there's other players like Circana, analytic partners.
Rio (34:24.493)
Yeah, they were shutting down the press room, I remember.
Brett (34:38.537)
so kind of bought, NPD when it was IRI and NPD combined, right? Yeah.
Rio (34:42.317)
Well, there's the old school MMM vendors like AP or Echometric Spread who I think have done, or you work with TransUnion, The old way was we'll look, gather all these data feeds, we'll gather your first party data, we'll gather your media data, we'll gather all these other data feeds, it'll take six months, a year, we'll build this big model, right? We'll run regression analysis on it and then we'll give it to you and then hopefully your agency does something with it. I don't know, is it...
Brett (34:49.571)
Rockerbox.
Brett (35:02.927)
Yeah.
Brett (35:10.094)
Yeah, for the next year's planning cycle. it seems like, you know, and then with Meta coming with Robin and Google coming with Meridian, it seems like which are open source approach. mean, my point was that there's a broader shift happening. Marketers want faster, more operational MMM. Some people call it hypergranular MMM. Right. Do you see that? know, because to be able to make faster changes to their media budget allocations, which oftentimes to Rio's point have taken months and months, you'd have a six month data wrangling and then analysis.
Rio (35:38.165)
or even translating an MMM plan to actual how like the spend gets broken down and then like how it gets, you know, even affect how it affects pacing goes into platforms is I think it's as much of an art, maybe more of an art than a science spread historically, right.
Brett (35:52.367)
Yeah, yeah. And I mean, that, you seeing the same in that? We'll end this with a question, Joanna, eventually. Do you think this is a trend that's certainly you see as driving some of the growth of your ability to sell data into these players? And what do you see happening generally in that space from all your conversations?
Joanna Drews (35:59.367)
Yeah.
Joanna Drews (36:12.238)
Absolutely. when we started the company, like I mentioned, we were always data as a service. I think we were too early to market. Everyone wanted an easy button. We met the criteria of what was being said publicly that didn't exist. But then when we approached those people, we got
Brett (36:23.108)
Yeah.
Joanna Drews (36:34.268)
That's too much, right? And I would say even from a roadmap perspective, if you asked me this question six or nine months ago, I would have answered it extremely differently. But fast forward to today, today's marketplace and what it's preparing for, everybody just wants the data and they want to do what they want with it, right, themselves. And what we're also seeing is that a lot of brands are moving away from agencies and even...
Brett (36:35.77)
Yeah.
Brett (36:52.399)
Yeah.
Joanna Drews (37:02.034)
even some of their vendors and building things themselves, whether it's a retail media network or otherwise. So essentially the way we see ourselves is we're positioning DIY data across the board. And what's happening in the marketplace is that in the last nine months, that's accelerated significantly.
Brett (37:22.126)
Well, and so to your point about in-housing, mean, because you're saying brands, are you saying carte blanche that it's all, they're in-housing every part of the marketing life cycle or is it the measurement piece specifically? Where are you seeing some of
Joanna Drews (37:36.082)
I would say they're in housing a lot of automation regarding trading and they need a data. Yeah.
Brett (37:41.433)
Yeah.
Rio (37:42.945)
Definitely a measurement though, I'm seeing that as a, maybe that's where you start if you want in-house, you start with measurement.
Brett (37:44.889)
Yeah.
Yeah, suit.
Joanna Drews (37:48.316)
Great, but you need measurement in order to make those decisions, right? And that's our position within it. This is the data set to help you guide how to build your...
Brett (37:52.537)
Yeah.
Rio (37:54.157)
What you
Brett (37:59.074)
Yeah, to add fidelity to your existing data sets. Yeah.
Rio (37:59.682)
Yeah, Joanne, it sounds like the hypothesis here would be that by providing this panel-based person-level screen exposure data, really, like from HIFO metrics, like this will improve, I mean, like this will, you'll build better measurement models. If you want, for MMM, this would be an incredible input, I would imagine, to actually, instead of just having these modeled assumptions, right, you can actually look at what's happening and use that to tweak models or maybe build...
Joanna Drews (38:01.596)
Crap.
Rio (38:28.599)
build new ones entirely,
Joanna Drews (38:30.948)
Yes, I think to tweak model, I'll say this, I think our data set takes your model to the modern age. So let's say that we're feeling modeling 2.0. And the reason why I said that is because all models without our data is riddled with holes. let's say, I'll give another example, an advertising example. One for Breeze Commercial, one production, but how many permutations does it have, right? Now with AI.
before AI was so prevalent. Several cuts, you swap out the fragrance, you put it in different markets. We are the only technology that's able to measure that. And if you don't have our data set that measures all the permutations across your TV, your tablet, your phone, your ages, your demographics, your regions, then how possibly accurate is your MMI?
because ACR can only measure a fraction of those ads and can't tell you the difference of fragrances that are being displayed.
Brett (39:31.352)
Yeah, and so this would be a reach in frequency in potential engagement, but reach in frequency, let's say, of media exposure to...
Let's say if you've got 100 variations of a particular ad within a campaign that are broken down by geo and audience types, you're able to actually distinguish each one of those 100 different ads, right? Tie them back to where they were viewed, right? Which is interesting because that's kind of a creative performance play, right? Yeah, that's important. it's, know, love, Rio, you gotta validate whether or not these claims that Joanna is making, that they're the only ones.
Joanna Drews (39:57.658)
Exactly.
Brett (40:08.865)
that track this, but I'll trust you for now, but I do think it's pretty important. I think it leads me to this question is, is there's a ton of talk and we've heard it ad nauseum, especially when we talked to the mobile guys, right, and the gaming guys, which almost feels like a back...
to the future kind of conversation because they talk about clicks, they talk about performance. Yeah, we were in a performance channel, performance, performance, performance, clicks, performance, performance. And it just reminds me, it's like a PTSD back to last click attribution in Google search, the early days of Google search. how are you, like, how do you think about that conversation? There seems to be an obsession with that right now and I don't know, I'm trying to cut through the noise here.
Joanna Drews (40:28.487)
Yeah.
Rio (40:28.749)
Well, it's all about performance, right? Yeah, it's so funny, yeah.
Joanna Drews (40:39.794)
Great.
Joanna Drews (40:52.304)
Yeah. Yeah. So I guess my performance against what? Like, what are you comparing yourself against? Right? And think of yourself.
Brett (40:59.052)
Yeah.
Rio (41:02.317)
Well, if you're a mobile, it might be easy because it might just be a download or an install, right? mean, it's, it's, which is hence the obsession or myopic focus and performance, Brett, because I think it's easier, but for everything else, certainly for television, it's a different ballgame.
Joanna Drews (41:07.237)
Yeah.
Brett (41:12.098)
Yeah.
Brett (41:15.607)
Yeah.
Joanna Drews (41:16.006)
Right, but if you're a large brand, right, it's whether it's television or mobile, it kind of doesn't matter. You're responsible for all of it, right? And if you what are you how can you how can you measure that performance? If you're if your baseline doesn't include all of that behavior, then how could you possibly measure your performance? Right? What are you? How are you quantifying that delta? So that's how we position ourselves against.
Brett (41:24.846)
Yeah.
Brett (41:41.848)
Yeah.
Yeah, and you might be quantifying it in favor of channels that are further downstream. know, somebody's exposed to an ad and then does a search.
right, and they end up committing an act in search, clicking on a blue link, whatever it might be, and I know that's a little outdated, and then that channel, the Google search channel gets all the credit, even though the original exposure or numbers of exposures are what might really be driving that brand awareness. Yeah, and so that's where I feel a lot of those conversations sort of break down, because it's like we're talking about an entire spectrum of kind of human experience with advertising, right? And you could talk at, you know,
Joanna Drews (42:06.063)
Exactly.
Joanna Drews (42:18.171)
Exactly.
Mm-hmm.
Brett (42:22.811)
Grand versus performance, that's a common conversation. People don't operate, we don't operate through a funnel. No human being kind of goes through a funnel. But you do have an initial exposure event, a re-exposure event, and a visual, right?
Rio (42:37.037)
Well, yeah, Brett, it's all related. There was a fact that James Barrow, he was an open market pod like this last week. He made a really good point. He was saying that they've noticed when advertisers stop spending on the TV spots, right, that over time, not even a lot of time, over just a few weeks, the cost for search ends up going through the roof, right? So like people's initial exposures that are generating awareness, are having a downstream impact and they end up paying for it anyway, right?
and maybe even more in less precise way. I thought that was really interesting. like, John, looking at measurement, which is, think, very important here, the trend seems to be, as I mentioned earlier, going from just panel-based to panel plus database. And Nielsen's focused a lot on that in recent years. And then you're seeing that more big data players entering the space with more combined approaches. Do you see us heading towards like,
Joanna Drews (43:06.672)
Yeah.
Brett (43:25.336)
Yep.
Rio (43:34.433)
Will it be one replacement currency? Will it be multiple replacement currencies? It seems like there's a lot of brands are taking multi-currency strategies as well as platforms thoughts on that.
Joanna Drews (43:44.733)
Yeah, I think that there's gonna be multiple ways to transact. What's not going away is more automation, right? And as an industry, we've been working to automate it for 15 plus years. I personally have been working on that through various roles and otherwise. And I think what's happened is that CMOs are more challenged than ever.
to make the right decision and prove that their decision is optimized. And with that, it's created a more competitive marketplace that we're here to fuel. So we're here to fuel all the automation and we're here to fuel the measurement of performance. Just provide that base layer so that Delta can be quantified. But whether you choose CommScore as a customer of ours, whether you choose CommScore as a currency, or you choose to bring it all in
Brett (44:31.383)
Yeah.
Joanna Drews (44:40.05)
and measure your performance.
That's the interesting part about our space right now and where the interesting questions are about where things are headed. How much of spend is going where and what is driving it and what's driving comfort to switch off traditional currencies or not.
Brett (45:05.514)
Yeah, and you guys just announced a multi-year partnership with Countscore, right? Or is that, how recent was that announcement?
Joanna Drews (45:10.354)
We recently announced our partnership with NBCU. Yeah, and then we launched.
Brett (45:15.7)
NBC you yeah and
Rio (45:18.477)
They're now licensing your core product, correct?
Joanna Drews (45:22.832)
Yeah, everyone licenses the exact same product. Where data is a service, we have an API, and you can do whatever you'd like with it except resell it. So the use cases, we have the largest CPG brand in the world as a customer of ours. That Febreze example I shared is why they like our data, because of that accuracy that we can provide within their modeling rate, every single one of their ads and every permutation.
In the case of NBCU, we work very closely with Brian West, who heads content research. And going back to that Yellowstone example I shared, right?
a yellowstone in all of its permutation needs to be understood. And as consumers, we don't turn on our TV to find ads that are measured by other companies. We turn our TV to find content. And how do we get to that content? And then what are we exposed to during that time should be the questions that we're asking. And that's why our customers love our data, because those are the questions we're answering.
Brett (46:30.968)
There we go. So, Comscore, that wasn't an agreement. I thought that you guys had a partnership agreement for person level audience measurement for CTV.
Joanna Drews (46:40.498)
Yeah, we have a public press release with them. Yeah, Yeah, I said, sorry. I I assume part of this is going to get cut. But I mentioned that Comscore was a customer when I was talking. No?
Brett (46:43.858)
Okay, okay, yeah.
Brett (46:52.823)
Okay. Yeah, no, no, no, no, was, yeah, I was just in re-
Rio (46:56.299)
No, recent, yeah, just the recent announcement was for NBCU. That was the big one,
Brett (47:00.022)
was for NBC, the most recent. Yeah.
Joanna Drews (47:02.5)
Yeah, I'm stalling because I don't think either one of them mentioned how many years and I don't want to get it wrong.
Brett (47:07.97)
Okay, got it.
Rio (47:08.759)
Yeah, don't worry about it. So looking at like, your main customers would be, sounds like platforms, measurement companies, just e-sell direct to brands too, the big ones that have in-housed.
Joanna Drews (47:20.464)
Yeah.
Yes, so we service the entire ecosystem. We don't service a subset of it because whatever questions that you have about the performance of your content or advertising, you need a holistic view on the human experience. There's only so much attention that we as humans can provide, right? And understanding that attention is...
what needs to be driving all of our decisions. Whether you're an ad tech company and need to figure out what your next layer of products are going to be. So going back to the unmeasurables, we're the only ones that can measure navigation, multi-screen viewing, streaming, and otherwise. So that's a whole new layer of revenue potentials for ad tech companies. For programmers, again, we're the
company that can measure every single permutation of a piece of content and understand how consumers are finding that content and what are they doing afterwards. For brands, know, we discussed that at length. But also currency providers like Comscore and otherwise, we provide that human layer, that personification that Comscore previously didn't have, but we also provide that unilateral
streaming to linear that allows them to enhance their CTV and streaming products. So the use cases are...
Rio (48:49.143)
So Joana speaking, looking at linear versus CTV, I know that CTV, the promise was more addressability, more targeting, more precision. you look, all it's really created is more fragmentation. It's also created a lot of, I think, suspect CTV inventory, but it seems like that some of that is working itself out, right? As people get, as brands get smarter, media planners and buyers get smarter about that. But thoughts on this fragmentation, like how is it made, I think TV,
Brett (48:58.594)
Yeah.
Brett (49:13.365)
Yeah.
Rio (49:19.209)
uniquely hard. Now granted display has its own measuring and attribution issues, but just looking at TV itself, what is so hard about measuring TV and video, especially right now with all that's going on?
Joanna Drews (49:25.424)
Yeah.
Joanna Drews (49:31.698)
What's hard is that we as a consumer see it as a device. What's happening on the back end from a measurement perspective is you have apples, oranges, bananas, kiwis, and the list goes on. And the stitching together of that is not reflective of that consumer one screen experience. And we don't have a mirror.
consumer behavior that media companies can react to. That mirrors Merkey because every single different way of measuring streaming versus linear versus video gaming, it can't tell you the human experience. And the technology that's used, all the legacy technology that's been used thus far prior to the hypha,
It's the same permutations and approaches being force fit on new platforms, right? So you can't take a schedule data methodology of measurement that worked for 100 years in diaries and linear and force it into a space where there's no schedule. You need a new approach. need a new.
Brett (50:37.855)
Yeah, people are going into like an app. It's like an app experience. It's like a giant phone on your wall, right? Where you're going into a very specific, generally walled garden, right? A Netflix, an Amazon Prime, and you're going into their content library.
Joanna Drews (50:44.068)
Exactly.
Rio (50:53.141)
Yeah, even I think my TV in my office, Brett, I don't even have, I just use, it's a smart TV, I just use Wi-Fi, so I'm always in apps, right? It's total, I'm not using it all. mean, the one living room we still have, we still have cable, but it's wild that like, so.
Brett (51:00.012)
Yeah.
Brett (51:04.661)
Yeah, and that's that overwhelming trend from consumer consumption of content on the television glass has clearly changed in our lifetimes. you know, I was old enough where I used to flip the channel, but there was a table of contents. It was your guide and everybody across the country, you know, different hours watched, you know, the ABC News, watched certain content.
Joanna Drews (51:24.336)
Yeah.
Rio (51:26.029)
Well, it's crazy, think younger people may not realize that we actually used to get a TV guide in the mail or in a newspaper each week and it would have what was listed and everything. It was linear, everything was, you know, had set times and then, yeah.
Brett (51:30.486)
Yeah.
Brett (51:34.229)
Yeah, so it's a lot more predictable. Now it's so fragmented in terms of the applications that you're going to within your experience.
Joanna Drews (51:34.738)
Yeah.
Joanna Drews (51:42.611)
Well, but I'll also challenge you, Brett. Do we know? Do we know how fragmented it is? Because the technology up until today isn't measuring all of our experiences. So we actually don't know much about what's happening within the home. We're working off a subset of a full picture. What, I mean, I'll put it on YouTube. How much linear TV do you watch?
Brett (51:53.28)
Yeah.
Brett (52:02.634)
Yeah.
Rio (52:06.573)
Well, is Joanna, this is, this is what I think so interesting. Sorry to interrupt too. This is, think so interesting about this whole debate is Nielsen, no one ever suggested Nielsen was perfect, right? But what it, what it did was it said, we can give you some assurance that what you're buying is at least directionally correct. Before Nielsen, no one could buy television. No one, very few would buy television spots. didn't know what they're buying. At least this would say, here's more or less what you're buying.
hear the households you're targeting, hear the ratings for these shows, hear the time slots you're buying for. It would give brands some assurance and it could quantify the value of what they were buying. You could put a price on it, even if it wasn't perfect, didn't need to be, right? I think with digital, we've gotten so addicted to thinking we can measure everything. you remember Brian Quinn was on you there, he was talking about how, or Alex Meruka, remember he was talking about how like people on the, we can actually measure what parts of the touchscreen on the smartphone people are actually touching.
when they're within an app. We can get that precise. But you can't get that precise within what people are watching on their home television. You just can't, right? And maybe you don't need to. Maybe it just needs to be directionally accurate. Maybe that's actually enough,
Brett (53:04.599)
Yeah.
Joanna Drews (53:17.476)
Yeah, I hear you. What I will say though is what Nielsen provided prior to the digital era was a independent, neutral, third party, equal measurement, right? And let's not single them out. Let's include every measurement company in the space, legacy or otherwise, here in the US or globally.
Brett (53:32.279)
Yeah.
Joanna Drews (53:41.299)
their tech stacks haven't been able to continue that equality and that independent and third party expectation across digital and streaming to equivalize it to linear. So I think that's where trust is eroding. example, YouTube gives their top 100 shows to the ACR providers.
Brett (53:54.037)
Yeah.
Brett (53:57.41)
they tried. They tried for decades.
Joanna Drews (54:08.594)
Like, where is the neutrality or independence within that? Shouldn't the measurement provider be dictating what the top 100 shows are on YouTube? That's not happening. So when you're asking what's making measurements so hard, it's absolutely a technical problem.
Rio (54:24.429)
So they're getting it right from the walled garden, right from YouTube, right? He's kind of grading their own homework and saying how well they're doing. I mean, most people watch YouTube and like, but that's interesting. Okay.
Joanna Drews (54:26.736)
Yeah, then separately, yeah, but then separately, yeah, but separately, the walled gardens, all the credit in the world. When I was at WPP, when Meadow was growing, they mean they made that easy button for buyers. It couldn't have been easier, right? And now in the streaming space and every other screen that
Brett (54:27.444)
Yeah.
Brett (54:46.892)
Yeah.
Joanna Drews (54:57.362)
consumer experiences in, they're very well aware of what is measured and what's not measured and who they're allowing in and
and what they're saying about their own performance and who they're giving it to, right? So again, without that independence of a technology that doesn't need YouTube's help and can measure linear equal to streaming until that exists, which it does here at HIFU, until that exists at greater scale, I should say, because it does exist.
That's where trust comes from. How can you trust a company that is taking self-reported data from another company and go...
Brett (55:36.083)
Yep. Yeah. And only a sub segment of this, of total reported, of total tracked data, right? A hundred seems like a very small number compared to the total content that's being served on YouTube, right? What happens if you want...
Joanna Drews (55:47.826)
Yeah, but then going back to it, like, is it a small number? Is it a big number? we don't know. We don't know.
Brett (55:51.274)
Yeah, The universe, the universe of YouTube, how big is that? is this going to... Yeah, and if I'm a different type of advertiser that's looking at a different audience cohort, let's say I'm SMB or mid-market, I'm a local advertiser, and I'm looking at cohorts that are sort of in a long tail of content that's not in those top 100, I may not have exposure or be able to track or see that.
Rio (55:56.374)
What's a number they've decided on that suits them, it sounds like.
Joanna Drews (56:13.446)
Yeah. economic, and like the medium buying market marketplace, it's just a market, economic marketplace like any other. No me any other marketplace where a company grades its own homework to determine the value of their product. Right.
Brett (56:23.404)
Yeah.
Brett (56:30.326)
Yeah.
Rio (56:31.821)
So it's a good point about Nielsen, whether it's perfect or not is irrelevant. saying the fact that was independent is what made it valuable and that really hasn't existed today in the evolved space where you have CTV, have linear is only a subset of it, right?
Brett (56:39.35)
Yeah.
Joanna Drews (56:42.886)
Right, correct.
Brett (56:51.595)
Yeah, and in walled gardens have only kind of like exaggerated that problem, right? Because because you have to trust because they have a data moat around, you know, the content in the in the advertising.
Joanna Drews (56:51.986)
Exactly.
Joanna Drews (56:57.358)
Exactly!
Joanna Drews (57:03.482)
Right, and in every marketplace, whether it's insurance or produce or media inventory, you need a third party independent arbiter of deciding the value of that product. And neither the buy or the sell side should have any say in regards to how that is measured. And that's the fundamental problem.
Brett (57:25.621)
Yep. Yep.
Joanna Drews (57:27.122)
And that's why I started Hypha, because we're purely here to just show you exactly what happened and then let the marketplace layer on larger data sets, proprietary data sets, transact on it, and genetically use a currency, whatever makes sense for your brand you should do. But as an industry, it's incredibly important and I would argue ethical that we get the counting right because we like to
Brett (57:55.371)
That's what it gives out to, right?
Joanna Drews (57:57.307)
Right, because we're responsible for the distribution of information across our society. That's correct.
Brett (58:02.39)
Yeah.
Rio (58:02.637)
Yeah, the conflict of interest is the risk, in sense that you're saying. Whether the Walden Garden is grading their own homework, is passing those grades on to ACR providers, it's, I mean, it's the recent publicist acquisition of LiveRamp, right? mean, can they really be Switzerland anymore? That was their biggest value, was providing that independent identity spine and measurement and integration with all the publishers. They could really...
Brett (58:05.909)
Yeah, yeah, and I think...
Rio (58:29.719)
do attribution like as an independent party. So I think that's a really valid point, independence.
Joanna Drews (58:33.33)
Right. Right. today we talk about higher quality data improving our ad experiences, our consumer experiences, being the independent person being more aligned with the media company that they're being exposed to. But let's imagine a world 10 years from now where HIFU never existed and the current data erosion that exists today is what we're all banking on.
from now, that murky mirror is even more murky. So the walled gardens, the media companies, they're all wholly reliant in that scenario on their own data to tell you what you should like rather than buying a data set that shows them what you like.
Brett (59:19.733)
Yeah, well, it sort of backs into that point. Yeah, Rio, data is the new operating layer that's going to be driving a lot of how headless technology interacts with that operating layer, agentic or otherwise. And to me, that supports the argument that you don't necessarily need one company defining the currency. You just have to have the good data.
Rio (59:26.114)
Yeah.
Joanna Drews (59:28.924)
Thank
Joanna Drews (59:47.377)
You just need to be able to trade streaming, linear, and digital, knowing their exact value to you as a company. And it needs to be independent, and needs to be trustworthy. And what you decide to do with it and how you trade is the free market, right? We don't have it in the United States.
Rio (59:57.888)
Easy one standard.
Brett (59:58.07)
Yeah.
Brett (01:00:01.632)
Yeah.
Joanna Drews (01:00:11.858)
The free market decides how to trade and the value of the inventory. How can they do that without a neutral third party representation of that inventory? They can't. They can't. The answer and the disarray that we live in today is the result.
Brett (01:00:21.205)
Yep. Yeah.
Rio (01:00:28.279)
sense. Want to move on to quick hits?
Brett (01:00:30.845)
Yeah, let's do it. All right, so I'll start. What gets misunderstood more? Attribution. This is a tough one. Incrementality or causality.
Joanna Drews (01:00:44.594)
I will say incrementality because I think culturally within our ecosystem there are people who see digital as the baseline for that and there are people who see traditional TV as the baseline for that. So when people talk about incrementality, I think it should always be followed by a series of questions.
Brett (01:01:10.45)
Yeah, that's for sure, yeah.
Joanna Drews (01:01:11.555)
Dan, what is the baseline of your belief system, right?
Brett (01:01:17.153)
Yeah.
Rio (01:01:20.301)
What's the most overrated marketing or measurement metric that people use?
Joanna Drews (01:01:41.426)
Okay, the most overused measurement technology is the self-reported data by the big plot.
Rio (01:01:49.805)
That's fair.
Brett (01:01:50.453)
Yeah, well that was an easy answer.
Joanna Drews (01:01:51.943)
So when you launch a campaign, you click, when you're on one of those platforms websites, you click for measurement and you think you're getting neutral third party measurement, they are controlling that pixel and they are self reporting it to the.
Brett (01:02:08.26)
yeah, yeah for sure, yeah.
Joanna Drews (01:02:10.308)
And I think that's a big dirty secret no one likes to talk about. But that is, I think, the biggest shame in our space.
Brett (01:02:19.063)
Yeah, for sure grading your own homework. One thing AI will make better, and we've talked about this a little bit today, in measurement in general.
Joanna Drews (01:02:28.722)
The one thing AI will make better in measurement, it's going to give access to, it's going to allow for more individuals and professionals to have access to analytics, data, dashboards, DIY. These things used to be very large, heavy lifts. Now you can put it in cloud, can put it in chat, GBT, it doesn't matter, but.
Brett (01:02:51.244)
Yeah.
Joanna Drews (01:02:58.66)
If you're curious, you can actually go and figure it out. Whereas before, the old guard made it very stodgy, expensive, and hard.
Brett (01:03:13.717)
Yeah, it sort of democratizes insights and as long as there's sort of like an underlying data congruence, know, there's an underlying everybody that equal access to data within, let's say an organization, then people can derive insights from that data. And so it does democratize. And we talk a lot about the hyphenated job, right? The death of specialization. Because if people are given equal access to data within an organization, you know, behind a firewall and can probe and
derive insights from it, you can arguably do the job of many people, right across.
Joanna Drews (01:03:48.497)
Right, exactly. And I think that's the puck that we should all be paying attention to regarding where our marketplace is headed. is, what companies have access to that kind of data? What companies care about quality of data? Where does the integrity lie? Because at the end of the day, new insights is what gets attention and
Brett (01:04:07.947)
Yeah.
Joanna Drews (01:04:18.098)
So because we've been data dry as an industry for a long time, so as all these unmeasurables and new insights are coming out, they're getting the most attention. But do they have the ability to sustain and be sustainable into the future and be used because it has the integrity and quality behind it?
Rio (01:04:41.357)
Joanna, will all TV eventually be CTV and if so, when?
Joanna Drews (01:04:48.05)
believe that we have the infrastructure as a country to support that today. So I think there's a federal component to it that needs to be fixed. And I think that's going to be really interesting because there's certain normal season comforts that live in the broadcast world that streaming doesn't have to address today. And as all of this evolves and as our federal government is going
to support our behaviors as consumers. They're also going to have to service us all as citizens and there's a public service related to putting information out there. So right now, it's all over the place, but if we are going to head to an all streaming household world, we have to prepare for that on many levels.
Brett (01:05:27.83)
Yeah.
Rio (01:05:28.909)
Yeah. Yeah, I don't think we have the regulatory framework to support it now. Yeah, agreed.
Brett (01:05:33.845)
Yeah, that needs to evolve.
Brett (01:05:43.352)
Yeah, you've got to some level of what we'll call it equal access or equitable access to some level of content. Let's say it's the PBS's, the airwaves, right? In CTV environments, oftentimes you don't have that because everything is behind a paywall, right?
Joanna Drews (01:05:57.105)
Right. But don't forget the gutting that occurred during the Tyson fight. Right? Like that's that's an infrastructure problem because
Brett (01:06:06.134)
Yeah.
Rio (01:06:06.401)
Yeah, not everyone had access to Netflix and even those who did it crashed many times because Netflix could not handle it.
Joanna Drews (01:06:10.052)
It crashed, right? So in an all streaming world, when the aliens come, how is the federal government going to tell us?
Brett (01:06:10.955)
Yeah.
Brett (01:06:17.801)
Yeah, yeah, exactly. So, bye.
Rio (01:06:20.545)
How will we know the world is ending? Right?
Brett (01:06:23.305)
Yeah, exactly, So, hyphometrics. This is my last question. This is more about...
Joanna Drews (01:06:23.954)
Okay.
Brett (01:06:30.751)
The name, which I think is interesting because the Latin root hypha is it's the building block of mycelium, mycelium, right? Which is sort of that underground web of when you see some, and I think you illustrate this on your site. So that underlying sort of fungal or the base, the building blocks of that fungal web that sort of powers ecosystems of plant life and, that nobody really knows much about or sees outside of the specialists. is it where, like, how did you guys come up with that name? And then how are you part of the hypha network?
Joanna Drews (01:06:32.113)
Yeah.
Joanna Drews (01:06:38.098)
Exactly.
Brett (01:07:00.715)
because I know you're called the hyphen network company so I wanted to see what the interconnectedness of those things was.
Joanna Drews (01:07:06.834)
Well, so HIFAA is the original in nature HIFAA is the original neural network and HIFAA as a company the way we position ourselves is we're here to facilitate a common language across the entire trading ecosystem. within that in nature that what that purpose that that neural network serves is it lives underground and it lives across the roots of trees and
Brett (01:07:10.902)
Yeah.
Yeah.
Brett (01:07:23.446)
Yeah.
Brett (01:07:35.499)
Yeah, some of the largest organisms on the planet Earth are driven by that. Yeah.
Joanna Drews (01:07:38.419)
Exactly. And what it allows is it creates a common language across all of those organisms so they can speak to one another. So for an example, there are forests in Africa where if a giraffe starts eating the leaves of...
Brett (01:07:54.401)
Yeah.
Joanna Drews (01:07:55.475)
of a tree. That tree lets the rest of the network know that that's happening and the other trees change their taste to become bitter to deter the rest of the giraffes. that is the genus.
Brett (01:08:08.439)
prevent the extinction of their fellow trees in the sea. Yeah, that's a it's it's really pretty fascinating. I yeah.
Joanna Drews (01:08:13.274)
Exactly.
So that's what we're here to do. We're here to promote a common language so that no matter where you're sitting at the table of a negotiation, you have the truth right in front of you. And that's what you're.
Rio (01:08:18.381)
That's interesting.
Brett (01:08:31.063)
Yeah, yeah, nothing to do with the hyphen network, which is kind of an environmental climate justice company, right? I searched it up. It was a yeah, there's a there's a company called the just hypha. Yeah, no, it's called it's called hypha. It's a baby. It's a bit. It you know, it came up in Google search under the name hypha. But I thought, oh, I wonder if there's any interconnectedness, you know, between these two things.
Joanna Drews (01:08:38.256)
I know.
Rio (01:08:43.629)
You're making stuff up, Brett. You just...
Joanna Drews (01:08:51.266)
You're so, you know what? This is, it's got, you guys aren't allowed to use this, but like the word hypha has gotten like popular. It's like cooler now. So these companies are popping up. That's a new, like what you just found is a fairly new company. There was one like three weeks ago that popped up and these are all, these are all true.
Brett (01:09:11.039)
Yeah, hypha.network. so it literally is that is the URL, hypha.network. Anyways, but pretty interesting. Yep.
Rio (01:09:17.239)
Well, this has been fun. Great, great conversation. And yeah, I've got to run to a meeting that's like five blocks away.
Joanna Drews (01:09:17.554)
These are all trademark and friendship.
Brett (01:09:22.551)
Rio's got a meeting to get to. It's the end of a Friday. Well, thank you, Joanna. It was a great conversation to dive deep and for everybody that made it this far, visit us at www.signalannoyce.ai on YouTube. We're on TikTok and Instagram Reels now. Lots of shorts floating out and being distributed across channels as well as Spotify and Apple Podcasts.
Joanna Drews (01:09:26.321)
Yeah.
Joanna Drews (01:09:29.936)
Have a good meeting, Ria.
Brett (01:09:48.123)
and lots of articles as well being written by executive contributors, Rio, me occasionally when I have time. So check us out. We'll see you next time. Thanks.
Rio (01:09:58.423)
Thank
Joanna Drews (01:09:59.997)
Thanks guys.
Comments