Data privacy, data compliance, AI, and ChatGBT are some of the most important topics for local media right now. In a recent episode of our Leading Local Insights Podcast, I had the opportunity to dig into each with Timur Yarnall, co-founder of Neutronian.
A subject matter expert on data, Timur is a successful executive and technology entrepreneur. He’s the co-founder of four start-up companies that focused on data privacy and consumer trust, including his most recent projects with Neutronian and Neeva.com. He also co-founded MdotLabs, which was acquired by Comscore, an advertising and media evaluation company.
We discuss the implications of media privacy and data compliance in the local advertising space and examine the use of AI in an environment where privacy is constantly challenging targeting abilities and impacting the effectiveness of local advertising. It’s a fascinating discussion that cover important ground.
You can also read an edited transcript below.
Hello and welcome to BIAs Leading Local Insights Podcast where we focus on the trends, technologies, and activities driving local media advertising. I’m Leyla Chatti, senior media analyst at BIA, and I’m here joined by Timur Yarnall, who’s a senior digital products executive and technology entrepreneur, who is the founder of Neutronian and has dual role as head of business development of Neeva.com. Today, we will discuss trends regarding media privacy and data compliance.
First, let me give you a little bit of background about Timur. His father, Dogan Cuceloglu, was of famous Academy writer and professor who taught and public speaking on media psychology and Turkey. Carrying on his father’s teaching on personal development, Timur persevered in mechanical engineering and co-founded four startup companies focused on data privacy, and consumer trusts, also, including one that was acquired by an industry leading advertising and media valuation company. Also, prior to founding Neutronian, Timur was also the Senior Vice President of digital advertising and corporate development at Comscore, after it acquired a bot detection company he co-founded.
In the last few years, the need for increased marketing capabilities, efficiency, effective targeting, attribution, and customer retention have been really the focus of marketers. So without any further due, let’s jump into the dynamics, the functions, the implementations of these emerging technologies that make these aspects possible. So two more welcome. Thanks for joining us today.
Can you start by telling us more about how you came about launching Neutronian and where the key name came from? Also, why did you see such a need for such a product?
Yeah, thank you, Leyla. I’m glad to be here. Thanks for the full background, I think as you pointed to my father, who’s a national icon in Turkey, and a very well-known psychologist, and really brought I think family psychology and modern psychology to Turkey, really influenced my thinking and, and both around entrepreneurship, always encouraged me to take risks, be an entrepreneur and to write. So, I think that was a thing so that led to my both like the left or right side of my brain with engineering, as well as you know, jumping into the startup world, and not being afraid to take risks, and I think, potentially embracing failure as part of the path as opposed to the endpoint. And so that led to the startup life and I think has influenced, you know, business and life overall.
And as far as Neutronian, we are a data privacy analytics and data quality scoring platform. So our goal is to be like the S&P or Moody’s, for scoring data privacy and data quality, meaning local audiences that are having ads targeted toward them. We want to make sure that there’s a proper global framework in place to make sure that that’s the audience that the marketer expects, that the audience itself is aware of how they’re being tracked, and that the local media platform and provider is protected, saying we follow these standards, we follow this framework. We are being transparent and are following applicable compliance laws. And you know, how that came about, I think multiple levels: one, I have a core belief that data privacy is a fundamental right and part of democracy. And I think we’ve seen globally, the impact of that when those rights are eroded, and when they’re supported. And so that’s just a really, personal passion of belief for me.
And then, as you mentioned, my third startup MdotLabs was acquired by ComScore, and I ran the audience verification teams at ComScore for several years: right there, I faced firsthand the issue in finding high quality data partners. Everyone would come in saying it “we got great data, It’s ethical, it’s privacy first, it’s all first party” and there was no way of measuring it. I mean, we had to dig in and essentially put everyone through an M&A mergers and acquisitions level of vetting before we could find was it actually good data? As a complainant in most cases, I’d say 80%. Plus, we would find issues that would have us walk away from the data partner. So certainly, my time at ComScore influenced that quite a bit.
And then I would say my second company, broadcast interactive media where I first met BIA and I worked with local platforms across, you know, Gannett, Abeelo at the time, I think we worked with you and hundreds of local TV and radio stations, that very much obviously influenced my thinking to that time, even prior to doing bot detection. But seeing, you know, local faces face special challenges, and has special value in terms of the local content generation, the local news reporting and by large, that’s been discounted, I think, at the broader market.
A lot of what Neutronian is doing is trying to highlight lower volume publishers that are exceptionally high quality, because the market has moved towards valuing scale only, and not valuing that. So in a sense, the market has moved towards viewing everyone as fast food when there are, you know, organic and much higher quality level of content and audiences out there. So, all of those things inspired Neutronian and I think it’s something we’re still driving towards.
That’s very interesting. Thanks for giving us some background.
What are some of the challenges now that Neutronian is there and operating? What are the kinds of challenges that you’re seeing today? And what are the new trends that you’re seeing in Martech?
Sure, so Neutronian launched three years ago, in March of 2020, the day before the shutdown order was issued in San Francisco for the pandemic. One massive challenge we faced was just for the first 14 months, we couldn’t meet with clients and prospects face-to-face. So again, this is my fourth startup, I hope to never again launch into the teeth of pandemic but I’m proud to say we have persevered. We have grown, we have fantastic clients and partners that are very committed to the business and to the vision, partners like DoubleVerify. We recently announced Dynata, a major survey platform, as a client and as a partner, Bombora on the B2B side.
Among my stars are the most committed customer base, and my co-founder and co-CEO, Lisa, who’s based in the Netherlands. We’ve persevered through this.
The other I’d say challenges just besides the tactical is that data privacy for a long time has faced cynicism and lack of attention. If we rewind 10 to 12 years ago, just like when brand safety first rolled out, when websites were first faced with, we want to vet your domains to ensure the content is brand safe. Nobody cared, it took a while to get momentum. The analysis was initially manual. Agencies were doing teams of manual vetting of domains to see if they were safe for brands that have now evolved to become standardized, automated. It’s gone from basic binary 01 to is it brand safe or not to involve very nuanced machine learning and detection systems that can do sentiment analysis on post publisher pages and so data privacy scoring faces those same challenges right now.
People wonder why it matters. Is it important for my brand? Well, you must do it manually, does anybody really care, etc. And we’re going to see that same trajectory two to three years from now, huge steps forward in automated scoring, along with regulatory change, as well as making it part of an opportunity and showing that data privacy correlates to actual better performance for marketers.
But you know, stepping back, there’s still a huge amount of skepticism. I tell the story, which was originally featured in a publication, I told it to another editor. At CES this year, I was meeting my first meeting at CES, which is always a little bit of a shock. You know, you come from the holidays, and you’re a citizen.
My first meeting with a major identity platform, I won’t name the name, but everybody knows this platform. I was meeting with six of their executives and five other people in the meeting. Their CEO was out on the call, but their chief product officer looks me in the eye and says “Look, this is all baloney. Anyway, nobody cares about data privacy. This is all just for show. None of my customers ask about it. And we do it so well that we don’t have to worry about it” and I was shocked that he would admit that in front of a meeting. But I turned right back around and said your customers don’t ask about it because they don’t want to know, right? Like your whole value proposition is that you’re essentially indemnifying them from the risk and you’re taking the risk, and we know that. Those are the huge challenges that we face it. But that is changing rapidly.
I think with the Sephora fine with the FTC actions, the recent fine against BetterHelp for sharing data and, in a post Dobs world, privacy means something to a much broader group of people. And there’s people who are willing really to fight for that. And we’re fully supportive of that, I think, when, when a consumer or when somebody can be prosecuted for looking for relevant health care information, I mean, that’s something that needs to be addressed by the ecosystem. And we need to recognize that if we miss target someone, they happen to look at an ad and they can be investigated because of that, something’s gone completely off the rails. We must do a much better job of defining privacy and protection and that’s very important.
Thank you so much. That’s really great. I think there’s a lot of room for that and we need to dig deeper. Now I wanted to change gears a little bit as you mentioned in your time at ComScore.
Can you tell us a little bit more about your bot detection company MdotLabs?
MdotLabs was, as so many kind of inventions are, born out of necessity, not a value. My second company actually, Broadcast Interactive, was a technology provider and an ad network provider to local media. We launched it as a kind of a local version of YouTube with video and helping broadcasters get their video online way before they knew how to monetize that video. And getting a video online, we also realized they really needed help. You know, this is 2011/2012. Even the idea of a pre-roll or a mid-roll was something we had to explain to people. We were helping with the advertising side of that at Broadcast Interactive. We started seeing weird traffic patterns emerge, and we started seeing user behavior. That made no sense! At one point, we got a notice from Google that will look like you have bought traffic on your site: we had no idea what that meant.
I went out and I found an academic and a cybersecurity expert, at University of Wisconsin Madison, where I was living at the time. He looked at the data and said, these are bots. And I said, I’ve got no idea what that means. I learned about it quickly and was like “Oh my gosh! This is an existential threat for the ecosystem.” And we turned that into an opportunity and launched MdotLabs as a bot detection platform. And, you know, fundamentally doing data analytics, you’re identifying bots and you’re identifying the flip side of humans. And similar to the story I just told about brand safety is what’s happening with data privacy.
We started with a rather simple, almost manual process. But you can imagine the way that bot detection works and the way that we did it, which is a very specific type of analytics and data science, called anomaly detection. We’re looking for anomalies in the data that would flag non-human behavior.
When it started, I was looking for simple traffic patterns, such as Do we see a user with a clearly misconfigured user agent browser string? Do we see a user that appears to be clicking or surfing 24 seven all day, every day, clicking at regular intervals. Those are very simple flags of bot activity that even a layperson can understand. And then from there, you actually start building up more and more and more nuanced and more sophisticated layers of bot detection, that take into account different methods and different data streams. And you get to where we are today where once MdotLabs was acquired by comScore, along with Human, which used to be called White Ops, we became part of the MRC standard that defines what IVT is today: invalid traffic and invalid traffic can be broken into general invalid traffic, which is kind of brute force bot attacks. And then very sophisticated IVT SIVT, which is more around nuanced and more targeted pieces that might go after a particular publisher or vertical.
So MdotLabs was a fantastic experience. It really was my strongest introduction to Cybersecurity and kind of that hybrid between security and privacy. And now with my role at Neeva, really looking at where generative AI comes into play. It was just a fantastic experience! So very grateful for it and yeah, it was a wonderful team at MdotLabs and we became the core data science team for ComScore. That data group is still the foundational layer of ComScore`s activation business which just got renamed I think as Proximic by ComScore, which is wonderful! It’s also humorous because we acquired Proximic back in 2015, so it’s interesting to see ComScore resurrecting that brand name and building that whole offering around it.
That’s such a fascinating story! Even the whole technology of it and just working there, I remember, there were a lot of emerging technologies that were being adopted. So, I’m very interested to learn more about it. Can you share more?
It is interesting and just to interject, our team came in and said, Look, there’s amazing data at that conference. And we built out the ComScore device graph at the time and this is back in 2015/2016. And already building device graphs, you’ve got different players that had that end, in terms of privacy and analytics. I mean, it’s scary the lead and the type of data that these companies have. We found that we could build a device graph that could track an amazing number of people at a very granular level, to the point where we needed it. We tested it on our own phones and the location tracking capability, and the level of what could be there was relatively mind blowing.
In terms of knowing that there needs to be a data privacy framework, and we need to move beyond I mean, that was something we saw firsthand, but very careful to follow rules. ComScore. I would say, is one of the most careful and Neutronian actually does not have a commercial relationship with ComScore, but ComScore rates very highly in our reports, because they’re very buttoned up from a forward-facing standpoint.
It hammered home just the level of data that the walled gardens and even large publisher networks have about the everyday consumer is mind blowing. And when that data leaks into the bloodstream, and becomes available to everyone, including foreign adversaries. When you think about the fact that index exchange was based in Moscow up until about 18 months ago, it’s very interesting to think about. I’m talking about data privacy and governance and how that impacts how we view Tik Tok now, etc., and bitstream usage. These are vital, democratic issues, and they really impact how MarTech is perceived.
Absolutely, thanks for giving us a bit more background on it. I mean, it’s just fascinating how we’re all evolving into that space, eventually. We were just mentioning Neeva.com, which is your most recent project, which is an ad free privacy first search engine. You said they’ve raised $80 million from Sequoia and making big waves into the generic AI space.
Can you give us some more background on its capabilities? Where do you see the pros and cons and AI in this?
Yeah, absolutely. It’s a big topic. I think the gist of it is that, as I’ve moved to a co-CEO role at Neutronian, partially that’s because I believe that the changes in search capability and generative AI does hold the key to what’s going to allow Neutronian to scale, search and discern privacy practices, etc…because we are analyzing and crawling publishers all the time. And then here, you have Neeva, which is literally a search engine, crawling engine that’s got incredible capabilities. There’s a clear tie between Neutronian and Neeva from a vision standpoint that I see, so I’m very comfortable with both.
Neeva is a high-powered AI team that has raised 80 million from Sequoia, amazing technology. You can use the search engine right now on Neeva.com. You can download the apps, both Android and iOS. And it’s amazing to realize there is a private alternative to Google and even to DuckDuckGo. You know, DuckDuckGo talks about being private, but they monetize everything through Microsoft ads, and they show ad results before they show actual real search results. Whereas Neeva is all about the user, it’s all about answers and not ads, and generative AI, which is the simple version of generative AI in search.
We’re not just going to do links back but like ChatGBT, when we put in a user query, Neeva goes out crawls multiple sites and comes back with an answer. It’s really mind blowing to think about that capability! And I’m sure many people listening to the show have already experimented with ChatGBT’s amazing power, yet incredible risk as well for false answers for what are called hallucinations, for misinformation to be spread.
One of the things I’m very proud and passionate about is that Neeva has already focused on user privacy, has focused on generative AI, but has already proactively inserted publisher and local media. We inserted citations into our generative AI results so that a user can see where the answers are created from, you know, one, two, or three, four sources and can verify for themselves, you know, are those sources trustworthy? We’re not just going to come back with a general answer, we’re going to tell you where it came from.
To me, the potential for publisher partnerships, we will be reaching out to local media partners very aggressively soon, to partner with Neeva to use our API’s. We think we can enable this type of search experience within sites, which is going to be very exciting to marketers, advertisers, national and local and as well as hopefully, local viewers and readers. And so yeah, that’s what Neeva is about, I think of Neeva as really an amazing AI lab that has yet to really broaden the commercial applications beyond a direct-to-consumer approach.
My role with Neeva is to create the industry partnerships with publishers, with enterprise and data, and grow it that way. So, it’s a wonderful fit, I’m very excited and happy to be to be part of that team as well!
And how does Neeva identify and manage risk versus let’s say, the performance of AI in its search engine?
There’s a ton of methods that we do on that, I mean, we’re using co validation of other results, we’re constantly taking our own sources to use as baseline sets, make sure we’re on target and we are reaching out to other companies that are building and are testing our own large language models. So, you may have heard this incredible change in search is all underpinning by this change in the actual model, what you come back with is called large language models, or LLMs.
The explosion in AI capability has allowed for these LLMS to become so much more powerful to check for a wider array of potential results. And I think as far as risk and misinformation checking, the main thing is to integrate with a wide variety of sources that we believe we can know and trust, as well as to make it apparent to the user that they’re also responsible for that and that they can be part of it.
If you go to Neeva right now, there’s both citations in the answers. But one of the greatest things that we have is something called a Bias Busters, where a user can toggle sources. And if they want to see what a left leaning source is, they can toggle to the left. If they want to see what a right leaning source is and see the difference, they can toggle to that. So again, I’m very proud of being transparent about that risk and I think that’s an important step.
Thank you Timur. That’s very cool. And I’ve got only one more question.
Where do you see the ad world going? In your own perspective, especially in terms of digital channels, targeting local consumers? Now I know AI brings a new level to innovation to the space, like we just talked about, and privacy is going to be an issue going forward in terms of first-party data and targeting. Fraud will also continue to be an issue. What are the most important opportunities for improving local media and ad markets?
Yeah, so I think there’s probably four major overlapping trends that I point to right away for local. One, I believe strongly that privacy is not for just a few people. I mean, it can be dismissed as only a few people care about it. I think privacy is now mainstream, it’s here to stay! We see that based on Apple who has pivoted its entire marketing message around privacy. And so, when the most valuable company on the planet takes that stance, I hope people do wake up and pay attention, although unfortunately far too many are dismissing it merely as a marketing tactic. But privacy and showing that users and viewers can have their privacy choices honored, I think is very important.
Secondly, the importance of local journalism and the awareness of misinformation, and the damage that can do, is also a very strong trend for advertising to engage with that.
Then, the changes in identity: So, the issue of cookie deprecation and just where does identity come from, will continue to be a massive trend. And that’s where this concept of publisher first-party data, I think it’s well known. But every local media site and property should be thinking about how do they get their users to engage? And how can they transparently take in first-party data that they’re going to be responsible for, and use that to build really good ad experience?
And then the fourth trend is generative AI. In terms of that engagement, how are people going to engage with your site?
In this new world are they just going to go to ChatGBT? Our view at Neeva is that generative AI needs to be much more like an Ironman suit for local reporters and not a destructive force. So, we want it to be an enhancement. Let’s call it a bionic man or a bionic woman. And I think that trend is very important. So those are my trends. I’d love to follow up more. I think we could talk for days and days, right but I really appreciate you having me on. And I look forward to working with BIA quite a bit more and just really appreciate you taking the time to learn.
Thank you so much Timur, we really learned so much about it and we’re probably going to dig more into the subject going forward. And thank you for your time today. We really appreciate your time.
To our listeners, if you are a BIA ADVantage client, we encourage you to go into advantage and check your local market and for local business opportunities.
Please let us know if you have any questions and our team is here to help. Send us an email at firstname.lastname@example.org and we will get back to you right away. And finally, for everyone out there listening, thank you for joining us for this edition of BIA`s Leading Local Insights Podcast. It was a pleasure talking with Timur today. And we will continue covering trends in Martech and AI as this space continues to evolve and grow in so many ways. Stay tuned and have a good day and we’ll see on the next episode!