Data Quality and Transparency

Learn more about our certification here:

Read the transcript here:

Kristina Prokop:

Hi, I'm Kristina Prokop, the CEO and Co-founder of Eyeota. I'm very happy to be hosting the latest instalment of Eyeota's Audience Data Sessions today on a topic that has always been very important to the company and to me personally, as a co-founder of the business, and that is data quality and transparency. Today I'm joined by Timur Yarnell. He is the CEO and co-founder of Neutronian. Neutronian is focused on independent data quality measurement, and we've been working with them over the past months, going through their certification process. So we thought it was a perfect opportunity as we close this up to invite Timur to talk to us a bit about the path to founding the company, the mission behind it. Um, and a little bit about his experiences as a serial entrepreneur. So thanks for joining us today and enjoy.

I'm very pleased to have with me today, Timur Yarnall, the CEO and co-founder of Neutronian. Welcome and really looking forward to our session today. Just a quick background, - we are partners. We are currently going through the certification process with Neutronian and today we want to spend some time talking about Timur's path to founding Neutronian and the founding principles behind it, about data quality and the importance of an independent auditor on market in this field. So, welcome. And why don't you start by giving us a little background about yourself?

Timur Yarnall:

Sure. Yeah. Thanks, Kristina. This is, this is fantastic. I think the quick background on me, you know, other than being a dad of two great kids and living in the San Francisco Bay area, this is my fourth startup and my background is in engineering. And I've been passionate about entrepreneurship I think since university days. So my first two companies were late-90s, mid-200s. They were web content management and video content management platforms respectively. So doing actual publishing and starting to really get integrated into monetization strategies, I think from the earliest days. And then my third company was an audience verification or a bot detection company that we started as a result of seeing you know, false patterns emerge in my second company, essentially seeing getting attacked by bots. And that's really what was the inspiration for the third company. So that's the main entrepreneurial background I did. My third company was acquired by ComScore and I was at comScore for about three and a half years where I ran corporate development and digital product for some time, but have always been passionate about entrepreneurship and data quality and so thrilled to have launched a Neutronian just about just about a year ago.

Kristina Prokop:

Great. Great. Just a side question before we get into the nuts and bolts of Newtonian. What was one of the hardest lessons that you learned along the way between all those between all those companies?

Timur Yarnall:

One of the hardest lessons? Well, the hardest was the for the first company was the crisis around the.com crash and 9/11. And, really, you know, capital efficiency is was something that was not hard. So literally seeing money end up on the floor and learning that it's better to actually take less money and ensure that the model is right. I think that's probably one of the first lessons. The second lesson is that it's, I'd say the obvious one is that it's important to choose your cofounder as carefully as you choose your spouse. So I'm fortunate that I've had, I'm good friends still with all of my cofounders, but it literally is one of those, you know, it truly is a marriage. I'm proud to say that I'm friends with all of them, but it's not doubt at that level. So I could see you smiling and thinking the same thing.

Kristina Prokop:

Yeah. You know, I think you end up, especially in those, in those early years, you probably spend more time with your co-founders than you, than you do your spouses. And my husband always used to joke that he was husband number three. Yeah. So totally, totally understand that. And it's one of those things that's true, you know, through, for better or for worse and you know, those, the challenges are just as challenging sometimes, if not more.

Timur Yarnall:

Absolutely. I find a relationship in the company truly is like a child in its overall sense of, absolutely.

Kristina Prokop:

Great. So before we get into some more questions, I think to frame the conversation so everyone knows about Neutronian and what the purpose behind the company is. Can you just give us a top-level of the problems, what the goal of Neutronian is, you know, why you founded it and what the, what problems it's solving on market?

Timur Yarnall:

Absolutely. Yeah, so there's really two, I'd say quick elevator pitches that we give from Neutronian. One is that we are a SAAS platform that delivers essentially a credit score for MarTech data. So we want to allow platforms such as Eyeota or others in the marketplace that if they, they want to say that we've got fully verified data, they can, likewise for data sellers that, that want to say we've been verified in advance. They can do that as well. So that literally like a credit ranking. The second elevator pitch is more like we are the organic data quality standard. So again, a platform could say we want to go the route of Whole Foods. They can actually say we've been certified organic by an outside body. Likewise, a marketer can purchase organic and a data supplier very importantly can be rewarded for farming organically as opposed to only being evaluated on scale and treated like a fast-food merchant now. So that's the background. I think, you know, we're very passionate about the issue because of the issue of not having a data quality framework in the ecosystem affects everyone. It affects marketers, agencies, platforms, data suppliers, everyone in between. And ultimately it really impacts consumers as well, who, who don't necessarily have trust in our ecosystem, I think to say the least. And that, you know, for, for me, you know, coming out of Stanford in 1995, 1996, when it was really a bunch of starry-eyed optimist about how we were going to change the world with the web. You know, I still have a big part of me that believes in that and it's sad to see the lack of trust that has developed in our ecosystem. And that's definitely a core part of our ethos.

Kristina Prokop:

Yeah, it's interesting. That segues into my next question very well because I was, I was doing a little bit of research before our talk today and I saw that, you know, dating back even to 2013, there are articles about you fighting the good fight in the industry, you know, for different projects and companies you were working on. Is that, is that something that, and it kind of came out of a certain, a certain product you were working on or this because it seems like for many years now there's been at least elements of what you've been doing that keep moving towards this you know, pulling out, pulling out the good, you know, the good high-quality parts of the ecosystem and focusing on that.

Timur Yarnall:

Yeah, I think for me there's three things that really impacted for me personally and my cofounder Tom, who's fantastic and CTO passionate about this as well. But for me, it really boiled down into three things. There's one is I think ethos from my parents and just how I was raised, like maybe always relate a little bit to the underdog and we're just very focused on you know, I would say focused on ethics, not so much morality, but I think ethics is a key part of my upbringing. I'm the second, I think the second piece of the ethos is that my engineering background really in mechanical and systems, it really is a system engineering approach. So you know, system engineering, my inclination is to want to work on the system as opposed to maybe just incrementally extract from the system.

Timur Yarnall:

And there've been some great articles about how, you know, for example, maybe our political system in the US could use some, you know, system engineering or things like that, but very similar, very similar approach in terms of wanting to improve it overall for everyone. And then the third that really I think is the most relevant and impactful is that at my second company. And, you know, we had worked our tails off for eight years to build a great video platform and we were helping local media properties figure out how they were going to execute on monetization strategies. And this was in 2010, 2011 when, you know, monetizing video was very nascent and you know, we were having success and had just started to expand our, our revenue network outside of local media, which tended to be a very conservative, didn't really use traffic generation strategies.

Timur Yarnall:

And in late 2011, early 2012, you know, we were hit with a clawback notice from Google, literally out of the blue, you know, as a smaller company, literally had no idea. And we got a, you know, several hundred thousand dollar clawback noticed and it was just lights out. I didn't sleep for four days. You know, it was, it was going to be... I ended up hiring my future co-founder the University of Wisconsin Madison and had him look at this data and he's like, look, you've been attacked by bots. Look, you've, one of your partners has been using bots to drive traffic. And that just really you know, it was kind of the six stages of, of denial, guilt, anger, acceptance, but, you know, turn it into an opportunity and we first built a platform to protect ourselves and then we realized that it was a massive opportunity to launch audience verification that way. And that's, that's really what triggered the most recent, the 2013, you know, writing about improving the ecosystem that way.

Kristina Prokop:

Yeah. And what was it about when you were looking at founding Neutronian? Was there, what was it about the data space? Cause obviously we know we have a lot of, although our digital space, a lot of people put that all into one bucket. There are endless numbers of you know, parts of the ecosystem. What was it about data that made you, made you hone in on this as the main focus of the company?

Timur Yarnall:

I think the emergence of, you know, with, with my third start at MdotLabs, you know, we were acquired by ComScore in 2014 and very grateful for the experience at ComScore. Met so many good people and really learn quite a bit about the systems in place to I think to process data in an ethical way. And I don't have to go about to, to pull together a true perspective on audience, but right at that time was, you know, we were, if you remember in 2013, 2014, 2016, we were right in that massive shift of people-based marketing, you know, from, from inventory base to, you know, the concept of buy audience anywhere. And I think that concept of, of taking a user and kind of tracking around the web, it just really resonated both as an opportunity and a tremendous risk.

Timur Yarnall:

And I've been studying it since then. I think the risks that became apparent were risks to publishers in terms of not necessarily being rewarded directly for their relationship with the user, but also even I think the emergence of other risks such as fake news in terms of you know, being able to vet quality data. Seeing how that was tied to you know, bots you could, you could have an entire fake audience that you were, you were trying to track. So just looking at that, it, it continued to intrigue me and seeing the impacts, but also then the opportunity of showing a marketer what can happen when you engage with, with a quality audience. And fortunately, I think even within ComScore, seeing with some of the tools we had seen tremendous success with, with key engagements of within say a core brand safety environment that really inspired me to show what was possible. So I think that's it.

Kristina Prokop:

Yeah. Is it, you know, I think from our side, we part of the reason why we found it so interesting is because the, you know, the fact that, that this came up and why we jumped at the opportunity is because, yeah, there are some things that are pretty clear, like I think explaining somebody to have a bot, like how about works and how that becomes fraudulent is, I mean, it's complex, but it's the concept of it can be understood. And I think data is such an area where everyone has their own interpretation of what data quality is. If you ask somebody what's high data quality, you'll have, you'll have three different answers from three different people. Mmm. And it is an area that people don't necessarily understand the mechanisms of how data is collected, who's doing it properly, what the different methodologies are.

Kristina Prokop:

There's a lot of complexity to it. And so I think you know, from our side when we saw, when we saw you guys coming to market, I thought that was a really interesting area to, to move into this independent, you know, into an independent auditing of the data space. So with that getting, I think now we probably should get into the nuts and bolts a little bit before we, before we lose all our time. Can you share with everyone exactly what the process is? What you guys are auditing for? You know what the process is like to go through for a data, for data provider?

Timur Yarnall:

Sure. Yeah. And I think, you know, it's wonderful that you've joined and you know firsthand, you guys, your team is actively going through it. But we are a SAAS platform company so we do have this audit process that is manual to begin with. But we really view that audit process as an onboarding to our SAAS platform because we, we look at this as a three-stage process absolutely, to it to get to a true measure of data quality. There's the first step is to organize what a data provider is saying that they do or also what a marketer wants, right? So there's the discovery process or what we call a disclosure process and that is getting those thoughts organized. There's then the audit process. So you know to say, well you've said what you do, are you actually doing it? And that is a bit of a point in time measurement as we have members on our team that walked through that and do, you know, do testing, review data samples specifically.

Timur Yarnall:

And we'll get into the five general categories that we break out as an audit. But after the audit there is a third step obviously as we're, we're just getting into it with, with Eyeota and with our other clients. There's an ongoing monitoring capability. So with GDPR and CCPA, you know, we don't think it's enough just to have once a year or even twice a year audits. We believe that having continuous monitoring of data samples to look at new sources, to look for statistical anomalies, et cetera, and to really help our data partners monitor their own customer deliveries to make sure there are no misses. That's a key part of the framework as well. So just wanted to spell out those three steps of, you know, disclose, audit monitor, but the audit piece itself, which Eyeota is in the midst of completing now, you know, it is a both a manual and a data review.

Timur Yarnall:

We have our Head of Customer Success walk through the process in a kickoff call. We then have a questionnaire that we deliver through our platform and it is a rigorous Q&A, written Q&A session that we break out into five categories. And within each category there's 10 to 15, what we call aspects within each one of those are always evolving. But those five categories are consent and compliance is the first one dataset characteristics. So is it fully modeled? Is it more first party, a mixture of et cetera? Performance would be the other piece of the data that we would look at. The data processing, actual methodology and security is a key part. And then fifth would be the transparency, the source transparency of those databases. So currently we are doing a certification process. I think as you know, it's more of a binary meets standards or not, but within those five categories, we actually have metrics that we have a score that we keep internally to get to a, essentially a pass/fail. In the future, we will move to more of a scoring capability, especially when it's talking about a specific purpose on a particular segment level to get fairly deep in the weeds there. But that's the overall processes. Like it typically is anywhere between a six to 10-week process with a team working on it. I'd say, you know, several hours per week.

Kristina Prokop:

Oh yes. We can, we can, we can speak to that. I think for us where it was a, you know, it was a good process. It was a good process. It took, it's not that it took longer than we thought, but we, it was also, it was, it was good. It was good going through it, going through the steps from our side as well. Got us. It helps you also get kind of like the GDPR process. It helps you, not only does it go through the certification, but it helps you, helps you get a couple extra you know, pieces of documentation in place and get your ducks in a row on making sure you can you have every, you know, everything in place in terms of audit trails and everything else. So it's been, yeah, I mean it's been a great, yeah, it was a lot of work. But, but well worth, well worth the effort. Mmm.

Timur Yarnall:

And we don't want it to be a one way communication. That's good. Your team has been fantastic in terms of that, but it is a chance to improve on both sides and particularly as an early partner. I know you helped us as well, but, but I think we hopefully, you know, we bring a lot of experience in this. We've done these types of audits. You know, Tom, my co-founder had done SOC 2 audits. I've done data quality review and I built out products. So we hope that that do that value. And it certainly seemed like your team was very receptive to that. And we, we appreciate that because we want it the process overall.

Kristina Prokop:

Yeah, no, absolutely. Absolutely. So if you look at the if you look at the response from the industry so far, have you had, what has it, what is the and what is the support from the industry been? I know, I know that you have some, some, a really good mix of backgrounds on your, on your advisory board. What is the general response been from market and the support that you've seen so far?

Timur Yarnall:

Yeah. The response has been nothing short of overwhelming in terms of both our investors and advisors. I, I, you know, again, this is my fourth startup. We have close our first round of financing and are doing quite well, even in the midst of the crisis. So I feel very fortunate in that, but our investors include a great mix of Silicon Valley and Silicon Alley, New York-based angels. So we've got you know, top venture capitalists like Hans Tung who invested directly. We've got Eric Franchi of Math Capital, you know, Amar Goel of who's co-founder of Pubmatic. Auren Hoffman of SafeGraph et cetera. So that's been amazing. But on the advisory board, you know, folks who have been willing to share their perspective both from the buy and the sell-side has also been overwhelming and humbling. Humbling. So Steve Katelman from Omnicom, Amanda Martin from Goodway Group, we just announced that Mark Ginsberg who runs marketing for the SMB business at American Express has joined. Serge Biscard, who ran the Ogilvy Paris offices and now is at another company in the EU, so that perspective. And then we actually just announced that Deven Sharma who's a former president of Standard & Poor's has invested in the company and is taking an advisory position. Well, so just, you know, amazing response. And I think it just indicates that, that folks have seen for a long time, this needs to happen, but the time is now with GDPR and CCPA, you know, this type of framework needs to exist for the benefit of everyone.

Kristina Prokop:

Yeah, no, definitely. Great. So I think, you know, it's there's, I think the one last thing I want to touch on there is, you know, obviously there's a lot of there's a lot of industry interesting talk around it, but what I thought was really interesting is not only is it in the industry, but you've got you know, even publications like Harvard Business Review now releasing articles about the importance of data quality and things like that. So it was obviously, you know, it's a topic, it's a topic that has, is a top of mind for a lot of, a lot of different sectors, a lot of different you know, a lot of different areas of areas of business. And I think you know, it was, it was mentioned, one of the articles that I was reading what I think is so critical and, and you know about looking at data quality is because unlike measuring an impression, whether it was fraudulent or not, you have, you have so many products that are built off the backbone of data.

Kristina Prokop:

So you have everything from artificial intelligence to model, to lookalike modeling to do you know, even just the, the, the simple targeting of advertising. But if you don't have that base correct and you know, high quality that can throw all of your other, you know, all these other products that are built off the back of it of. So I thought, you know, this article was really interesting and I know we had spoken to before. Were there any, were there any interesting takeaways that you took from that you found in that article?

Timur Yarnall:

Well, the Harvard Business Review article. Yeah, it was fantastic and timely. I mean, we didn't we weren't involved in it directly, but it spoke exactly to what I think what we're all working on. So it pointed out the issues. It was a, a good study in terms of using some some walled garden data as a baseline and then measuring with survey samples against the reality of in certain datasets. And I appreciated their approach because they didn't necessarily try and demean any particular data provider. But they did point out, I think a key takeaway. They did point out that the average accuracy of demographic segments that they saw was only 42%, which is less than the chance of just flipping a coin. So they really did use that as a good chance to show what is needed here.

Timur Yarnall:

And we participated shortly after that. We participated in a ANA working group you know, for, with several brands and, and some agencies directly in terms of trying to get a overall framework for them to look at data quality. And it was a good, a good output there. It wasn't, it's not our methodology directly, but we, you know, we kind of melded in with a lot of other views on the committee. So I think that was a positive step. And the HBR paper was certainly well, but yeah, this is, this is a broad issue and it's, I think it's safe to say it's not just within MarTech. This is you know, issues with insecurity within credit rankings. You know, so many, so many things that we talk about AI and the, the, the positive and the negative impact that algorithms can have on people's life.

Timur Yarnall:

I mean, when we unknowingly or knowingly introduced bias into our algorithms and you have things like, you know, I think that the story that came out was that there was some very, very successful, very well known women who were getting denied credit card applications because of a gender bias that came in. You know, that just is the ridiculous thing that happened from a data perspective. So we have so much work to do in many ways. But yeah, that was a, that was a good thing to focus on.

Kristina Prokop:

You picked a good area to keep yourself busy for many, many years to come.

Timur Yarnall:

That's right. Yeah. And it is, that is the thing, you know, my co-founder and I especially, you know, Tom was he was CTO of, of a brand safety engine called Proximic. He was also quite senior at Amazon AWS and managed the top five AWS accounts for some time.

Timur Yarnall:

So, you know, Tom and I jointly could have done many things and, and could have built a very successful consulting business in cloud computing if we wanted to, for example. But this is an area that we're very passionate about and we know we can have an a very positive impact. And that's, that's what he and I care about doing with our lives and our careers at this point. It's obviously a great business opportunity, but more than anything we want to look back on how we're spending our time. And so we've done this with a real positive impact and that's why we made the choice to do here and why we've been able to invest in the company ourselves and also bring in, I think very, very smart motivated investors too and, and partners.

Kristina Prokop:

Great. Well, I can't let you go without asking the question that everyone asks. How has COVD impacted your business?

Timur Yarnall:

Yeah, yeah. COVID. You know, it makes, first of all, heart, you know, my heart goes out to everyone I know that ha that has been truly impacted by it from a health perspective and an economic perspective. We were very fortunate that we closed our financing in January, February and that we are a seed-stage company. So very fortunate in that we've actually seen continued interest in demand for our services that we had put out. We've, we've continued to see inbound expressions of interest because after, I think the initial shock of the crisis, you know, this issue remains and it's probably even more important as we said, I spoke to a market and they say, you know, marketing dollars are even more precious now than they were four or five months ago. So the cost of being wrong has just double or triple. So that's actually been good for us. And then we're actually fortunate that we've also had additional investors come in so that we have, I think, you know, clear line of sight through 2021 interim what we can do. So that, that's how it's impacted us. But I feel fortunate, frankly, to be seed stage because it's not it's a, it's a little bit easier to navigate maybe when you're, when you're a bit smaller.

Kristina Prokop:

Yeah. Yeah. I mean, it's not an easier for any of us, but but we also have, we're all thankful we're not working in industries that are, that are impacted as, as heavily as many. So Timur, thank you so much for joining me today. And I want to say a true thank you for forcing the discussion around data quality. I also agree that it's, it's something you know, incredibly important for, for the future of where, where our business is going and where the ecosystem has to go. So awesome initiative. Wish you all the best luck. We are supporters and we'll, we'll try to, we'll bring more supporters your way. But you know, if anyone has any questions please feel free to reach out to us about our processes, the how the process went for us or out to Timur. I'm sure you guys are all proficient enough to find us on LinkedIn or contact us directly. So thanks so much for your time and say healthy and safe.

Timur Yarnall:

Yeah, I do need to, I do want to just acknowledge you for being among the first of our partners and to really force the dialogue as well, so that that needs to be acknowledged and we will do that. But yeah, thank you very much in you and Eyeota are to be congratulated on that.

Kristina Prokop:

Thank you. Okay. Have a great rest of day. Talk to you soon. Bye.

Timur Yarnall:

Bye