And actually what the decision factors were made in terms of what ad I saw versus what my friends saw or my wife or my parents saw. And then from there, jumped even further down the media and data rabbit hole, working for Neustar, running their product marketing organization, focusing on identity, focusing on how we’re resolving identity, how we’re using identity to power advertising and marketing effectively and efficiently. But also again, further down the rabbit hole of identifying what actually is happening behind the scenes, how the proverbial data sausage is being made, how it’s being managed and how essentially it is turning into the ads that we see. And potentially a lot of these other issues that we’ve started to see over the years, programmatic advertising and other kind of nefarious ways in which ads essentially using our information to drive information in front of our screens and our eyes.

And so I worked at Neustar for about seven or so years. It’s a TransUnion company. I want to make sure I want to say TransUnion in that as well. And then decided to look to where else could I spend my time and my expertise and my knowledge to do a little bit more good inside the four walls of marketing and advertising. And so I decided to work for InfoSum. And InfoSum, for those of you who are not aware, is a data collaboration and clean room technology and infrastructure. And the goal for us is to safely allow organizations to use the intelligence they’ve collected and harnessed from their consented consumers to drive better experiences, to drive better analytics and drive better performance without actually sharing, leaking, exposing any information whatsoever, ever, and eliminating the movement of data across the ad tech ecosystem. So really gone on a very unique journey from a new person in the ecosystem, learning about identity, and now essentially figuring out how to protect identity day in and day out. So happy to be here.

Kyle Holloway:
Yeah, Devon. Great love hearing your background and in today’s ecosystem, folks like you are so important that have taken an opportunity to see a larger swath of the ecosystem and how that’s interrelated because the fact is there’s a myriad of capabilities at play. And so that’s awesome that you’ve had a chance to deep dive in different segments there. Just playing off of what you were just talking about in particular, where you’ve ended up here at InfoSum and talking about clean rooms, clean rooms in and of themselves within the industry is a generic term. I mean, you hear it a lot, but it’s not really clearly defined. And then you hear things like InfoSum. Are they synonymous? Is just a byproduct or a type of platform related to InfoSum? I mean, give our listeners some context there.

Devon DeBlasio:
Yeah. Great question. And we’re still working through the clarification and the taxonomies behind data clean rooms. And so InfoSum, born out of Europe, born out of GDPR and really born out of a company called DataSift. We essentially keyed the term data collaboration about seven or so years ago. Being able to identify the ways in which you could run really complex computational actions and use cases across connected data sets that you did not have ownership or control over. And so we started that journey many, many years ago in the greater Europe region, working in the limitations of GDPR, figuring out ways for organizations to still use the intelligence that they and their partners have been able to harness across them being a B2C organization, working with their customers and consumers. And so we were a data collaboration platform for a very, very, very long time and still in Europe.

More broadly, we are seen as a data collaboration platform, less so a data clean room. Data clean room really, I think, started to be popular in the United States. And just like any other acronym or technology term in the US, more so than any other region, from what I found, needs to latch onto something, needs a specific terminology to create some level of understanding around, which was fine. And so we started to don the data clean room nomenclature, and started to help working with the IAB Tech Lab, working with MMA, ANA, all of these other industry organizations to figure out better ways to define the standards and practices of data clean room. So I would say that InfoSum is synonymous with data collaboration with privacy protection and data security. We take both of those things very, very seriously and treat them very separately between privacy and data security.

And I would say that we are one of the founders of the clean, protected data ecosystem and are really working hard to make sure that other actors coming into this space are using the right technology, using the right processes when they make the claim of being a data clean room and working very hard to make sure those standards stick and make sure that organizations who are investing their time, energy, and dollars into this new and very powerful technology, they’re investing in the right thing. They’re getting essentially the guarantee of non-movement of data or protection, end-to-end privacy, and complete data security because you can’t really mess with privacy. I like to say that a lot. I use a different expletive when I say that. But we are very, very serious about the protection and privacy of consumers and the data and the businesses that are responsible for that data. And so I would say we’re synonymous with all of those things, but we really need to drive the industry forward to understand a better definition and standards for the clean room industry.

Dustin Raney:
Thanks for mentioning that about how you guys grew up in Europe, really as a company. And it’s got to play to your advantage to a degree to have to meet business needs inside of GDPR in a much more regulated environment. And then before really bringing that to the US and offering, you have the same kind of capabilities when US is playing catch up there. Are you starting to see the same kind of adoption here that you saw in Europe before? What are the trends that you’re seeing there?

Devon DeBlasio:
Well, we are seeing adoption at individual, organization-by-organization basis. So because there is not a federal or national privacy law like there is more broadly in Europe with GDPR, it’s not as great of a forcing function. And because in Europe I feel like they did not develop the same types of programmatic and automated practices we did in the United States because GDPR has been around for quite some time. And programmatic, I think caught on really, really quickly in America and grew really quickly. Outside of America, I feel like it wasn’t as fast a ramp-up period. And so there wasn’t enough time for other regions where privacy is more prevalent to have some of these baked-in processes, baked-in expectations to weed through and work your way through to break the mold. And that’s what we’re seeing in America was really breaking the mold of the traditional way or the known and agreed-upon way to buy and sell media, to use information about consumers, to deliver some level of an experience or expected, perceived advertisement for me as a customer or anyone else.

And so in Europe, yeah, we’ve been great for us in terms of an accelerator, as an expectation to have this non-movement of data or not having to sign a data sharing agreement. That’s very, very prevalent there. In America we’re starting to catch on. I would say really this year we started to see the terminology of data clean room, the interest in data collaboration, really start early in the year. And we’re starting to see a lot more RFIs come through. Over the fence, we’re seeing a lot more agencies and analysts and consultancy organizations who are responsible for driving education within their enterprise-level advertisers and publisher organizations to learn more about data clean rooms and data collaboration. So we’re starting to see it catch on more greatly. We’re seeing very large organizations announce that they have built a clean room initiative or built a privacy-first data strategy in-house.

And so it’s been great, but we have not yet seen the use cases run at scale, which is what we’re starting to see very, very slowly in America. And so people have maybe signed up for a data clean room or testing multiple data clean rooms, which is fantastic. We recommend testing multiple solutions now before it’s too late down the road. But they’re not using it at scale. They’re not really testing the waters. They’re not really smashing the keyboard to really figure out what the limitations of some of these solutions are. And that’s what we are hoping to see over the next 18 months or so, specifically in the US as what we’ve seen in greater Europe.

Kyle Holloway:
Yeah. Do want to just geek out for a minute with all of the technology changes and impacts that are going on with movement out of on-premise more into the cloud and just a lot of the different aspects of what clean rooms could and may entail. Talk to us a little bit specifically around InfoSum. What does your tech stack look like and where does it fit within the broader ecosystem?

Devon DeBlasio:
Great question. Yeah. So we’re agnostic in terms of data, in terms of identity, in terms of cloud infrastructure. And so our goal essentially, and what we’re doing today, is that we can work with your data where it lives. So if your data is in AWS, if it’s in GCS, if it’s in Azure, if it’s in IBM, your data can remain there. And essentially what happens is we have the ability to have a data streaming process so we can stream your data into a file management stack. So it’s a UI of your data. And what you have to do in order to get your data ready for those, the listeners who don’t really understand the process, is that you have a bunch of data living in different silos or in different places. You may have a CDP that’s helping organize that. You may have another identity player that’s helping organize that or doing everything internally or with an agency.

But the goal is that you need to get your data to a point where it’s in a file, whatever a file that may be, and you essentially want to use that file to collaborate. So this set of data for these set of customers, I want to make sure that I’m connecting it to if I’m a brand to this particular set of publishers’ data sets so we can run intersection, so we can identify which of our customers we share. And then we can explore the attributes and behavioral characteristics so we can work together to devise some level of strategic campaign for targeting, for measurement, whatever. And so in order to get to that point, you have to do a lot of pre-processing. You have to do a lot of normalization. And we handle all of that. So we have a transformation engine that can stitch the data together, manipulate it in a way that it gets to a point where it’s exactly what you need for collaboration.

We have a normalization process, a global schema that we use so that every data set that resides within our platform, or is used within our platform that resides within the organization’s own infrastructure, can be connected immediately. There’s no back and forth connections that you have to do or conversations you have to do to get the data to the right place where it’s Nth matchable. It’s immediately matchable within our platform. And then we use what we call bunkers. So bunkers is a secure environment that houses only the mathematical representation of the data. It gets really in the weeds if you want to go down that road. But essentially what we do is we take all of the PII, all of the keys. We encrypt it, we remove all the actual identifiers themselves, a completely pseudonymous view of that data set. All the attributes remain in tact because that’s what you want to see in terms of the characteristics of those individuals or households. But none of the actual keys, none of the identifying information is ever available or ever seeable.

And what we do is that every partner that has a data set that wants to collaborate, they have their data in a bunker. And then you say, “Hey. I want permission. I want to send a permission request to your data set so we can work together.” Great. You grant the permissions and then you can do a match and you can have 4, 5, 6, 10 different parties working together, all in their own bunkered environment, essentially working together, all permissioned. And essentially you can generate a match. The match happens instantaneously. You just essentially drag and drop which data sets you want to collaborate with. You get a match rate of what the overlap of your customer base looks like. You can explore the attributes and behavioral characteristics of that overlap.

If you need to expand that overlap with an identity partner like an Acxiom or someone else, you bring the identity partner in. If you don’t have a common key, they create the common key or they expand the match rate, increasing the match rates. You have an enrichment partner inside of that environment, you can bring another secondary data partner into the party, and they can essentially add additional characteristics about those consumers. And then you essentially build your audience. You can build your audience on our platform. We have audience-building capabilities similar to other organizational processes and platforms out there. Build your audience, build your segments, activate those.

And then same thing happens for measurement, working with measurement partners, ad servers, DSPs, SSPs. Same process. Everyone’s got bunker data, pulling the data in, connecting the data together without sharing the data, generating some level of an analysis with an incremental reach, overlap, frequency, attribution, and then everyone can essentially extract the insights without actually sharing any of the underlying data set. Through all of that process, there’s differential privacy applied. There’s noise, redaction, rounding applied, federated learning applied, all the cool secret sauce stuff that allows you to do all of this without actually ever exposing the underlying data set itself.

Dustin Raney:
And that’s what you call geeking out on data.

Devon DeBlasio:
So there’s this encryption process. There’s the streaming process. There’s the data, the sketch that we’re creating of the data set. There’s differential privacy. There’s multiple layers of our privacy tech that we’re using to protect the data over and over and over again. A lot of those have patents associated with them as well. So we can go into InfoSum-specific nerdy secret sauce stuff, or we can talk more broadly about what characteristics clean rooms operate in holistically. Happy to go either way.

Dustin Raney:
Honestly, Devon, I think the way I would like to see this go, maybe a slight right turn into use case. So you mentioned several use cases. Your conversation just now talked about overlya analysis across multiple parties. You talked about activation. You talked about attribution and measurement. One area that the entire industry is scrambling to figure out when cookies go away, and I think clean room’s inserting itself into this use case in a unique way. The question that we often hear, and I actually have myself, is the monetization part of that. How does it work? Who gets charged between an advertiser and a publisher? How are you seeing these contracts happen in place of a typical bid stream transaction when you’re activating data? So can you give our listeners some insight there on how you see that happening?

Devon DeBlasio:
Yeah. Like I said before, in terms of being agnostic, so we don’t have media. We don’t have identity. We have to bring those parties into the process. And so let’s make it easy. Let’s talk about let’s say a four-party engagement. You have a brand, you have a publisher, you have an identity partner, and you have some sort of a data partner that’s there for enrichment. That can either be the identity partner or standalone.

And so what happens is usually the brand wants to identify audiences within the overlap of shared customers with the publisher. They want to make sure they’re getting the right information in front of the right users at the right time, the classic right ad, right time, right place, right person, all that stuff. And so what they do is each party has their data in a bunker, like I said earlier, and they generate the intersection. Throughout this process, at least with InfoSum, no party’s paying for any of the computational costs required to do any of the heavy lifting of the processing or the intersections or any of that.

And so what happens is you essentially have the brand leading the charge. They’re in there. They’re essentially either writing their own scripts or they’re using our UI to generate the overlap, building their audience, enriching that with an identity partner to make sure there’s a common key. Maybe the brand has only email addresses and the publisher only has phone numbers. The identity partner comes in, builds that bridge. They can make a match. Great. We have audiences in there being enriched by second-party data partner. And then what happens is an audience is created, and the audience is then pushed either directly to the publisher if they have their own buying environment or it’s put into the seat of either the SSP or the DSP, depending on the relationship between the publisher and the brand and how the brand wants to buy or bid that request.

And InfoSum’s not taking any dollars and cents from anyone involved in that process. What would happen is we have multiple party agreements where the identity partners essentially gets paid by the brand using the identity graph outside of our platform. Second-party data partner also gets essentially paid by the brand or the publisher, depending on the relationship outside of our platform. That’s happened in a separate commercial agreement. And then really, we’ve just facilitated the connectivity, the insights and the activation pipes, and then the brand and the publisher, essentially they would go through the normal supply midstream processes to generate the bid, make the bid request. And if they win the bid, they would essentially pay the bidder. And that would be routed then to the publisher like it would normally in any other practice.

The beauty in what we’re seeing in our platform is that we’re seeing a lot less of the programmatic ecosystem come into play because that’s still difficult in terms of all the different parties involved and making sure you’re having a clean execution and a clean ad being served without a lot of other parties’ hands in the pie. And so we’re seeing a lot of direct publisher-to-brand relationships where the brand is buying directly with the publisher. The dollars and cents are going right to the publisher. The publisher can essentially offer up a higher value inventory set that’s either built custom for that brand, which means that would have a higher CPM. Most likely they can affix a higher dollar figure because it has greater interest to their particular brand. And the Brand then essentially is going to get higher ROAS and ROI out of that because it’s made custom for their customers, custom for their specific users they’re trying to get in front of.

And we see that commercial agreement really happen happened directly between the brand and the publisher, kind of a callback to the old school days of just calling up someone on the phone saying, “Hey. I want to make a placement on your newspaper and let’s get it going. Here’s my dollars and cents directly to you. I want this ad specifically on this time or day.” It’s similar now where we’re actually having those brands and publishers actually interface with each other. Less people are getting involved in between that process, which means more dollars and cents are going to the publisher and more value is coming back to the advertiser. And it should be cheaper for the advertiser at the end of the day as well.

Kyle Holloway:
So are you seeing a particular profile around the publishers? What types of publishers are coming on board and what’s the adoption of that?

Devon DeBlasio:
That’s a great question. So publishers who have good amount of first-party data. We’re talking about first-party deterministic matches in most cases here. We can do probabilistic matches as well. We can talk about that or table that for another discussion. But most likely it’s going to be a publisher that has a good amount of first-party data that they can match with the advertiser’s first-party data. And so typically, we’re seeing a lot of CTV publishers. We’re seeing streaming audio. We’re seeing gaming as a very, very large piece of the pie there, and retail media networks, but also any blank media network that you’re seeing. Entertainment media networks, travel media networks. These organizations that are essentially trying to build their own ecosystem to compete with the larger world gardens that are out there.

They’ve gone down that road thus far because they knew that cookies eventually were going to be completely decimated, we thought next year, whatever. We don’t have to touch that elephant in the room. But they have started to go on down the path. And they’re just continuing to chug along in terms of, “Hey. I have first-party data about my subscribers, about my customers. I want to be able to monetize that safely and securely with full control, with full visibility and transparency so I can allow an advertiser to work with me. I can build a custom set of my inventory that matches their specific set of users, but I can also identify other partners, whether it’s in another space or a retail media network working with a CTV or an MVPD organization, and essentially allowing inventory to flow through those lines of site as well.”

So we’re starting to see that effect really catch on in terms of the ecosystem, but also broadcasters. iTV and Channel 4 in the UK have been one of our biggest sets of clients in terms of their ability to say, “Hey. I don’t want to work with advertisers anymore who have first-party data unless you use InfoSum as the data collaboration platform. I don’t want to touch your data. You have to go through this neutral intermediary. You essentially bunker your data. We bunker our data. Then the advertiser has complete control over how their data’s managed, how the data’s permissioned, and they can work with our set of really robust audiences across our inventory. And then we can build custom audiences for them.”

So they’ve built their own products around our service to facilitate really, really great engagement, really powerful ads that can be served across their broadcast and streaming networks with full control over the privacy and security of the data set. So that’s been another huge area of value and growth for us over the past few years.

Dustin Raney:
Devon, one of the areas that Acxiom and InfoSum were strategically coming in together and coming up some kind of innovative new ways of addressing some of these problems, especially as it relates to reach, you get rid of the elephant in the room, the cookie deprecatation and all that. Well, you got to replace that with something. So we have some exciting announcements that are going to be hitting the press before too long around match multiplier, our ability to augment that match between the advertiser and the publisher, or maybe two advertisers doing an overlay analysis or anything like that. Do you see this type of capability really being a key part of InfoSum’s future value prop? Are you guys building out an app marketplace? What does that look like? And if you don’t mind, just talk a little bit more about the middle layer, the augmentation layer.

Devon DeBlasio:
I’d love to. Yeah. It’s a hugely important component of what we do. Again, like I said, we don’t have the data. We don’t have the media. We’re purely the infrastructure and the privacy-enhancing technology. And so we have to rely on very, very strategic partnerships like us with Acxiom. And again, I don’t know when this podcast goes live, but maybe it will be after the announcement or before the announcement. But we have something very exciting on the horizon that we’re essentially pushing out into the ecosystem that any essentially organization that currently is working with InfoSum or wants to work InfoSum now will have access to the great identity and expansion solutions that Acxiom and others have to offer.

And it’s extremely crucial because many organizations have first-party data. Great, but they don’t have every single key attached to those individuals. And nor should they. They collect the information they need to do what they need to do to drive experiences and positive experiences for their customers and consumers. And we can rely on bringing other parties to the table to help them extend the value of that data. So identity partners like Acxiom does a great job in terms of saying, “Hey. We can essentially safely give you the keys that will allow you to essentially create a bridge and a match with this other party,” whether it’s brand-to-brand, brand-to-publisher, multitude of those that can be involved, just so you can actually have a conversation. It’s almost providing a Rosetta stone for identity inside of the InfoSum walls, allowing these different organizations to speak the same language.

Once they’re speaking the same language, they can strategically drive some level of value for themselves as well as for the customers and consumers at the end of the day. And so it’s extremely impactful and important for us to have that middle layer, the strategic partner and channel partnerships sitting within our platform. So identity partners, measurement partners, data partners, all sitting there ready to go and ready and willing to provide value to those organizations who want to be able to work more seamlessly together.

And we wanted to make it function very similar to how they’re used to today. So every organization has their own bespoke data stack with multitude of three-letter acronyms that they’ve essentially invested in over time. A lot of those won’t be valid once addressability is no longer viable because the cookie went away, where privacy legislation has become to a point where you just aren’t willing to share that data. And so we’re building a net-new ecosystem, a non-sharing, a non-movement of data ecosystem on the backs of Acxiom and other identity partners and other data partners and measurement partners, all sitting within our platform.

And now with the launch of platform Sigma that we just announced in late July, we’re making it even easier and more fluid for organizations like Acxiom to build on top of our infrastructure, to build on top of our APIs. And that allows for apps to be built inside of our app library using our app exchange. So developers can get their hands “dirty” in a clean room. I don’t know how you want to slice and dice that statement. But what we’re saying is that developers can have greater access and greater control and flexibility over how they’re building applications inside our framework, inside the clean room. So if an organization wants to build an app specifically for them and their partners, they can go ahead and do that.

Acxiom can go ahead and build an application in our platform that’s available to everyone or only a certain set of users. That allows our partners to monetize their services and solutions inside of our platform in a safe and controlled way, but also other organizations to integrate clean room technology and data collaboration technology more fluidly across their tech stack. So they can easily start using privacy-enhancing technology without feeling like they’re missing something or they have to relinquish something, or they’re going to miss a gap in their tech stack. And so that’s what we’ve done with Sigma opening up our platform, allowing organizations to build on top of it. And hopefully, Acxiom over the next coming months is able to build really cool stuff for their clients and our shared clients within the platform itself.

Kyle Holloway:
What I’m hearing is it’s the aspect of bringing the app to the data versus bringing the data to the app, right?

Devon DeBlasio:
100%.

Kyle Holloway:
So you guys are enabling a platform where brands can maintain the security and the control of their data, but then have apps available that can leverage that data and execute in these collaborative fashions. So it’s really cool. I like that net new ecosystem description there.

Devon DeBlasio:
And agencies are a huge part of it as well. And I think agencies have really started to join the fight in terms of, “Hey. We know we need to be the bastions of good, the bastions of privacy-centricity.” And we’re seeing a lot of agencies, ones that have their own identity spines, ones that have their own ecosystems also build on top of our platform because a lot of times, the brand and the publishers aren’t doing the work themselves. They’re relying on a trusted intermediary or a trusted partner like an agency to do some of that work. So the more that we can build in a transparent manner of, “Hey. This organization’s building an application for you, but they’re doing it within a neutral, agnostic, decentralized platform like a clean room,” you know it’s safe and secure. You know you can use it with confidence versus the old ways of doing things where things were murky and not transparent. And maybe they were building things that necessarily an organization wasn’t totally aware of.

Dustin Raney:
Well, Devon, it’s been an amazing podcast. Can’t believe we’re already hitting the 30-minute mark here. Time flies when you’re having fun. And one, it’s always great to have another comrade in the rebel alliances, as Kyle likes to talk about it on data for good and just to do the right thing, right?

Devon DeBlasio:
Yes.

Dustin Raney:
We’re in a very important place inside the context of data-driven marketing and identity. So it’s always good to talk to people who take that very seriously like yourself. We do like to wrap up with a standard question. Loved your intro and going to love to hear your response here. It’s the year 2035. What does consumer engagement look like? Do you see a scarier, Orwellian-type future, or do you see brighter days ahead?

Devon DeBlasio:
That is a great question. I wish I had a canned answer that I could just pull out. But man, I want to hope that it’s a more consented environment. So there’s multiple types of consumers out there that people want the creepy experience because they don’t see it to being creepy. They just want super engagement, super customized experiences. And they’re going to give all their data up for that experience. And then there’s the people who are like, “Hey. I wear a tinfoil hat. I don’t want you touching my data. I want no experience whatsoever. I’m going to use DuckDuckGo all the time, whatever you want to call it. And I think I’m hoping to see there’s a more tailored experience and there’s a baseline offering for all types of consumers out there. I think the baseline today is that we’re going to collect your data really without your consent.

Yeah. We’re going to send you some consent flags and there’s certain information that you’re going to be able to control how it’s collected and not. But I think it’s still very opaque. It’s not regulated at a national level. And so I’m hoping in 2035, each one of us has our own consent management platform baked into every single piece of technology that we own. We can control how our data essentially is monetized, that we get a slice of that information. “Hey. I don’t want to pay for my cell phone bill anymore. As long as you collect the data and you can use that data anyway which you want, I don’t want to have to pay for cell service ever again.” Stuff like that. I think zero-party data is a little bit of a joke. I think there’s a way for us to enhance what can be done in terms of giving consumers greater control.

And I think advertising should reflect that. And I think that we should have a baseline awareness campaign that can go out. Everyone can see it. Great, fantastic. But there’s areas in which a consumer can raise their hand and become more involved in how they’re bespoking their experiences, similar to I think how gaming is going to revolutionize the future of advertising. I think that’s going to be the first area of where we’re going to start to see customers and consumers be more in control of their ad experiences day in and day out. But 2035, that’s a ways away even though it’s not really.

Dustin Raney:
Devon, thank you so much for joining us today. And for our listeners, you can find all of our Real Identity podcast episodes at acxiom.com/realtalk. And you can also catch Devon hosting the InfoSum podcast, Identity Architects. The link is included in our show notes. So go out and take a look at that. Man, I know that we received a lot from this today, Devon. I definitely want to have you back. Thank you so much. And look forward to doing this again sometime soon.

Devon DeBlasio:
Very, very honored to be here. Thank you so much for your time and looking forward to see how this turns out. Thank you.

Dustin Raney:
All right. Well, thanks everyone. We’ll catch you next time.