Podcast: ft. Rob Armstrong - Cookieless: Part 1 - The History of The Cookie

• Before digital came into play, people were just aggregating properties/ad real estate to sell to companies. for instance, "We will run our swimming pool ad in the lifestyle section of every local paper because it's the summer time."
• Print ad specs were standardized.
• Over time companies realized if they used computers, ads could be more efficiently managed.
• Since its inception in 2009, real-time bidding (RTB) has become one of the most popular ways to purchase ad inventory online.
• As ad media moved from physical space to digital space and analytics became more accessible.
• The early 2010s were an exciting time of creativity - many new businesses emerged in the online ad space.
• Retargeting soon became a strategy and a shiny new object. Today most marketers have budget for retargeting.
• Audience innovation and lookalike models came next...powered by cookies.
• A cookie, created in the mid-90s, is a file with a few strings of information that gets dropped on your computer. It that allows websites to provide you a better experience. Without a cookie, a site doesn't know who you are. They help recognize you and tie your activity back to previous activity.
• The difference between 1st party and 3rd party cookies.
• Ad buying became as dynamic as the stock market - there's a ton of science and logic behind it.
• When you're on the cusp of innovation and are successful, it looks chaotic. The fact that it is chaotic is evidence you found something important.
• As part of a digital technology shift, you tend to push the limits of technology.
• Importance of free and open content in society. Can we save the open web?
• A sneak peak in topics discussed in Part 2 of this series.

Adam Kerpelman:

My theory is that there are the people that are legit gamers. They really like sitting down and playing a board game for a long time. Most people have fond memories from growing up with them and they think they like board games, but they don't actually.

Brian Jones:

Yeah. They like a 15 minute game of Sorry!.

Adam Kerpelman:

Right. They like Cards Against Humanity, right? When I say board game, I'm talking about strap in, it's going to be four hours of fun in my mind.

Brian Jones:

That's too long.

Rob Armstrong:

But TEN's nice, you can change it to an hour, an hour and a half.

Adam Kerpelman:

Exactly, it's like enough.

Brian Jones:

Too long.

Adam Kerpelman:

Hey, it's The Data-Driven Marketer, sponsored by NetWise. I'm Adam.

Brian Jones:

I'm Brian.

Rob Armstrong:

Hey, I'm Rob.

Adam Kerpelman:

Welcome back for another hang in the data basement. Thanks for joining us, everybody, and special thanks to our guest today, Rob Armstrong, that third voice you heard in the intro. We specifically went and got Rob for this one, to kick off a three, maybe more, part series that we're going to do, about the cookieless future of data driven marketing. From there I think-

Brian Jones:

Everybody's favorite topic.

Adam Kerpelman:

Whose favorite topic?

Brian Jones:

Everybody's.

Adam Kerpelman:

Well it seems like it lately. But yeah, I think maybe the place to start is just with a quick intro, Rob, from you, kind of your background. Why did we seek you out to talk about this stuff?

Rob Armstrong:

Sure. Yeah. There's quite a bit to unpack. Hopefully we can get it done in three sessions. So I'm a product person in the industry. I've been in product management for last 12 plus years. My first role was actually at the first ad tech company, which is older than Microsoft, and has since been acquired and part of Mediaocean, where we were trying to build a digital campaign management tool and got very much disrupted by programmatics, so we'll come back and talk about that today, and then I went into startup land. I worked at DoubleVerify and a couple other startups and, along the way, got the opportunity to co-found Bombora, where we took internet traffic, at that time, a relative commodity, and turned it into a lot of B2B, unique data for intent, and I'm now SVP of product at IOTA, which is an international mostly focused company, although we have a lot of business in the U.S. as well, making audience data available at great scale from data suppliers to platforms, and we also do a lot of onboarding of data as well for brands.

Brian Jones:

You said something interesting in there that I'm curious about, you said the first ad tech company. What do you consider ad tech? We're big in defining terms on this podcast right now, so everyone can know the thousands of terms in marketing.

Rob Armstrong:

So Donovan data systems was started in the sixties or seventies, we'll have to check Wikipedia up, but it was started before Microsoft as a green screen, mainframe driven technology for ad agencies to manage media buys, and really stepping into replace a fax machine really, and so DDS was very prominent for print and radio and TV ad buying, and grew to be over a thousand people.

Brian Jones:

Wow.

Rob Armstrong:

Yeah, and so I stumbled into the industry, the finance. I was in finance before, out of college, and everything was going to problematic places with Lehman Brothers going under, and Bear Stearns is the price of a cup of coffee as a stock, and so stumbled into this company called DDS, and it sounded exciting, digital media technology, but we quickly found out that the pace of change in digital was actually going at a very quick pace, and linear media planning was not necessarily the end all be all, or a very difficult thing to continue. Yeah, so learned a lot.

Brian Jones:

Neat.

Rob Armstrong:

[crosstalk 00:04:36] every people there.

Adam Kerpelman:

So-

Brian Jones:

So that was like the first digital purchasing of media, even though it was non digital media that was being bought? It was like a transaction system for print media and radio?

Rob Armstrong:

It was a transaction system yeah, and it was effectively... Basically you'd have a campaign, you'd have a budget, and then you would have placements that you would buy, directly with publishers or ad networks, and I remember I wrote... This was a company that was not necessarily agile yet. I wrote like a 60 page spec for how to do package buying, where you would buy one line item from a network that would have a bunch of different placements and creative sizes, and so all this was in a linear workflow where you define it up front, you have a budget, it runs for a period of time, and this type of transaction, real time bidding, was very new, novel, and it just grew so quickly, and disrupted... At the time, there were a lot of ad networks, and the ad networks, just what they did was they helped arbitrage and consolidate supply.

Rob Armstrong:

If I'm planning TV, I know, relatively speaking, the TV channels that I would buy from, is you think about buying ads from websites, I mean, there's hundreds of thousands, if not millions, where you could have your ads, so. Ad networks came up to help consolidate different types of sites together have very large audiences, so you could call one person and make a pretty big buy, and that's kind of the realm that we were building this iDesk product into, when all of a sudden real time bidding came up and really changed everything.

Adam Kerpelman:

You said a funny thing early on in there that's one of those terms that I was like, "Oh yeah, that did mean that before it was co-opted by the other thing." Green screen. You say that now, and everybody thinks it's the way that you put a funny background behind you on zoom, but it used to mean literally the monitor that was green, that our like DOS equivalent showed up on so that you could talk to your mainframe, or terminal is another term, right? But now even terminal is just a program that we use to just-

Rob Armstrong:

It's just a color scheme.

Adam Kerpelman:

It's just a bunch of things, it's not... But it's named for this thing that used to be... Anyway.

Brian Jones:

No, that's-

Rob Armstrong:

No you're right. There was a mainframe, it ran in the basement, and even our technology, which we were building in Java, had to rely on certain data fields and components from the mainframe.

Brian Jones:

Yep. Even in the seventies, we were hanging out in the data basement. I like it.

Adam Kerpelman:

Data basement.

Rob Armstrong:

That's right.

Brian Jones:

We have like a...

Adam Kerpelman:

Data lived in the basement. So yeah. I feel like maybe the place to even back it up to reframe the whole thing, and then we can kinda walk in order up to where we are now, is to maybe even cut out the digital piece, right? You were talking about ad networks and linear buying and all that kind of stuff, and the next step is how digital starts to eat it, which I think you've already alluded to, is not necessarily the way that everybody thinks. But what did we mean by advertising before all this digital stuff started, it come online?

Rob Armstrong:

Yeah that context is really... That's a good point, Adam. I remember I was at DDS for two and a half years, and this is again, a company that has been focused on traditional media and building products to help facilitate that buying process for decades, and digital was very attractive but it was still very novel, and I remember at some point one of the executives celebrating the digital was now bigger than print in terms of spending and market share. So this was even before Facebook, and I mean Google had their search activity, that was still very large, or they had very large market share, but apart from Google search, there was no other really tech social platform concept either, and so the way that things evolved from there, it, as I said, with real time bidding, it just drove so much efficiency, that for every single ad impression, you can put a value on it, you can be very precise about whether or not you want to participate on it.

Rob Armstrong:

And then the other thing that real time bidding entered in, enter stage left, is AI, maybe one of our other favorite subjects to talk about, but the process of algorithms optimizing your campaign and making those decisions in real time, all these micro transactions happening, created so much efficiency, and one of the challenges with ad networks was the publisher was getting 50 percent-ish of what an ad network was receiving revenue, and what was the ad network doing other than just aggregating supply. So they weren't injecting a lot of value into the whole media product delivery other than aggregation. So once that was resolved, that then opened up a lot more investment that went into places like audiences, or places like ad verification, where or it took some time.

Adam Kerpelman:

That part of it's super interesting, because we think of ad networks in a certain way in our digital, sort of data driven space now. But it used to literally just be people aggregating properties, so it would be, "Hey, we've pulled together all of the local newspapers across the country, and if you want placements in all of them all at once, you can just come to us, pay for the ads to go to that space." But there's no... The value add was just, "We've done the legwork to pull that information together so that you don't have to go find each paper individually," for example, but there was no, like you said, there wasn't this value add piece. So the interesting thing that has happened over time is, like you said, first in this sort of green screen era, people went, "Hey, if we use computers instead of pieces of paper in a filing cabinet, we can more efficiently manage this network."

Adam Kerpelman:

But then like we're seeing, like we're constantly talking about on this podcast, once the software ate that, we eventually realized that we could do interesting things with it. More than just have a Rolodex that lives on a computer instead of a paper, right? And so, then the network started going, "Okay, well we can actually have a value add here, where we are calibrating this buying experience for the ads and stuff," so that it's not just literally like, "Okay, we're going to run in all of the local papers. We're going to run our milk ad in every local paper for this period of time, because it's the summer and we think that's a good time to run this ad, and that's it strategy done." Like...

Rob Armstrong:

That's right, and I think within... So this was 2009, 2010, by 2012 and 13, a few years into to real time bidding, there were such advanced tactics being used... It was pretty manual, but it was very effective. For example, looking at weather and saying, "If there's a lot of snow hitting the Northeast, I want to heavy up on Jeep ads, or winter tire ads," and guess what? Those performed really well.

Adam Kerpelman:

So that's a really great and really cool example, because we have no idea what the weather's going to be, right? You say, "Oh, it's summertime," and so... I think about my background in entertainment. You can plan your summertime movies, right? People want to go into theater, it's nice and cool, the kids are out of school. You work a cycle around the idea that Thanksgiving, Christmas, and summer are always stable, and that's the entire movie release schedule. But you couldn't suddenly be like, "You know what? We're going to drop Batman a month early because it's raining," so more people are going to stay inside and watch TV. I mean they kind of could now with streaming, but they're still probably not going to do it with movies because they're such hulking, expensive beasts.

Adam Kerpelman:

But I think the other interesting shift we're talking about here is... I keep imagining that scene from Zoolander, where they're like "The files are in the computer." It was a giant shift to move from computers making the buying experience for traditional advertising, print, TV, like you said, linear, more efficient, to the idea that the ads are actually on the computers, which is the world we live in now. The advertising we are talking about, certainly with the data driven approach, is on the computer, and so it's that funny shift where some of these companies fell off because they didn't have the foresight to be like, "Yeah but eventually the ads are just going to run on the computers, right?" And it's like, there's some set, always in these waves of evolution with these things, some subset of companies that go, "No no no no no," and they end up disappearing.

Rob Armstrong:

Sure, and it's fun to dwell with you on this time period, because while I lived through it I don't sit back often and reflect on it too deeply, and in our day to day and with our industry, it moves quickly, so we see things incrementally and we're definitely riding a big wave now of, and a bit of a roller coaster, about what will this world look like without cookies, and we have time to talk about that in the future. But I mean, at this time, it was quite novel and exciting that the industry had standards of what are the ad sizes that are available to buy, that there are standard ad... They all had names, like medium rectangle.

Brian Jones:

Totally remember those. Yeah.

Adam Kerpelman:

Right. Right.

Rob Armstrong:

Right? And that was a huge achievement, as an industry, that we got like standard sizes, which is fascinating, and the other thing here, which relates to the algorithms and also relates to the cookies situation is, with linear, one of the inherent challenges with it, apart from it being this all upfront effort to then deploy a lot of capital, once you deploy the capital, it's running, and really the only thing you have to do is like binary. I turn it off. I turn it on. With the realtime bidding components with measurement, there's a lot of feedback loops that happen every day, every hour, that optimize how I'm spending, and can create entirely new strategies spun off of what I thought would be the plan. Mike Tyson's "Everybody has a plan until they get punched in the face." I mean the ability to adjust and course correct is so important to media now, and so inherent to how campaigns are run and... Agile marketing, right? It didn't used to be that way.

Brian Jones:

How did you see, as this transition was happening, right as ad media was moving from a physical space into a digital space, and the analytics were really becoming more accessible and in higher volume and higher velocity, how did you see the businesses changing? Both the publishers, right, places where you can run ads, and companies that could do advertising? What was happening from a business side?

Rob Armstrong:

I would say there were a lot of very successful entrepreneurs who were pioneering during this time. With the new components it's like a brand new toolbox, kind of like the internet opening up. So I had the privilege of being at DoubleVerify for a year, where DoubleVerify was creating this ad verification product. There were actually three or four other companies that were in the same realm. So suddenly because now when you're buying against a set of users, or not talking to an actual ad network, or a direct site, you don't necessarily know where your ads are running, and so the whole idea of having and paying somebody to shine a light on where your ads are running, that seems so counterintuitive, because don't I know where I'm buying my stuff? Don't I know where it's going to run?

Rob Armstrong:

But you really don't, and so I remember coming up with the above the fold below the fold, I mean that was one of the things that verification shined a huge light on, was that actually when you buy ads, a lot of times they don't get seen because they're deeply embedded on the page, and so what are you really buying? And then, this whole idea of what is a viewable ad took several years for the industry to really define, so. I think it was a time of a lot of creativity, and on the audience side, this was a season where something like retargeting became really known as a very effective strategy. Maybe even more than it should have, given that you're already reaching people who have come to your site, but I mean the metrics on engagement on ads for those who have come to your site and you do retarget, I mean, it performs so much better than anything else, and so retargeting emerged as this new, shiny object, and now is effectively... I mean, most marketers have budget always for retargeting.

Adam Kerpelman:

Right. This is-

Brian Jones:

Yeah, and that touches on an interesting, another transition here, I think, where in traditional print media, radio, TV, there really was no way to know the individuals, right?

Rob Armstrong:

That's right.

Brian Jones:

We had analytics, we had metrics, we had Nielsen ratings, where some people would fill out Scantrons and stuff. But digital, all of a sudden, initially, also wasn't really looking at that, but very quickly we realized, no, there's deeper intelligence here by understanding who the person is behind the ad, who's viewing it and what their behavior is. How did you see that evolve? Because that's a direct... Retargeting is a real early manifestation of tracking, right? And identifying the consumer of the ad.

Rob Armstrong:

Yeah, and I think a lot of this started very innocently, truly, in the sense of, "Well, we're coming from TV. We can target age and gender, maybe, ish. We get that's like an index, and now, all of a sudden, I can go after auto intenders who have just recently signed up to get a test drive." How powerful of an audience is that? And that concept, any time over, in any type of vertical. So CPG, those who are buying shampoo products. I mean, it gets really granular, so. That created a lot of value, and it also created a lot of innovation on audiences. The other thing sort of separate, but also powered by cookies, is the measurement.

Rob Armstrong:

So those feedback loops are dependent upon being able to say this subset of users you've served the ad to, they have clicked, they have gone to the landing page, they have done other things on that landing page, and now you take some gross data of signals in AI, and now you can create lookalike models on that, so people that look very similar. So all of these things came about and they were very powerful. Now, the industry has not been regulated very much, and this is where we get into some of the downfalls of cookie or problems, but no one was really holding the industry back. So one of the things that became obvious, and was a very strong sort of line to draw, was we're not going to assign Brian Jones... We're not going to know his cookie and share that information. We're going to make cookies anonymous, but in the last few years, we've realized that, what is anonymous and how anonymous is that? And if you know that they are of a certain age range, and live in a certain zip code, and have this type of job title, well maybe you can identify them just based on the amount of information you have that, in and of itself, is anonymous, but combined together, could actually identify someone.

Adam Kerpelman:

So I think this at point we've finally got to the point where you said cookie, so we should back up a tiny bit to explain that, but I kind of have one thing I want to hit first, I think is interesting. So you mentioned your background before jumping into this stuff was financial. Do you think that gave you a better view, as you started to look at this in terms of bidding? And because that's kind of what happened with this ad stuff. When real time hit ad buying and all of that behavior, it suddenly became a market that was as dynamic as a stock market or a financial market, where it used to be a much slower sort of, different behavior. I wonder if your background, if it was easy to go, "Oh, once this happens, it becomes this instead of this, and it's actually going to look a lot more like..."

Rob Armstrong:

There was a real science, and logic, and math problem behind it. MediaMath had this math men branding of themselves for a number of years during the time that Mad Men was so popular on prime time TV. I think that allowed the industry to absorb a lot of great talent, and if you think about the industry spawning in New York city, where wall street is so dominant, as a tech person, first of all, if I'm an engineer, I got to wear a suit every day to my office. Second of all, if I mess up, and my code has a problem, it could be millions of dollars, if one record in the database gets corrupted. I go into this world of ad tech where there's billions of records, each of them is worth fractions of a penny. If one of them disappears, oh well, we have more, and so I think it appeals to a certain type of person.

Rob Armstrong:

I wouldn't say that finance specifically helped me, but what attracted my mind to finance was solving those deeper problems of logic and math, and that advertising afforded not only those same types of problems on a greater scale, in terms of tech so harder to solve, because there's so much more data to sift through. There's a lot more noise. But then also the culture of we have fun at work. Not as cutthroat, and I think I've really adored our industry for that, that we like to try to have fun, and they have this term frenemies, and that's really true. Even those we compete with, we would consider friends, and I think that's a real... A nice thing to have for work.

Adam Kerpelman:

The other thing I thought of, as you were talking about that. You mentioned the ad verification piece, just needing to verify that the ad ran. It's another one of those funny things that, as you get a shift in media, it's like, oh, this is a problem that we used to have with other forms of media, and it kind of... You ever hear about publishing companies? The reason they exist is because as radios started to pop up as a place to play music, we needed somebody whose job it was to literally send operatives to sit at a radio station all day and check a box if a song was played, so that they could collect royalties, and those publishing companies still exist, and still take a toll, even though it's all been digitized. So they don't have to pay for a workforce to go anymore. They just literally get a download of a Excel file from every radio station in the world, run a little math on it automatically, and then take 20% of all of music artists' money across the board, and that's BMI and ASCAP. That's what they do. It's ad verification.

Adam Kerpelman:

Which is funny, because those legacy things get to continue to exist if the function is still needed, and it is still needed, but slowly you're starting to see like Spotify push the publishers out, right? And it's kind of like the same thing with the ads, right? First there's this emergent need to have a verification service, and then it just becomes part of the technology, and then it's just built into the thing, and nobody really talks about... I mean they do, but it's an afterthought relative to the idea of, "We need verification now. Hey."

Rob Armstrong:

Right. It's a given and there's very clear standards now, and I think, Adam, what that also touches on is just sort of a general non-specific to media, just general startup, the startup experience of when you are on the cusp of innovation, and you are being very successful, when you're inside trying to build things, trying to answer customer questions, it looks very chaotic. It looks the opposite of success, because those problems are very real, but they're based upon the emerging discovered use cases of a client, and the fact that you're getting stretched so much is evidence of the fact that you've found something that is really really important.

Rob Armstrong:

But with any startup you have such limited resources to conquer all the things that you know you need to you today, and mentally it's just a very difficult thing to be in that I still struggle with today, of fortitude and resilience, and always making sure, you know, every day you try to make your best choices, but DoubleVerify at the time, our product was reporting, and our daily job to do reporting was taking more than 24 hours. So we had like a real crisis of just getting our basic, what customers were paying us for, done, let alone new features, let alone anything, and now DoubleVerify's multi-billion dollar company, right? So that was all very good sign, but at the time it did not feel that way.

Brian Jones:

Yeah. A lot of this has been emerging, and I guess this is a trend that will likely continue, maybe forever, with computers, but as part of a digitized technology shift, you tend to push the limits of the technology, right? So as our computers were emerging and digitizing things, and as they get faster and faster, the things we want to do are not doable quite yet by the machines, and so you're always at this point where computers are millions of times, maybe billions of times, faster than when we started, when we first had digitized computation, and we're still at the point where it's like I'm just on a call yesterday with the engineering team like, "Hey, the computers can't perform fast enough. How are we going to engineer around that?" Right?

Brian Jones:

And that's a big part of ad tech, and as we get into where it exists now, and how cookies come in, and how ad tracking and all this bidding happens, it's astonishing technology, right? It's some of the most sophisticated technology we've built as humans, right? The speed and the pace at which knowledge is going back and forth, and decisions are being made on these giant networks, so. It gets really fast paced and interesting, from a tech perspective right here.

Rob Armstrong:

And you think about an environment like that. I think that's why I, as an industry with there being so much value unlocked, transformation of how media's bought and sold, demand from... and real spend happening, real dollars floating in, to now where we are today, where we look back as an industry and we're like, "Man, we have not done a very good job of engaging consumers." We can serve them ads every day, we're really good at creating the ad product for those who spend on it, but this is a murky practice that is not simple for anybody to understand, and we've been really bad at engaging consumers in any form or fashion, and there are some use cases, the one that... I mean, as someone who's built a lot of the products that do tracking and do ad targeting, for a time I was quite, I don't know, and maybe it was a bit of denial about... And just knowing my own experience.

Rob Armstrong:

There was never any intention of nefarious activity. It was only whatever the audience is called, that's what we're trying to target. We're not really trying to go and identify anybody here, and what's so bad about getting an ad about cloud computing or an ad about a new car, if that's what you're genuinely interested in... But the story, I forget, I was doing some research about this, and the story of an expecting mom who has a miscarriage, and keeps getting ads about her upcoming child, and she can't do anything about it to turn off these ads. That is heartbreaking, and that is something that I think now, as an industry, we're working towards solving with the cookieless world that's coming, so.

Adam Kerpelman:

What-

Brian Jones:

Can you, as a bit of a tease for kind of the next episode where I think we're really going to go deep into, what are cookies, what is programmatic, can you describe the shift from how ads we talked into kind of digitizing here. What physically changed for advertisers from the time of like context driven advertising, or just still buying media on websites, to programmatic? What is programmatic for people?

Adam Kerpelman:

Yeah, I think... The thing I was going to say right before that, which gets to the same question is, you mentioned retargeting and it's a funny... it's a new emergent term from a thing that we were able to do. But in a way it's not any different. Eventually, we crossed this almost like uncanny line, where suddenly people are going, "Okay, that's weird." But we've been retargeting as a tactic forever. Retargeting just used to be, hey, the same kids that watch He-Man probably also watch GI Joe, and so we're going to run it on both places, and we're going to run it for long enough to make sure that the kids have all seen the ad for the toy that we're trying to sell them. Nobody's creeped out by that, and so it's interesting because with almost all these technologies, you have this transition to a thing where suddenly we go, "Wait a minute, maybe this needs to be regulated. Maybe it's too powerful now."

Adam Kerpelman:

Which yeah, I think gets to what is that transition technologically? And I think gets to the question that we want to end this episode with a bit of, which is what is the cookie, right? We're hearing about cookieless. We're hearing about the cookiepocalypse. We haven't really addressed it on the podcast yet. What is a cookie, other than a delicious morsel?

Rob Armstrong:

So it's an actual file. Yeah. It's a file. It's like a notepad with a few strings of information in it, that gets dropped onto your computer by somebody that you often don't know who they are, and what that does is... It was originally to allow websites to provide you a better experience, and that's still a very key use case of cookies, which is, I go to a site, I log in, I want them to remember me, or I leave something in the shopping cart, and I go back an hour later. So, in terms of state, without a cookie, the site doesn't know who you are when you return later. They don't know that you've already been through these pages. If you go to a new site and you've read three articles, if you go back in an hour, do you want to see those three articles in the prominent place? Or do you want to see the next best things that you are most likely to be interested in?

Rob Armstrong:

So that's what the cookie helps do, is it helps recognize and tie your activity back to previous activity. Now where cookies have... There's kind of a line drawn between first and third party, and don't ask me where second party went, because... What we call we first and third. So first is the site I'm on, so I'm browsing New York Times, there's a New York Times cookie, and they know what I've done on New York Times, but there'll also be other cookies that come from advertising, that will tie what I'm on also browsing The Drum, and looking at Car and Driver, and browsing other sites, so. That's where more data... Data is the new oil. Data has fueled the great value exchange and growth of programmatic, and this is why, as much as cookies are a problem, you see a company like The Trade Desk developing a universal ID is because they know what's at stake without cookies, and without a way to get to audiences, the open internet won't exist as we know it, and it'll all be closed, walled gardens like LinkedIn, Facebook, Twitter, et cetera, and they have a ton of data about your activity within those platforms, oftentimes much more than a cookie ever would, about those sites you browse, and so there's...

Rob Armstrong:

The problem with cookies is they never got regulated, as you said, and the other thing out them is, so if I have technology, and I'm an ad network, I'm trying to build a set of users or an audience system. I want a set of users interested in cars, and I have really great data about cars, and auto. If I want to exchange my data with a DSP, we both have to create a cookie on that user and exchange IDs in order to transact later and share information, so that my data can be used to target you. So at least two parties have to be creating cookies, and so in the world of ad tech, where we have many data sources, many platforms, that just creates an exponential floodgate of cookies on devices, and that now starts to drain your performance on pages, the loading times, and that creates... that has accelerated the angst about, "Who are these companies and what are they tracking? What are they doing?"

Adam Kerpelman:

Right. So fundamentally, the cookie is just a tiny bit of code that a site will drop on your device, so that, for example, you don't have to log in, like every time you leave a tab and then come back to it again, or something like that. But in this context, a bunch of different entities are dropping cookies, or looking at the other cookies, and then everything is kind of talking in this network that then adds up to, we're able to target ads really effectively. That's sort of the essence of programmatic advertising, right? We're able to know that, I don't know that you're Rob across the internet, but I can recognize that, oh, this anonymous identifier has, like you said, gone to this, a car website, a house website, a baby website, whatever, and then we can start to kind of put together a picture of your behavior and what you might be interested in.

Adam Kerpelman:

I think one of the things to call out, as we talk about the next piece, which is where are cookies going, is the part about this is important for the open web. Isn't just hyperbole that we're throwing out because we don't want cookies to go away? The problem with the walled gardens is if a Facebook, or whatever these walled gardens are, because they don't need to use the cookie to have that interconnected web of data, they just end up being a better place to run ads, and so if you kneecap the alternative, that's an open answer to that, all the ad money is just going to flow to those other places, because we can do better ads there and more effective advertising there, and so then the reason it's going to kill the open web is because everyone who's not that ecosystem is going to have more expensive, less effective ads, and so they're either going to have to charge you so that they can keep doing what they're doing, or I don't know what, right? This is why we need the other scanners for.

Brian Jones:

Yeah. And I want to be clear here too, because this is the part that'll be really interesting to get into, as well as the technology on the next call, but open web doesn't just mean some technology concept that only engineers care about. This is like all the websites that exist that aren't just Facebook, right? All the things that you share inside Facebook that Facebook's stealing the value of, and I'm choosing stealing because I don't really like Facebook and I would prefer the value go back out to publishers, that's going to be broken worse than it is now, and so there's huge implications for cookies, and I'm not arguing good or bad on cookies in any particular way, but how we evolve these technologies and how it impacts media, and information, and social discourse is hugely important here, so. This is-

Rob Armstrong:

Absolutely, and we are in the marketing camp, that's our day job. There's some very vocal folks about how important privacy is, and we're just trying to draw that line. I think a great irony is in Google's latest announcement of delaying cookie phase out, a big part of that, a big driver was actually a government entity out of the U.K. that has also been very focused on privacy. So wait, why would they want to delay cookies? It's because they recognize, Brian, what you were just talking about, which is the importance of free and open content for democracy, and for society.

Adam Kerpelman:

Yeah, right. That's kind of like, and the question of... So then the last question to hit before we get outta here, I think we already did pretty well, so the next episode, to tease that one, will be, okay, what are all the possible alternatives that ultimately save the open web if we can adopt them? But the last question is, where are cookies going, right? And some of it comes down to regulation, like you just alluded to, but then there's also this funny tension where regulatory bodies are actually fighting with one another, because one is saying, "Okay, we need privacy, and we need to do these things," quote privacy, but then there are antitrust ones that are saying, "But wait a minute, wait though. If we shut down... If you, like these other bodies, are going to Google and saying, yeah but if you pull the cookies, it's going to strengthen your monopoly on this behavior that happens because you have Chrome, so maybe don't do that. That might be a problem." And then Google goes, "Oh crap, you got us. We're going to punt for a year."

Rob Armstrong:

I think are going the way of the dodo. I think it's a matter of time. I think what comes on the heels of that, and with the industry and its history, and things we talked about before, I think there is going to be a dawn of innovation, like we haven't seen in a while, not only because of the size and magnitude of its disruption, but also because what we wind up with in the future is something that we can really build for, for a long time. I mean, having been in the industry for a while, it's been three or four years hearing that cookies will go away, so that is really blunted, I think, further advancement or investment in making them better, and we've been trying to figure out where are things going, and we're starting to see there's some first prototypes and signals of where things are going, but there's still a lot left to be defined, and I think it's a very exciting time for innovation. It doesn't feel like it yet, but it is coming, and so yeah, let's keep the conversation going.

Adam Kerpelman:

That's a perfect, perfect way... I don't even have to wrap this up as a host, because you just did it for us. Yeah, so everybody like, subscribe, stick around for the next one, that will drop eventually, which would be part two in our cookieless series, where we'll talk with Rob about some of the alternatives and things like that, that are starting to emerge, so that our listeners can go learn about those and be ready for that shift, once this cookieless shoe does drop. Otherwise, thanks everybody for joining us. This has been The Data-Driven Marketer, sponsored by NetWise. I'm Adam.

Brian Jones:

I'm Brian.

Rob Armstrong:

Nice to meet you, everybody. I'm Rob.

Brian Jones:

I love it. Take it easy, everybody.