The challenges of product management at scale: this week Paul Ford and Rich Ziade talk to Oren Mor, a head of product at Goldman Sachs and a former product manager at Google. They discuss his entry into the industry, making the technology behind Microsoft’s Kinect, his pivot to finance, and his return to tech at Google, where he spent years shipping ad products. They then go deep into the online advertising world, including ethical concerns around ad placements and the new taxonomies that’s creating on the web.
[Music plays alone for 15 seconds, ramps down]
Paul Ford Hi! You’re listening to Track Changes, the podcast of Postlight, a digital products studio at 101 5th Avenue in New York City. I’m Paul Ford. I’m the cofounder of Postlight and the cohost of this very podcast and I am joined by my fellow [music fades out] co-founder and co-host?
Rich Ziade Rich Ziade.
PF Rich, how you doin’?
RZ I’m doin’ very good.
PF You got a little cough.
RZ I got a little cough and it’s not—it’s weird it’s one of those it’s scratchy and I get coughing fits.
PF Give the listeners a taste.
RZ Nah, if I start I won’t stop. So—
PF Ok, you might hear that. [Rich clears his throat] Might have a little fun later—
RZ No. Tom will edit. Tom is our producer and he will edit it out.
PF That’s right. We can only hope. Rich, before we talk to our guest who I’m really excited to talk to, what [stammers] what does Postlight even do?
RZ Postlight is a digital products studio. But you could almost call it a digital product and platform studio. We build and design platforms and we build and design the products that run on top of them . . . apps, web experiences, mobile apps . . . internet of things! I just wanna say the internet of things for shits.
PF You know what it is? There’s just lots of transactions at the bottom of the things we do.
RZ Yeah, we can’t help but open the hood. That’s our thing.
PF Yeah, we like to—we like to think about that low level where the database is—a thousand things are happening a second.
PF Things are goin’ on.
RZ We actually—I wanna—can I continue to plug us?
PF Yeah, plug!
RZ We run a service that does nearly a hundred million requests a month and it’s free because we’re generous people.
PF That’s right. It’s called Postlight Mercury. It’ll turn any webpage into a clean, easily readable webpage.
RZ It turns it into data.
PF It actually powers a lot of the reading views all the way around the internet.
RZ Apps and stuff.
PF So if you go to mercury.postlight.com you can check that out.
RZ Enough about us!
PF Yeah, you know, so—This—this is a big subject we’re about to tackle. What are we even doin’ today?
RZ Well we’re gonna crack open a puzzle, and we’re gonna ask ourselves: what does one of the largest investment banks in the world have in common with one of the largest ad platforms in the world have in common with snowboarding in your living room?
PF I’m intrigued!
RZ Let’s do this.
PF So, we like to talk to people from very small businesses who are doing—
PF Cozy, little, tiny companies which is why today—
RZ We—I was walking by this little fudge shop in Newport, Rhode Island—
PF You said, “Hey! Hey, that guy looks interesting.” [Rich chuckles] And it turns out that the guy in the fudge shop was the Head of Product for Goldman Sachs.
RZ Oh, yoy, yoy. That’s—
PF We should disclose fully that Goldman Sachs is one of our clients and we don’t actually—we have a very clear policy of not advertising, you know, with the content of this show and not [right] sort of—this isn’t part of client service but this—
RZ But dude, if you’re gonna namedrop, yo.
PF Yeah. This is a smart person. His name is Oren Mor. Oren, welcome.
Oren Mor Hi, thank you.
PF Where do you even start? So, first of all—
RZ Why is Oren here? That’s, I think—
PF Well I think you can answer that.
RZ Well I did the thing where you go into incognito and then go to LinkedIn because I don’t want the other person to know that I looked at them on LinkedIn which I don’t know if that even works. I think it works.
PF This is the wrong place to ask.
OM I think you need to pay something.
RZ I think you have to pay something, that’s right. And I looked up Oren, I said, “This is an interesting person.” We had a great couple of meetings and his resume was fascinating. And we’re gonna get into it. And then we’re gonna get into a particular part of the world . . . on the web that’s talked about and really is a big puzzle today.
PF Well I think it’s worth noting: we talk a lot about product, right? Like shipping product and the kind—building products and platforms and so on. And if you were to look at someone’s LinkedIn profile and be like, “Well, that’s the real deal! That’s a real product person.” This is someone who’s been through—spent a lot of time at Google, other places as well. So, Oren tell us a little bit. So you’re a Head of Product at Goldman Sachs. What do you do all day?
OM So I’ve been at Goldman for seven months so I can’t claim to be a veteran of the company. Goldman is going through—like many other companies—through a—a transformation to become a more product oriented company, using technology not just to solve problems but to also service clients. And that is a big mindset change for a lot of people in a company that had—that’s been around for more than—over 150 years and needs to now become modernized in many ways. So, what we do is look at all the existing services and the way financial products are being communicated, distributed, and serviced with various types of clients and make sure that we build the right type of experience to those clients. Some of them are very sophisticated; some of them are less sophisticated and we have to make sure that they get the right level of transparency and service through technology.
PF So what—what kind of things do you build?
OM Just an example, we’re building a product. It’s a way for clients to trade directly through Goldman Sachs, so clients who wanna trade effects that’s foreign exchange. They wanna convert dollars to yen or any other currency, giving them a real time view of prices; letting them enter sophisticated ways of—of transacting between currencies and this is not—when you just go into a bank and convert money before your vacation. This could be large sums of money and you can move the market by doing certain things. So there’s a level of sophistication in just giving that level of transparency and being able to service that without creating any damage along the way. So it’s not as trivial as it sounds. There’s a lot of logic behind transacting and the types of sophistication and requirements that our clients have.
PF But at the end it’s sort of the regular, big, hard software problem, right? Like you’re building a tool that people will use every day.
OM Yeah. Look, it’s—Like everything in today’s world, software can solve it. I think this has the unique attributes of not just being a product challenge in terms of workflow and client services and all the things we come to expect now with any digital product but there’s also another level of sophistication when it comes to financial products. Those tend to be highly regulated; they have a lot of complexity; they have their set of requirements in terms of what to show and when; and addressing all these requirements and constraints while still maintaining a great client experience is a special challenge.
PF And so as Head of Product, you’re managing the team that’s building these things? Like what do you do?
OM I manage a team of Product Managers: great people who have the main knowledge in the financial products area but also usually were developers in the past or at least have a computer science background and understand software development to a large extent and can think through both the financial product aspects of what we deliver, and the product experience that our clients will have.
RZ So let’s rewind, and take you out of university [Oren chuckles], and you land where?
OM So that’s—that goes back a few years now. I am originally from Israel where everyone has to go to the military. Spent a few years serving in the military. I was a student and a soldier at the same time, and then I spent a few more years doing tech R&D for the Israeli military. After that I started a company called PrimeSense. You’re probably familiar with the technology that PrimeSense developed in the Microsoft Connect. It’s a 3D camera and gesture recognition—
RZ Right. I—I snowboarded in my [Oren chuckles] house [Oren laughs] using a Microsoft Connect.
OM You can fly; you can pet animals; you can fight dragons.
RZ I tried an exercise one and it didn’t work well.
RZ Anyway, clearly I see the dotted line to Goldman Sachs. The fact that you’ve gone from Microsoft Connect to Goldman Sachs, I think is deeply important. But we’re not—
OM [Laughing] And a few stops along the way—
PF Let’s take a few steps—right.
RZ Alright, so there’s gotta [chuckling] be something else in between. Alright, so, that became the Micro—which is—that’s an achievement. I mean that was a—at the time, really cool tech.
OM We had a lot of fun working on it. It’s I think is one of those rare technologies that people actually called magical which [sure] I still remember and makes me happy til today cuz it’s worth to do something that people—that have such an emotional reaction with people. That was a great experience and that experience also taught me that I know nothing about how business is run. I knew a lot about technology. And I knew about coding but I had no idea—the experience of raising money, starting a company, manag—scaling it up, managing people taught me that I need to really expand my business education which is why I went to Harvard Business School to get an MBA.
PF Sure, a natural next step, you know? You’re just like, “I don’t really know how this all works.”
RZ “I’m done with snowboarding in my living room!”
PF Anyway, enough making fun. Ok, so HBS: nice place, good school.
OM Good school; good people—
RZ Made friends.
OM Not good weather.
RZ Not good weather.
OM [Laughs] Yeah, and after that I really wanted to understand finance works so I ended up working for a family office doing investments for three years.
RZ Mm hmm.
PF We should explain to listeners what a family office. It’s a thing you might know about in New York City but not know elsewhere.
RZ You wanna give a quick, one second—
OM Sure! So, some families have the amount of wealth that requires someone to help them manage it. We should all have that fortunate problem but a family office is basically an investment company—usually a small one—that helps wealthy families allocate their funds in a smart way for next generations.
RZ Right. It’s large enough—the funds are significant enough that it deserves a set of people that are thinking about investment.
OM Yeah, there’s a point where, you know, the bank is—You know, you want more hand holding and [attention]—you want more attention and thought about what you’re doing.
PF You know there’s actually a—we’ll find the Wikipedia page and put it up with the podcast but like the Rockefellers actually—there’s like a floor or an office in Rockefeller Center and it’s known by the number, it’s like Room 801 or something like that.
RZ And it’s the family office.
PF It is the legendary—Rockefeller funds are managed there.
RZ Got it.
PF Ok, so—a little time in a family—So HBS, you get your business jobs and that’s pretty good training.
OM Yeah and then very, very fortunate to have had a chance to also learn about actual investing in a real investing environment. I started there because they wanted to do tech investing and this was 2010, the wonderful year where tech started the [stammers] early stage investing really started to bloom and a lot of things were happening but the type of investment DNA was very much towards things that are less risky. So I found myself doing a lot of distressed real estate [mm hmm] which, by the way, in 2010 was [stammers]—
RZ There was plenty of it! [Laughs]
OM There was a lot of it and the type of—and it was as adventurous as doing startups today. So that—I did that for a while and as that started winding down I realized that I missed technology; and I missed doing things that are innovative.
PF Right, your average billionaire grandpa doesn’t want you to throw your money at a machine-learning toaster startup and—
OM Well you know some of them I think it’s—some of them are, some of them less so, some of them have an allocation but it’s—it’s definitely for some folks it’s definitely a stretch [mm hmm] to invest in something that is—all the value is in the future prospects of someone’s idea and execution and not in some value that you can underwrite.
RZ Put a finger on.
PM Like real estate.
PF Sure. Well and rich people have done very well with real estate.
OM I hear—I hear it’s a good thing, yes.
PF [Rich chuckling] You get that—If you’re lookin’ at that hundred year time horizon, it’s—get that skyscraper—
RZ It’s good! It’s good.
RZ Yeah. Alright so, family office: three years, then?
OM Then—then going back to technology, I was contemplating between starting my own thing or working for a big company and I ended up after playing around for a few months with my own idea, I decided to work for a big company and started to work as a Product Manager at Google.
RZ Ok . . . and this is 2010-ish?
OM This is ‘13.
RZ ‘13, ok, so Google’s a very big company.
OM Google is a big, well-known company.
PF What did Google make of your private wealth period? Like I can see them being, like, “Oh great, the guy from the Connect! C’mon!”
OM I’m pretty sure I confused them in the interviewing process [Paul laughs]. I—I spoke fluent technology and what I did in my startup and the Israeli military is pretty much equivalent to product management work.
OM You know they saw that I knew what I was talking about but the investment chapter is—is definitely an outlier on that—
OM On my CV at that point.
RZ Ok, so—I’m gonna guess there are about 40 or 50 thousand Product Managers at Google.
OM No, actually not, so all of—today, Google is 75,000, at the time is what less than half. Google grew a lot in the last 20 years—
OM Around 30, 35 thousand total. Yeah.
PF And you’re all engineers. Like everyone at Google’s a software engineer at title, right?
OM That is a common misconception.
PF Oh! Please correct me.
OM Google is about—today Google is about 50/50.
OM Product Managers are considered engineers.
OM You have about 50% of the organization of actual Engineers, Product Managers, and then you have salespeople and legal, and support, and you have all the other functions that work alongside—
RZ Wow! That’s a big support side with sales side.
PF Well, I mean, it’s Google. Was your title Product Manager or was it something else?
OM Joined as a Product Manager [Product Manager, ok]. Progressed my career as a Product Manager.
PF And were you here in New York City or—
OM I joined in New York City. They have a wonderful office in Chelsea.
PF It’s vast! Every time I’ve gone into it, I have gone in the wrong entrance and then I’m a half hour late because it’s so big.
OM That’s part of the test, part of the test.
PF Well I failed the test many times!
RZ—over and over again.
PF That’s why I work at Postlight instead of Google. [Rich laughs] It’s such a big building that people get around my scooter. I’ve seen that happen there. So—
OM They still have it. I think it’s—they are starting to say, “Maybe you guys shouldn’t take the scooter.”
RZ Somebody got smashed into a wall.
OM Some people had some accidents.
PF The last time I was there it was so crowded that there was no way to do that. You can’t move around. Ok, so, Google: nice little company, another little startup and—
RZ What are you doing there?
OM I landed in a platform called Doubleclick, it’s an online advertising platform. It’s a business that Google acquired a few years back in 2008 and it was pretty big at the time but it grew to be even more successful with Google in the last few years as the main online, third party online platform for advertisers and publishers.
PF I mean for our listeners: when people say that Google is an advertising company what they’re talking about is Doubleclick. Like that’s the platform, right?
OM Well they have—Google has a few platfor—like ‘platform’ is a big word in Google. They have a few of them. They have—there’s a fundamental distinction between the search engine and display. So when you think about online advertising you have the search world where people advertise alongside keywords that people search, you know—when people search for things—the results have ads embedded in them.
PF So I search for coffee and then that text ad pops up and says, “Best coffee in the world!”
OM Right. And then on the other side is a display network which is very similar as to what you see on the billboards on the streets only on websites. You go on a website and you have advertisements for shoes and services.
PF That 300 by 250 rectangle.
OM The rectangles, the full pages, the pop ups. All types of wonderful creations.
PF So those all originated at Google. I mean the ads don’t but they—they are kinda hosted by Google.
OM Yeah so even here another distinction is there’s something called GDN, Google Display Network. Those are properties where people who have a website, let’s say you have a blog and you carve out a piece of that—of the page and say, “Google, please manage that for me. Manage that on my behalf.” That is something that then Google can control and serve against. That’s one type. That’s the GDN. Another type is when Google acts as a third party. That’s where Doubleclick comes into play. The New York Times—The New York Times is not gonna carve a piece of the website and say, “Google, it’s yours now.” They’re gonna put a piece of the website and say, “I will have a contract with Nike as an advertiser.” I know we don’t do advertisements but I’m giving you a few examples.
PF Mm hmm.
OM But Nike doesn’t want to now go around and make sure that the right ads on The New York Times are served for males in the New York area between 8am and 8pm and a different ad is served between 8pm and midnite.
PF Right. They make shoes.
OM They make shoes. And they don’t want to worry about aggregating all the advertising data. They don’t wanna worry about the slingshots. They’re just serving the ad to its destination. Making sure that it happened, making sure that some quality aspects are met. That’s why they get a third party, like Doubleclick, they upload their ads, they say—
OM They can say, “Go,” and in this world are two forms of ‘go’. There’s one called reservation and the other one is called programmatic. I don’t wanna get too technical but reservation is when you have a contract. The example with The New York Times and Nike. Two parties—
RZ They’ve agreed.
OM They agreed and now you just need to serve that contract. The parameters of the contract are placed into Google, into Doubleclick, alongside the ads and then Doubleclick just does the execution.
RZ Executes and takes a little cut.
OM Takes—charges a fee for serving and adds a lot of value through measuring demographics, measuring view-ability, will perhaps talk about what these terms mean. And the other side is called programmatic. That’s where you don’t have a contract. You have a [sic] aggregator of a lot of small sources of traffic. So blogs, small websites, websites that don’t want to—are not big enough to manage their reservations and actually sign contracts.
PF Nike’s never gonna call them up.
OM No one is gonna call them up.
OM No one’s gonna call, you know, someone’s blog and say—
OM Exactly. Exactly. So, Doubleclick—or that’s an entire ecosystem of companies that have exchanges and Doubleclick will buy on behalf of the user. They will look at exchanges. They will look at—we call it inventory. Inventory is basically when you go on a website and ads load—that takes a few milliseconds, what—the act of loading the ads, you’re creating an impression.
PF Mm hmm.
OM And your attributes—so, Google knows through your cookie profile that you are likely a male and they are connecting from the New York area. They will go to a—your impression will go to an exchange and they will say, “I have a male. He’s at this age group. And, by the way he’s really into sneakers. Who wants to bid on him?” And then, you know, a coffee company will bid something but Nike will bid higher because you’re their target market.
PF So it’s a marketplace.
OM It’s a marketplace. It happens in real time. And then once the bid is won, Doubleclick is responsible for—so Doubleclick is responsible for the bid, determining the right price to bid, and if—if they get it, they will serve the ad and they will measure a lot of things around interaction around the ad: if people clicked on it; if people engage with it; was it ever seen; was it ever displayed on the screen. What type of content is on the page. That’s where issues like brand safety may come up. So that’s—
RZ What’s interesting is a lot of people don’t know about the insights that Google [clears throat]—Google and others have just by you browsing the web. You know, they say, “I mean I went to a shoe site and I went to a couple of news sites and I saw a funny YouTube video. They don’t know I’m male but the insights are probably—most people sell short how incredibly powerful the insights are into location and also interests, right? Because you’re going to particular sites and the like. Which—that’s its own conversation. Now you mentioned brand safety. What does brand safety mean?
OM Brand safety started as a—as online advertising grew and people started to buy inventory or impressions on various properties and websites, there was a concern that, “I don’t know where I’m getting myself into.” So you have two sides of the equation: you have the publisher—so the sneakers blog and that says, “I don’t want inappropriate ads. I don’t want ads of content that doesn’t fit my brand image on my website.” That’s one side. The other side is from the advertisers, the advertisers say, “I don’t want to be against—I don’t wanna be—my content, my brand to be displayed inappropriate content.” And it started mainly with the concern of having ads on porn websites. That’s where the kind of [sure] the original requirement came from. You know, Disney does not wanna be displayed against porn sites. That’s the kind of extreme example. And there was a small ecosystem of companies that came to serve that need of advertisers. And they had different names—brand safety came later on. At the beginning the category came to be known as verification tools. If you think about online advertising in today’s world, I always think about it as the analogy to what the stock market was in the 19th century. A lot of the things that are clear today in the stock market, a certain criteria for how to be transparent around what you’re selling, how to conduct exchanges, the regulation around it, and the penalties if you deviate from those things are all very clear in the financial industry. They’re not as clear in the online industry—online advertising industry. So in the 19th century you could issue more stocks and you could do insider trading and it was ‘buyer beware’ mentality.
RZ It just wasn’t policed.
OM And today, I think, it’s getting better but the online advertising industry grew up with that mentality of ‘buyer beware’, and as the industry matures then the quality and the standards are increasing with it and brand safety in that area of verification tools are at the core of what is—what does it mean to have a good ad.
PF Well and I think it’s noting here, right? Like, Google is kind of a nation state and these are not small accounts. At this point. You’re talking tens, hundreds, millions, billions of dollars flowing through the system and so with that when the purchaser says, “I want accountability.” That’s a tremendous risk unless you have a good answer.
OM Right and look: [stammers] we saw in the last few months when some issues around fake news and hate content became more of a—became a topic. Advertisers became much more aware of the risk they bear when they advertise in certain channels. And that definitely creates a huge financial incentive for companies including Google to make sure that they have quality assurances for advertisers.
RZ And this is a problem that’s at a scale where you can’t just call your customer rep . . . to handle it in a reactive way, right? I mean it’s—it’s—
OM Right! So, it’s funny the industry started in exactly that way. The industry started by manually creating black lists. People would sit and say, “I don’t wanna be on that site, and on that site.” And create a list of websites and as you were bidding on inventory or as you were serving, there was the ability to block that. And not serve an ad or not bid on an ad in the first place which is the ideal scenario [mm hmm] but as the scale exploded, you couldn’t have people just looking manually at websites and say—
RZ There’s too many.
OM “This is good for me; this is not good for me.”
OM And machine learning algorithms started to be introduced into that space: looking at websites before an ad is displayed and categorizing that. There are a few companies that do that. Google has an internal product that is fantastic at looking at a website and determining what type of content is on that website. And that works! That works great when you’re talking about pornography or suggestive or violence or political or religion—there are certain types of categories that are easy—relatively, in today’s technology, relatively easy to detect.
RZ So what’s effectively happening is Google’s machines are reading over the website first and making sure that it fits a certain set—like it doesn’t—
OM It doesn’t violate what the client wanted.
RZ What the client wants.
OM So if the client said, “I don’t want religion.”
OM Every website that was categorized as religion is not gonna have ads from that specific advertiser on it.
RZ So that’s an interesting example . . . because we’ve been talking so far about offensive sites like that promote violence and pornography but religion—like I don’t really wanna take sides here, I just don’t wanna be on anything that has to do with religion.
OM Yeah! And it has no sides. It’s just about the content.
RZ Yeah, sure.
OM It’s about the type of content.
PF So, you know what’s interesting here, too, is that the earliest days of the web, right? It was all about—like Yahoo started as a categorization system for putting websites into a great big index. And that—that didn’t scale. Very quickly it turned out that you couldn’t actually categorize every website ever. And there have been lots of attempts to sort of organize things or get people to self organize. There was the big Semantic Web effort and things like that, like taxonomies would emerge. And what we’re finding here is that driven by the needs of the advertiser, a huge classification effort has happened. And Google’s able to point at something and say with some—I’m assuming like some statistically valid confidence what a site or what a page is about.
OM Correct. And it works in certain areas, right? So it works when you have—when you wanna look at pornogarphy or violence but it gets harder as you wanna get more specific. So violence, for example, is fantastic when you’re watching a James Bond trailer but [sure] it’s horrible when you’re watching an ISIS execution or other brutal examples [mm hmm] and now teaching an algorithm to say, “This type of violence is ok but this type of violence is horrible.”
RZ Very tricky!
OM Yeah. So that’s where the challenges lie now.
PF So how do you even begin to address a problem like that. What are your building blocks and components?
PF I mean machine learning in a very broad way.
OM So machine learning is the tool but you actually do have a pretty massive manual effort here in taking linguists in different languages to manually create the training sets in different languages. I mean, by the way, that’s not just a language barrier thing, it’s also a cultural thing. What is in one culture considered suggestive or erotic might be in another culture totally acceptable [sure]. So those nuances are captured through massive training sets that have to be continuously updated and then machine learning. A lot of machine learning.
RZ It’s a huge, huge endeavor.
OM It’s a huge and expensive endeavor.
RZ Does Google have thousands of people offshore that just focus on these things or how does—?”
OM Can’t tell you the number but it’s massive and expensive.
PF Actually all of the big internet companies. I don’t know about Apple necessarily but like Facebook does. Any major content driven, search powered play kind of has to deal with this stuff. I think it starts with basic standards around pornography and violent pornography and like, “We have to keep this stuff off the platform,” but then over time it becomes a tool connecting the advertiser to the person they want or assisting people in finding content and so on.
OM In general the quality—we use brand safety but the general idea of quality of ads is not just brand safety. We all know the websites that you land on the page and the page is swamped with ads, tiny ads [sure], and it’s not a good experience for the user; it’s not a good experience for the advertiser; it’s horrible for everyone but someone is making money in the process.
PF Mm hmm.
OM So a set—determining a set of qualities not just about the type of content but also around the density of ads; around where the ads are placed on a page. Many publishers found that if they stuff the bottom of the page with ads, they get paid for it . . . because—
PF Nobody scrolls down there but they are sitting there.
OM Nobody scrolls down but they get paid for someone placing an ad. But if you add the measure of view-ability which is a made up word used in the online advertising industry for just answering the simple question, “Was this ad ever presented on a screen?”
PF Mm hmm.
OM If you know that it was on screen: someone scrolled all the way down and had that on a screen for at least a second—
OM You’re already cutting off a huge amount of all of the online inventory. We’re talking about 50% of online ads.
PF Ooh! So that was a bad day, when the view-ability metric showed up.
OM Well it’s a good day for—for consumers.
PF Yes. Ok.
RZ So wait, this went down? Like they added view-ability—
OM I think the industry is in the process of adopting that and even—
RZ Ok so there’s resistance around this because obviously there’s dollars involved.
OM There are dollars involved; people are used to getting paid for something and that’s not going to continue. I think it’s clear for everyone that that’s the direction the industry is going but even in that direction the criteria around what view-ability is are, again, there’s no governing body for the industry so different companies invent their own set of criteria for what is a view-able impression.
RZ So this is still in flux? This is still being discussed and negotiated.
OM Yeah. [Stammers] There is a consortium of companies and they have a standard but some advertisers—some of the major ones in the industry came and said, “We’re not accepting this type of criteria. We don’t want one second, we want five seconds.”
RZ Mm hmm.
OM And on the hand you have companies that are effectively publishers like Facebook who say, “When you look at Facebook, view-ability for a video is three seconds.” So if you scroll slowly and a video starts playing, someone pays for it, even if you’re not watching the ad. The fact that it was on the screen for three seconds.
OM And that’s fine, that’s—and YouTube has a different business model with Trueview where if you don’t skip the ad, the ad counts, which is another way of dealing with viewability issues.
RZ If you don’t skip it, like usually it’s the five second countdown.
OM You have the five seconds, if you skipped it in the five seconds there’s no transaction. No one pays.
RZ No but wait, you can’t skip it.
OM No, sorry you have—but then you can skip. If you skipped in a certain window of time then the advertiser doesn’t pay but beyond that, the advertiser needs to pay for the ad.
RZ So if you stay an extra three seconds, something like that.
OM I don’t remember the exact number.
RZ Yeah, it’s all a ga—I mean this is sport, right? Like this is essentially a game where there’s a lot at stake, right? There’s a lot of dollard involved and you have these different actors with different interests that are playing around with us, essentially. And I think what’s interesting here is the lack of regulation. Usually this is where some governing body, whether it be some commission or something steps in because that governing body is not tied to commercial interests, right?
OM So the—the industry does have an entity—I’m not even sure how—like what’s the legal stance of that entity that created recommendation for standards but—
RZ But it’s—it’s not enforced.
OM It’s—there’s no—there’s no teeth behind it.
RZ Right. Exactly. Exactly. It’s like international courts [laughs].
PF Well it’s a standard, right? Like—
OM It’s a standard.
RZ Yeah but that’s—that’s nonsense.
PF No, no, but that—There’s no legal force behind it.
OM I think in this case the commercial forces are going to push towards convergence on standards towards something that’s gonna by—
RZ Agreed upon and—ok.
OM—widely accepted. Yeah. So, you know, I agree that sometimes a regulator are [sic] a good thing, maybe in this case it’s gonna happen by sheer market forces [yeah]. We’ll wait and see.
RZ Yeah, we should talk about another phenomena that’s kinda related to this which is relevant recently . . . and that’s sort of these protest movements that kick in where someone doesn’t like the content of a site or—
PF Oh, you know, there’s one called Sleeping Giants which is—They’re very opposed to Breitbart. And so what they’ve been doing is asking people to tweet about advertisers they see on Breitbart who probably don’t wanna be there. Big auto brands, things like that. And—
RZ It’s like, “Toyota! Why are you next to this horrible headline?”
OM Well that’s actually something that’s simple—that simple to solve because you can still use a blacklist. You can still say, “This particular URL or website is something I don’t wanna be on ever.”
OM And that’s just gonna be blocked.
PF So what they’re doing as an activist organization is teaching people how to get on to—Like, no one knows where their ads are going, you know, and so teaching people how to get on that blacklist, or to create that blacklist for themselves. You know, there’s—it’s a kind of abstract thing but I wanna bring it back a little bit. So, here’s this big platform, Doubleclick, and sort of the overall Google ad platform. You are a Product Manager but your actual job is to introduce friction into this process, so that not everyone can have every ad appear on every website. Like you are coming in here and saying, “Wait, we’re gonna create some—- There’s gonna be some gateways, some digital gateways.” What’s it like being the Product Manager of something that slows the whole system down?”
OM Um [hesitates] good point. Generally when you—when you introduce constraints to a system people—you know, that could create friction—
PF Mm hmm.
OM I think Google has done an amazing job at really understanding that the quality of the advertiser and the quality of the experience of the end user are critical and to be willing to get—
RZ Mm. That’s a long gameview right?
OM It’s a long game.
RZ Because the actions you’re doing are gonna actually, in the nearterm, equal less money.
OM Right. It’s on viewability; it’s on brand safety; it’s on fraud. Fraud is massive in online advertising and Google—just every time fraud is detected, it’s removed and people don’t pay for it which is, again: you’re losing money on the short term [sure] but you’re doing the right thing for the long term. I—you know, not every company has the deep pockets and patience to take that approach [sure] but—
RZ But they see long term value.
RZ And you know, that’s what’s at stake here.
PF Well and they’re doin’ ok. You know, if they take a little outta there—
PF They’re still gonna be able to make it fine in the next year.
RZ Well I think they also recognize themselves as an important actor, not just as, “Oh, we’re a company.” Like Google is a significant part of the piping here. Like that’s—that’s real and I think probably also they are very closely scrutinized as an organization [well it’s better for them—] that can be penalized—
PF It’s better for them to get to it before the government does.
RZ Exactly, exactly.
PF Without a doubt. Yeah. Absolutely.
PF I mean that’s just—giant orgs at that scale that’s the kind—that’s one of the motivations is that [right], “You know, I don’t want antitrust and I don’t want the FCC and I don’t—I wanna just get my job done,” and you know, if they do it they get to stand up and say, “Because we care so much about the integrity of your brands—”
RZ They’re getting ahead of it.
PF So it makes a lot of business sense when you’re at that scale to be just like incredibly paranoid, right?
RZ Sure, sure.
PF Ok so you—you were able to—you were shipping these sort of like almost gateways or filters in the whole system.
OM So that, you know, just defining—And again that goes to the part where people are not aware of all the problems that—even sophisticated advertisers don’t—you know, they think a lot about the demographics and they think a lot about their campaigns but the amount of thought given to exposure to fraud when you buy on different types of exchanges; the exposure to low quality of viewability; the exposure to brand issues—like all that is something that requires education and training alongside the technological solutions. So, thinking through the product, not just from, you know, the technology but through, “How do we get this out there? How do we—how do we get the word out through whitepapers and attending conferences and meeting with big clients and explaining that—and having—and building easy to use products?” All that together is—that entire package is the role of the Product Manager, not just defining the technology.
PF So every—every time I view an ad, it’s—the outcome—first of all, it’s the direct outcome of millions of lines of executing code all across the internet, just for me to see that rectangle on the screen.
OM The amount—the amount of logic that’s behind—and that pops on your screen is incredible.
PF And it’s simultaneously the product of a vast amount of communication between clients and ad creators and publishers and Google.
OM An entire ecosystem of people who profile users.
PF Mm hmm.
OM Not necessarily Google, there are third party companies that do that, an exchange that connects sellers and buyers, a bidding system, a serving system, and then behind the scenes a reporting system that collects the data measured out of your page and aggregates the data for the advertiser.
PF And all of that for that one in a thousand chance that I’m gonna click.
RZ Just so a donut can roll across the bottom of my article.
PF That’s exactly right.
OM Makes you appreciate the donut.
RZ I gu—I’m gonna stop for a minute and say, “Wow.”
PF You know what just happened as we had this conversation is it made far more sense to me why Oren is at Goldman Sachs.
RZ It’s an economy! He was an actor that had influence over an economy.
OM The analogies between the online advertising world and the financial world are abundant. You know, starting from an exchange through the type of—through transactions.
OM There are fundamental differences—ads are perishable inventory. If you don’t buy an ad it goes away.
OM You know, as you load the page if no one bids that’s, you know, it’s lost and a stock is just gonna remain there or the pile of gold or whatever people buy.
RZ I think you touched on it before which is it’s, like early days. It’s like a 19th century economy.
RZ It’s—there’s not a lot of regulation yet; it’s sort of the Wild West; the actors are trying to, you know, exert influence to protect their own interests, et cetera.
PF To be clear, we’re talking about advertising here not Goldman Sachs [all laugh].
RZ Yeah, exactly! Yes. Note!
PF Not like a 19th century—
RZ You’ve enjoyed Goldman as a client [others laughing]. Yes. Not 19th century economy.
OM I think there’s a lot of truth to that analogy as well. There’s the buyer beware and the level of sophistication that users need to have to protect themselves is not trivial . . . in online advertising.
OM Unless the industry matures that will—that will change.
RZ Oren, one sentence answer: what do you think of ad blockers?
OM Wonderful wakeup call.
RZ That’s great.
PF That’s a pretty great answer.
RZ That’s a great answer.
PF Oren, if people wanna get in touch with you, what do they do?
OM They are very welcome to—we’re—we’re building a big organization, an exciting organization trying to solve big problems that have a huge impact on the world and this is not an exaggeration, we’re trying to make resource allocation and investment decisions better for society, companies, countries, and individuals. It’s a major, major task and we’re looking for people who find that an intriguing problem on the product management side, on the engineering side, designers. Anyone who wants to be—who wants to take part in that adventure, we want you to join us.
PF That was fascinating!
RZ A lot of interesting insights here.
PF It’s like just watching lots of aircraft carriers moving around in a circle. Just big, big systems.
PF That goes very deep. Smart, smart fella.
RZ He’s pretty bright [laughs].
PF Seems to have learned a few things along the way. So look, you know, here we are at Track Changes, sometimes we have the ACLU, sometimes we Google and Goldman Sachs.
RZ That’s the beauty of Track Changes.
PF That’s why I love being in an agency in New York City because—
RZ You’re gonna see it all.
PF You get to see the whole world up close.
RZ You’re gonna see it all.
PF So listen, Rich, my name is Paul Ford, I’m the cofounder of Postlight. You’re Rich Ziade, the other cofounder of Postlight.
RZ Thanks, Paul.
PF What do people do when they wanna get in touch?
PF Just send that email.
RZ You got a question you want us to answer on the show; you want—you have thoughts, commentary.
PF We welcome criticism.
RZ We like to talk to you. So just reach out about anything.
PF [Music fades in] Give us five stars on iTunes if you’re just kinda sittin’ around the house with a lot else to do. Or even if you are kinda busy, go ahead, go upstairs, turn on that old iMac and give us five stars.
RZ That Dell XPS 1300.
PF Exactly. So, email@example.com and goodbye to everybody else. We’ll see you next week.
RZ Have a great week.
PF Bye! [Music ramps up, plays alone for seven seconds, fades out to end.]