Our SYNC Interview with Scott Knoll, CEO and President of Integral Ad Science
We sat down with Scott Knoll, the CEO and President of Integral Ad Science. The company works with advertisers and publishers to improve the quality of the overall user experience by better matching ads to quality content. It is not just about targeting ads, it is about making sure that the content the ad is shown on is both relevant and meets the criteria set by the advertiser. Scotts example is: you would never want to see a car ad, next to a story about a car crash!
Integral Ad Science was formerly known as AdSafe, a 2012 Big Award Product of the Year.
Interview Transcript
Russ: Good afternoon everyone this is Russ Fordyce from the Business Intelligence Group. Today we’ve got a really exciting interview with Scott Knoll of Integral Ad Science. They were a 2012 Big Awards winner for Product of the Year back when they were known as Ad-Safe and they’ve recently done a rebranding so we’ll hear from Scott about that. Thank you Scott for joining us and I appreciate you taking the time out of your day.
Scott: Thanks for having me. I look forward to chatting.
Russ: Tell me a little bit about how you guys started Ad-Safe and kind of the origin of the company. What was the original mission?
Scott: Sure, first of all I joined a couple of years into the company so I wasn’t part of the founding team so I can’t take credit for everything that was done. Certainly, the early work is what interested me in joining. The concept from the beginning of the company was-there was industry awareness around the problem with ads ending up on pages that were not intended to. It’s the car ad next to the car crash or your ad ends up on a pornography page and the brand is upset and the people associated with that placement are upset.
The reality is in the marketplace as we’ve shifted from buying by traditional media buying television show or tickets or magazines or a website to targeting audience; we’ve kind of lost control of where our ads are ending up. At the time that this awareness was generated in the industry around this it was all just after the fact awareness.
It was a report that pointed out the problem. What Ad-Safe recognized was that it was a problem for not just the buy side, but the sell side as well. A lot of times the sell side would be selling inventory they didn’t have control over and therefore had the same problems that everyone had. The original solution was let’s instead of just reporting on the problem, let’s come up with a way to stop it from happening.
The original technology which still exists today and is increasingly important on a daily basis is technology that before that an ad is served actually looks at the page and makes sure that the page is appropriate under the definition of what the advertiser wants and then only delivers it if it’s appropriate.
Russ: Yeah, so that’s kind of the secret sauce, right? Is getting I guess the quality and the audience right so that the ad performs for the purchasers …
Scott: Yes, although with the safety it’s making sure that the page itself is reflective of what the brand is looking for so if it’s an illegal download site for example most brands don’t want to be associated with something illegal so regardless of whom that person is they just don’t want to be next to something that’s illegal.
Russ: Alright.
Scott: There’s profanity-certain brands-family brands in particular want to make sure they’re not next to some article with hate speech or profanity.
Russ: Yeah, I imagine you don’t have a team full of quality checkers that are in New York or in the Philippines.
Scott: I laugh because a lot of times that was the original approach to it and it’s …
Russ: I imagine.
Scott: The scale is just too big and you need to be so granular looking at every single page because even with a big portal that has lots of inventory-some sections are on celebrity gossip and there might be some moron-you know young hip-hop music and there might be some profanity used or racial slurs used that you don’t want to be associated with so you have to have a machine learning type technology that can scan billions of pages and to date we’ve scanned almost like 10 million pages and [inaudible 0:04:09.3]
Russ: Are you doing that on a real-time basis or are you doing it kind of-are you indexing those pages –aka-Google and storing those results so that you know that’s a friendly page?
Scott: The story that I haven’t talked about yet is we do more than just ad safety now and some of the new stuff we’ll do on a real-time basis. In terms of ad safety we pre-score everything and it’s constantly updated so inventory that changes on an hourly or daily basis like a news site will continually call it to get the latest information. There’s no way for us to look at the comprehensive information that we do and do the complicated analysis that we do in time during an ad call to make that decision. So it’s something that we look at in advance.
Russ: You’re no stranger to this space-you’ve been in kind of advertising and technology for a long time-how long have you been with Integral?
Scott: I joined about exactly two years ago, but your point-I was at DoubleClick in 1998 and before that I was at Time, Inc. during some really early days on ad stuff so I’ve seen this industry go through a lot of changes. Most recently I was at a company that I had started called Aperture where we focused purely on audience targeting and trying to gather as much data as we could on the person based on purchase behavior, surfing behavior-some offline data that we could bring on anonymously online. What I learned in that was a couple of things-one is-it’s really hard to scale with data like that-particularly in some of these exchanges. Also I learned that it can be really inefficient to pinpoint just on the audience if you don’t know where the ad is ending up. Not only is it potentially dangerous where your ad ends up on a pornography page-which by the way when I was selling media as part of Aperture-I ran into those issues and they’re not fun to deal with when you get a call from the CEO of a company very angry particularly when you had no idea that that had happened either. I understood the problem but also what I understood is that a lot of ads are well-intended where you have the right person and you’ve done a lot of analysis ahead of time to note exactly the browser that you want to focus on but a lot of ads are just completely missed because they’re one of 30 ads on a page, they’re on a page with all kinds of other bright graphics and movies and other stuff going on or they’re just-you know it’s not a page that people necessarily spend a lot of time looking at or trust, therefore well-intended ad but totally misses the mark because it’s not in the right environment.
Russ: You mentioned that you’re working with ad buyers are you also working with the content providers in the site themselves to make sure that you’re calling appropriately? Are you working both sides of the equation there?
Scott: Yes, some of it is the quantities but more of it is just helping them solve these same issues. As I’ve said from the beginning with Ad-Safe before I joined recognized that this was a problem for both the buy side and the sell side. Half of our business is on the buy side and that’s working with agencies or direct with marketers. The other half of the side is the salespeople were selling inventory and so half our business comes from that side to help them. Its helping them understand to keep away from areas like unsafe inventory, fraud is another issue that even the sell side struggles with that we help them with. Also just understanding like what parts of their inventory works better based on where the layout of the page is and helping them understand how to value that and sell it more appropriately.
Russ: You guys are obviously in this kind of leading-edge technology. You’re developing this technology, how many other competitors are out there that you’re looking at or that are doing similar things either on the buy side or sell side. Are there other guys gunning for you?
Scott: We don’t know of any. I kind of joke but certainly there are-when we look at the playing field and we look at what we do we essentially compete in four areas and we see some competitors who do some of what we do like we’ll compete against one area or another but it’s interesting when you look at those four different areas and they’re essentially high-level-it’s measuring, it’s attribution, it’s the analytics decision and then the targeting decision. Of those kind of four areas there’s no competitor who we compete in each of those four. We think we’re very unique in how our approach to the marketplace. Even where we compete on people say in the real-time bidding or data for programmatic buying which I call targeting; everyone else’s approach is very much context-based. Let’s look at the words on the page and figure out what they mean and should we target based on whether it’s appropriate or contextually relevant. That’s one part of what we do-we combine so many other elements to the page, where the ads are, was the ad in view, how long was it in view, what was it in view, were there other images on the page, is the html code professional, was it written by someone in a garage with some sort of automation tool-all this data that we bring in decides the words-are so important to understanding-is this the right ad to target or not.
There’s no one else who’s looking at it from that perspective. I think it’s changing a little to be honest with you because I think the sentence market has been so focused on cookie data-which essentially is audience data that all the investment you look at from-you look at the lumen chart-almost all the investments are on people who are trying to aggregate more data or bring more data into play around an audience.
There’s almost nobody doing what we’re doing-which is the other piece of it-which is where an ad is going to be and from what we see in every single test we’ve done, every single kind of engagement we’ve done-that’s actually the most important piece-which is the actually have to see the ad for enough time in an environment where it gets their attention and they trust in order to have any impact at all. If you don’t get that right I don’t care how granular your charting is, I don’t care how much you pay for this cookie that has the exact history of buying you’re looking for. In a skin market absolutely, if they’re not going to see it-it doesn’t do you any good.
Our whole industry has been built on just trying to find the right person-we’ve forgotten about the fact that most of the time this ad is actually not seen or not seen in the right environment. That’s a big piece of-big, big important recognition that this industry is waking up to right now and realizing.
Russ: I guess that you’ve been in this space-I won’t say so long-but 1998 is pretty long-I guess we’ve lived through the banners, we’ve lived through kind of this-you know the text-based ads that Google has gotten into and others. How do you view the ad market changing in the next three to five years? Obviously mobile’s at play here-but how do you view it changing? I think we-I would imagine over time that the banner click through rate has either declined or stayed the same so how do you see that evolving over time?
Scott: It’s interesting when you put it that way because I looked back to the early days and some of the stuff that we’re talking about then are becoming a reality now so I think the vision was always there-it takes a long time to get there from a technology perspective-from an execution from a proliferation perspective so everyone’s using it. In general what we’re seeing-if you look at from a high level-what we’re seeing is an industry that started out with a little automation is becoming very automated and it’s really important that this happens because one of the challenges with online advertising is it is so complex in terms of the choices. Forget about the technologies you can use to understand whether it’s working-just starting with your choices. Television, even with cable, you’re limited in the number of choices you have.
Obviously print-the same thing and these are where the big marketers are used to spending. You go online and even if you say alright I want to buy women between 30 and 40-there’s still a hundred thousand websites that probably reach that sector-so let alone mobile versus whether you do a video ad versus other social and other ways to reach people in digital. There’s this big amalgamation of things we call digital-it gets more complicated every year as there’s new ways of reaching people and new ways of selling it.
It’s so hard to make decisions that if we don’t automate that-we’re never going to move this industry forward. The automation is really important-what we’re starting to see and where we’re going to change over time is initially the automation that runs around simple things that we can measure. Click-through-as you mentioned, the big automation now is around Last Touch. We have to find a way-we realized that click though isn’t a good measurement to get credit on. Most people don’t actually go to an ad click on it and go buy. It takes a while to sink in that they want something so how do we give credit?
We come up with this idea that’s called a view-througher or last touch credit where you try to go through the path that someone went to before they actually bought something and give everyone some credit along the way. In order to do that in a simplistic way or a way that you can make it work in the industry-you essentially you just sprinkle credit around or give the most credit-most people do-to the last place they saw an ad before they went to Google and converted it.
Whereas in that makes sense because it’s simple and you can anchor decisions on it. It’s really easy to gain and it’s a partial view of what’s actually happening. We’ve automated-but we’ve automated around the wrong things. Automation around the wrong things can create all kinds of problems-one in particular-like fraud-the reason for fraud is because we’ve automated around the wrong thing-that’ll ask people to game and creates this nefarious activity that hurts the industry.
I think as the industry moves forward we’re going to see more automation-I think that’s really, really critical for the reasons I stated earlier. But then automation around more than just one thing-automation around more holistic viewpoint of what’s working and more data sets to better automation. That makes sense.
Russ: Do you see the format of advertising changing at all?
Scott: I think in part of this it does. Again, the reason the format or the reason someone has an interest in moving from one format to another is because they look and see did it work or not? If you don’t-getting that right-did it work or not-you’re not necessarily going to move into the new formats that are working better. Again I use the example that we work with some clients or you read articles in the industry and there’s a very simple conversation going on right now on the surface- yet very heated debate-does an ad in view matter or not.
There are people that will vehemently argue that an ad doesn’t have to be in view to work. In fact ads that aren’t in view actually work better and on the surface you listen to that and you’re like what? That doesn’t make any sense yet their adamant in their argument and they show that their conversion rates and ultimately their ROI goes up when they buy inventory that’s not in view.
Until we recognize that there’s a measurement problem not a targeting problem or not just a targeting problem but also a measurement problem in deriving whether someone’s successful or not is going to be hard to figure out exactly what works the best. We say banners don’t work-we actually don’t know that-we just know the way we’re measuring it-a lot of banners don’t work.
Ultimately though I do think that as we’re looking at new things like what we’re talking about-the right environment-we’re going to find an equilibrium between the reach and the right environment where an ad is definitely seen in the right format that’s actually going to persuade someone to purchase something in the future.
I don’t think we know what that is yet. I think that there-again it’s so easy-unlike television where you essentially have one choice; you take up the whole screen and it’s just a matter of how long it’s in front of them-we have so many infinite choices right now-we’re still trying to figure out how to do that and how to standardize it but absolutely it will change.
Russ: I appreciate that. We’ve been focused on I guess talking with CEOs like yourself-a kind about it that the explosive growth curve that a lot of the companies are going through-especially the startups and I imagine you guys as well-how do you manage that kind of hockey stick growth as you guys get to run rate.
Scott: It’s a great question in one that you think you master and then-cause I’ve been through some heavy, crazy startup times-like DoubleClick-I was there from 150 people to 2,000 people in a two year period. It’s always challenging-we’re here-I joined when we were 15 people-we’re now 80 something people-we’ll have over 100 people soon and it’s-and we’re in the US now we’re in multiple countries-Europe and Asia and it’s never easy-it’s exciting, it’s fun to be part of; at the same time-you know quickly people are stretched and you realize that a lot of times you have people in roles who have never taken on responsibility managing-you are now managing big teams. I think that you know from my kind of experience and lessons I’ve learned part of it is just you can never communicate enough and even if you think you’re communicating enough it’s probably not nearly as much as you need to.
I think it’s just also trying to simplify the company’s message-not just externally but internally and just continually reinforcing that because as you’re growing one of the dangers of you’ll get into as a company is becoming unfocused because there’s like so many shiny new pennies out there that you could go after and so many new businesses that you can pursue because every conversation you have with a client or potential client leads to oh could you also do this for us, can you do that for us. You can’t focus and make sure that if you are doing something it’s not far from your core-that’s when you lose control of everything …
Russ: I imagine for you guys that’s even more of a problem for most-you’re working with some very large brands and the types of revenue I would imagine you’re getting from a few customers is pretty significant so I imagine they’re bringing you kind of adjacent to your core pretty regularly.
Scott: Absolutely, not just brands but also data technology companies who are helping power some of their solutions so it’s true and I think it’s also again-it’s not like you’re never going to do something that’s out of your core but you got to be very careful with your bets and what goes into that criteria-part of it is strategic, important part of it is revenue and part of it is just what is the opportunity cost of if we’re going to do this what are we missing out on.
We’re constantly looking that and evaluating it and I got to say something its gets harder-you know a year ago it was hard now it’s even harder because now we have triple the number of conversations we had a year ago and then triple the number of opportunities on top of that. It’s a constant struggle but you know at the end of the day that’s what keeps me excited and keeps me going. If I don’t think I had that-If I didn’t like that I wouldn’t be in this role. It definitely takes a certain personality to want to do that.
Russ: You guys have expanded your product portfolios significantly over the last two years and as you mentioned before it was Ad-Safe a security product to begin with and now you’ve got those other three arms that you’re talking about. Were those largely inspired by client engagements or did you see a need in the market or how did those come about?
Scott: That’s a very good question. It’s both-it’s inspired by everyday conversations we have with clients and seeing the challenges. It’s inspired by all the data we’re collecting and we’ve got a big data science team who’s looking at data saying oh my God only 20 percent of these ads are in view-how come-and digging into the details to figure out like what are the factors-to give you an example-view-ability-that’s a symptom of other problems. We dig into those problems to see what those are. Then part of it is just understanding at a very high level what are the challenges of this industry. What’s holding his industry back and then figuring out based on what we see-the data we see and what we do-how we positively affect that?
Part of that came from my old job when I was at my old company Aperture I was frustrated by scaling issues, I was frustrated by how inefficient exchanges were. I thought ultimately cookies are not the way to go. We got to find another way to target-we have to find another way of making advertising decisions that aren’t relying on a small sample of the inventory out there.
Our big vision is-what we’re probably-using a baseball analogy in terms of what we’re doing-we’re probably in the second inning of a nine inning game that might go into extra innings. We’ve got so much we want to do. That goes back to your challenge of you said about growing fast and making decisions. I can’t go fast enough in terms of more stuff we want to put out but at the same time we got to make sure that what we have out works and is successful.
Russ: You’re a pretty motivated guy-we’ve seen that here today. In the hour you get a week of downtime what are your interests and what are your passions?
Scott: Also that hour depends on what’s going on but I’ve got two little girls so they definitely take up a lot of that time and more than an hour as if you have children you’ll know. I’m told by my wife that I should do more age appropriate things but I’m also trying to do an ultra-marathon in two weeks from now which I’m behind in the training. I’m a big flat fisherman and the season just started so I’m trying to get out in the water. Can’t get everything I want done but certainly that’s on my wish list.
Russ: I wished you a nice unplugged weekend in Montana somewhere then to get the rods out.
Scott: I appreciate that.
Russ: Scott, I appreciate you taking the time. We’ve been interviewing Scott Knoll of Integral Ad Science formerly Ad-Safe. They were a 2012 Big Awards winner for their product of the year. We wish you continued success and really appreciate you joining us today and thank you for your time.
Scott: Thank you Russ, appreciate it.
Russ: Have a great day; alright, bye-bye.