← Marketecture: Get Smart. Fast.
Marketecture: Get Smart. Fast.
Stephen Yap on Perion One, AI Integration, and the Future of Cross-Channel Advertising
00:23:24
Loading summary
Transcript134 lines
- [00:10]
A
All right, welcome to the marketecture podcast where you can get smart fast with in depth interviews of technical executives. I'm Ari Papara, I'm joined today. This is Advertising Week, first of all. So we're excited to be here in New York in person. You may notice a little bit of a different format with my good old friend Stephen Yap, who I always take pains to prod on the podcast. But finally you're on. So Stephen, thanks for joining me.
- [00:33]
B
Thanks for having me. Yeah. Big week in Ad Week. Woo hoo.
- [00:36]
A
Yeah. Advertising Week. It's not Ad week. Ad Week is a trademark of a different company.
- [00:40]
B
Advertising Week.
- [00:41]
A
But I am probably going to go to the Ad Week session. Ad Advertising Week. Even though they're competitors, it's confusing. Why are you here, Steven? You're the CRO of Perrion.
- [00:51]
B
Yeah. As you know, after 25 years of combined servitude with Google and DoubleClick. Servitude. I was in paid service, so yes, yes.
- [01:02]
A
Okay, fine.
- [01:06]
B
I went over to Perrion. They're doing some really interesting things and again trying to get back to my roots in like I was at Doubleclick in early days. Google kind of building out solutions that meaningful and have an impact.
- [01:19]
A
How's the free food? Is there free food at Parion?
- [01:21]
B
There is some free food there, yeah. It's not as extensive as the Google cafes as you can imagine. Like there's no sit down dining hall as there is in Google or cafes. But we're small but mighty.
- [01:37]
A
You have to fend for yourself. It's how you stay sharp.
- [01:39]
B
Yeah, it's actually you're going back to the real world. It's kind of interesting.
- [01:43]
A
I know this is off topic, but some people say they join Google and they instantly gain 10 pounds like the freshman 15. And then some people do the opposite. They join Google and because there's salad all the time, they lose 10, 15 pounds. How do you feel about the healthfulness of free food?
- [01:59]
B
Yeah, so I think it's us just on a sliding scale. I think when you first start you're basically pillaging the micro kitchen for peanut M&Ms. You're taking those back on your commute with you, all the diet Cokes. And then at some point you kind of hit that point in the graph where it's just you get into self realization and you're like I'm going to have a cardiac arrest if I keep on eating like this. And so then you slowly start to kind of lose your weight and go down.
- [02:28]
A
Yeah, I always had peanut M and Ms. Speak to me because I even had an incident with them while I worked at Google where I was. I don't remember exactly the circumstance. I think this whole story's on my blog somewhere from back then. But yeah, someone tried to interoffice me. Peanut, M and Ms. Because I think I was complaining. The New York office didn't have any.
- [02:45]
B
Didn't have any.
- [02:46]
A
And the mailroom was unable to deliver. It is against the policies of the internal Google mail system.
- [02:52]
B
Look at you breaking Google policy from an early age. I'm sure that was a whole section in your book too.
- [02:59]
A
Like I said, I think I wrote it up at the time on aripaparo.com, but it didn't make it into the final cut of the book. All right, I guess we should talk about Parion, right?
- [03:07]
B
Yeah. Yeah, let's do it.
- [03:08]
A
Okay, so you've been there coming up on a year maybe.
- [03:12]
B
Yeah, almost a year. So I started in February.
- [03:14]
A
All right. And there was a pretty big kind of product change. You released Parion 1. So take me through what is Perion 1? What is Perion and what is Perion 1?
- [03:22]
B
Yeah, no, so for a long time. So Perion has been around for about 25 years. And so it was mostly just a holding company of multiple technology pieces. And I think, you know, it's. It's kind of, you know, the acknowledgement of. With the innovation of today and basically how ad tech has actually proliferated everywhere. There's obviously a pretty strong understanding that the world is just inefficient, Right. And you know, per earlier conversation around, you know, this. This kind of culture in technology, right? Watching my kids and things like that, of how they interact with technology, it is, it is an instant gratification consumer culture, like, oh, I got it on my watch, I got on my t. Got on my phone, my tablet, you know, what. What have you. And so I think where we're seeing kind of breakage points is around kind of how efficient and effectively can I get marketing out there, right across all the different channels and know exactly who I'm talking to, how I'm optim. Optimizing and how I'm actually spending my money. And so in looking at that, instead of having multiple products and multiple sales teams and multiple contracts and going to market like that, you have like four or five teams talking to the same person. We looked at it as a unification of kind of people, process and technology, right? Where it's like, all right, so we can bring the people together, get the processes Together, So it's a single contract. But the technology, what if you were able to actually log in and be able to start to media plan across all these different channels. Right. Instead of having to learn 16 different technologies. And so that was the impetus of Perion One. We launched that in February. Then obviously we've continued to develop our technology assets and then actually had the green bids.
- [05:16]
A
Right, Green bids, I forgot about that. Yeah. So let's break it up. Having it all in one place is great, but I don't know what all of it is. The 16 channels perion is known for. What Undertone search Green bids. Walk me through some of the key things that are part of that.
- [05:30]
B
Yeah, when you look across the channels, obviously everyone was familiar with the old undertone business in our, our industry. But you know, it's also kind of the marriage of. We had a business called Cold Fuel, Code Fuel, a business called Vidizoo. We have the old undertone business which was very heavy in kind of this high impact format. So really cool creatives. Probably the coolest I think I've honestly seen in my time in the industry. And then we also bought a digital out of home bidder. And so that company, which was Hivestack, is the best by far. Digital out of home technology. And trying to fit figure out, okay, so you've got ctv, you've got standard display, you have your search assets, you've got digital out of home assets. And then we obviously bought Green Bids, the custom Algo.
- [06:21]
A
It's custom Algo and Carbon.
- [06:23]
B
Kind of like. Yeah, so it's custom Algo Carbon and then it's across obviously YouTube and DV360, but as well as things like Meta. So being able to actually bring those all together and be like, okay, so here's kind of your overall media plan, here's how you want to buy it. Can you be more efficient and using kind of algorithmic AI and how do you optimize across all those channels? Because eventually you're going to get to a point in the industry which I feel like we're rapidly heading there. And I think that is the really interesting thing, which is how do you then leverage AI to allow AI to determine how you're distributed across all of these different channels and do so really agnostically. Because I think that's also a huge question out there with some of the larger tech providers.
- [07:14]
A
Yeah, I think when you have a big tech provider that has multiple products, they can put incentives to using them. They may not be in the customer's interest. They May be with your thesis about bringing everything together, how do the customers specifically benefit? Is it budget allocation optimization, etc. And what have you seen in reality?
- [07:35]
B
Yeah, so it's actually all of the above. Right. So it's kind of better optimization, better ways of working. Right. So we're bringing back kind of that, hey, we want you to do more with less. Right. So when you look at like our algorithmic business, your ability, you know, one of our biggest customers was basically talking to us about, look, your ability to make a million dollar budget feel like it's reaching $2 million is rarely telling for us. Right. Especially when you start to get into these really weird, you know, financial markets where, you know, you're getting price compression, the marketing budgets are being kind of scrutinized, which they seems like they always are these days.
- [08:17]
A
Yeah, sure.
- [08:18]
B
But you know, when you have a solution like, look, your budget might have gotten cut by 70%, but I can still make your budget feel and perform like it's at 100. So being able to actually apply that, and you're taking those algorithms and apply it not only to what you're doing in CTV and display and search, but also being able to do that across things like digital out of home.
- [08:38]
A
Yeah, sure.
- [08:38]
B
In your YouTube buys make a lot of sense because you're obviously using the same data sets to optimize them.
- [08:44]
A
And do you see that? Is the customer coming to you and they already have a media plan that includes say a format like digital out of home or is it opportunistic saying, oh well, I'm already doing say undertone high impact ads, I might as well do digital?
- [08:57]
B
Yeah, it's both. Right. Like, like honestly there's, I think there's, there's an opportunistic way of doing it which is kind of the land and land and expand type of model. Right. Where it's, hey, I'm, I'm doing wonders for your budget on YouTube. Right. Oh, you also do digital at home. Yeah, we see you do a lot of it. So we can actually do the same thing across, across these channels. And then there are the other ones where, where we know obviously we've got competitors out there in the market, but we can actually make them much more efficient. Like in, I know our European teams are doing this very, very well today, which are, you know, look, we really incredible efficiencies across both your CTV buys and your digital out of home buys using our custom algorithm.
- [09:37]
A
Right. So how does a custom algorithm. Well, I guess going into the depth, I wanted to explain It.
- [09:42]
B
Yeah, you know better than depths of me already, especially on the technology side. Shame on you.
- [09:47]
A
Right. But digital out of home and CTV is a nice combination. Definitely something that would be interesting. So how does AI fit into it? You kind of have to talk about AI whether you want to or not. It's a requirement.
- [09:59]
B
Yeah, no, it seems it was kind of the running joke which was the most used two letters in probably ad tech in the last three years.
- [10:07]
A
Other than feu. Sure.
- [10:11]
B
We continue to run in the same circus. Yeah. So I think. So first off, I think there's two kind of polarities of AI. I look at it in the marketplace. The one which is everyone talks about these big, bright, beautiful universes that are controlled by AI. Reminds me of the big data conversations of old.
- [10:29]
A
Oh, I remember those.
- [10:30]
B
Yeah. It cost a lot of people their jobs.
- [10:32]
A
Give us all the logs, even though we don't have any employees.
- [10:34]
B
Employees who could use the log or infrastructure. It was like when cloud was just getting started. No one could do it and everyone got fired. And that famous meme about the CMO's job is like 16 month tenure or something like that.
- [10:46]
A
Totally. Well, with AI, you don't need to.
- [10:50]
B
No, it's not true.
- [10:51]
A
Okay, tell me about AI.
- [10:53]
B
That's not true.
- [10:54]
A
Tell me about the CMOs.
- [10:55]
B
CMOs out there. That is not true. That is not true.
- [10:56]
A
The CMOS typing into the prompt, I want a unified solution and then boom, you pop up and everything's good. Yeah.
- [11:02]
B
I mean, our Google AdWords account is legit.
- [11:04]
A
It is.
- [11:05]
B
No. So you have that on one side of the polarity. The other side of the polarity is just a simple application. Right. And a lot of that has to do with how are you optimizing, how are you ingesting data into those systems and then how you're working across those systems. Right. So think of it as like actually the AI being the connective tissues. It's just kind of looking at each one. Right. I think that's also been a problem, or at least it is becoming a problem within the industry, which is we talk about all these silos and the silo story, but we're doing the same thing again. Everyone's creating AI on top of their silos. I got AI here, I got AI here. I'm like, great, we haven't solved any of the problems. I think the things that we're looking at, quite frankly, are, as I mentioned, being able to leverage our AI across all of the optimization architecture that we currently have in every channel. What does that really look like? Well, when you apply it, the easiest way to explain it would be as the campaigns are running and they're running concurrently. The system's looking and going, oh, well, we're ingesting weather data and it's going to be negative 40 degrees in a snowstorm in New York City. So this week I'm going to pull my digital out of home budgets and put them into ctv, because no one's going to be walking the streets of New York.
- [12:20]
A
That's cool.
- [12:21]
B
Everyone's going to be sitting on the couch. Right, right. And so being able to actually execute on those things, I think through AI, that's going to be the reality, because I do think ultimately we're going to get to a world where the CMO, who's still very important, not eliminated by AI, not eliminated by AI, but the CMOs, they're going to be looked at as almost a portfolio manager by the cfo.
- [12:44]
A
That makes sense.
- [12:45]
B
Which is like, here's your budget, here's the return I'm expecting. Do I necessarily care where you spend it? No, what I want is the return. And so obviously it's going to be a little bit more complex with that, with obviously brand suitability and things like that baked in. But I do think you'll get to a place where it's a nice marriage between people and technology, where you're going to be able to be much more efficient in terms of what you want to do, but also have a little bit of that control.
- [13:09]
A
Is your AI built on top of the green bids? Is that like Fulcrum?
- [13:13]
B
Yeah. So we combined all of our AI assets into kind of Paron Algo, which is basically the center point of all of our algorithmic aspirations, if you will.
- [13:27]
A
Right. And that works Both on Parion 1 and on other platforms.
- [13:32]
B
Yeah, yeah, yeah, yeah. Because you got the DV360 and Meta. Yeah, yeah, right, yeah. And so I think, again, I think it does speak a large part into the way in which we want to position ourselves. I mean, I've long held the belief that ad tech kind of lost its way during this journey, which is everyone wants to own this concept of owning the customer.
- [13:50]
A
Sure.
- [13:50]
B
It's like, oh, who owns the customer? And you look at it and you're like, in reality, no one does.
- [13:54]
A
Customer owns themselves.
- [13:55]
B
Yeah, customer owns themselves. And by the way, if you think you own them, open up the door and you see 10,000 of the same people like you waiting outside to meet with their customer. So I think ad tech really needs to be more of an enabler, an underlying technology. So it's like, look, I can fit in to at whatever point you're at and whatever part of the journey you may be at versus you have to come into my workflow, you have to use my data. And by the way, I'm going to restrict what you can give me because it's against my policy or so on and so forth.
- [14:26]
A
So do you think I want to explore this idea of the CMOs as the portfolio manager and also how it relates to AI? Do you think that the CMO is now looking to contract with one entity, their agency, who's going to be the AI ringleader for them? Or are they going to look for their own in house tech? That's AI. Or is there some hybrid.
- [14:45]
B
I think that has yet to play out, to be honest with you. Right. I think everyone, we made the joke earlier when we started the conversation. Everyone's got an AI story. No matter who you are in the marketplace, including obviously the large holder companies.
- [14:57]
A
They're investing a lot.
- [14:58]
B
Yeah, they're investing a. Of money.
- [14:59]
A
Ton.
- [14:59]
B
I mean I think one of the greatest moves they made unknowingly, right. Was to get ahead of the, you know, there's a whole thing called cookie deprecation where we're going to eliminate some sort of tracker that's out there on the web apparently.
- [15:12]
A
Yeah, I think our audience is familiar.
- [15:14]
B
Yeah, there's this big thing, you should look it up, but it's not happening.
- [15:19]
A
Anymore, so we're fine.
- [15:21]
B
I have no recollection of that, Ari.
- [15:23]
A
I did write about that. Anyway, go on.
- [15:28]
B
So I think with the agencies end of the, you know, kind of falling into was like, hey, we're trying to build our identity graph for the field of the cookie deprecation. What they landed with these huge data companies, right. Which has been pretty incredible, right. Because if you start to feed your AI aspirations and you're using that type of data to feed it, it can be incredibly powerful. Now I have yet to see kind of an example from any, from a lot of the different companies as to where they specifically are. I think everyone's got an example of how they're using it and leveraging it. But again, I think as marketers explore, a lot will be determined by kind of where their appetite is. Right. If they want kind of controlling of it, or if they're really using an agency partner in a global framework, what does that look like? And things like that. Because again, what we often forget is the AI is governed by the data that feeds it. The data is governed oftentimes by political and policy entities that are often different across the world. Right. The GDPR's, the California Privacy. So I think as this plays out, I think you'll see a different model reveal itself. Even though I'm not quite sure what that might be.
- [16:37]
A
Yeah, that's a good point. I think that there are applications of AI. Like creative is an obvious one. It's not clear that they all converge into one giant brain, advertising brain. You might have a creative system. That's where the agencies have been investing so much then. And on the media side, it might be a very different kind of approach.
- [16:58]
B
Yeah, yeah. And I think at some point they're going to have to connect with each other some way.
- [17:04]
A
Some way.
- [17:05]
B
Right. Because there's, I think what is also when you start talking to a lot of the marketers out there and the agency folks out there is there's a need for consistency. Like I can't be inconsistent in this channel and then be super efficient in this channel. Right. Like, so I do think ultimately you'll get to a point where it's just going to be a forced model. Yeah. But again, I think, you know, everyone who's kind of building these AI entities, including us, what we're trying to figure out is, look, we, we're, you know, we, our aspiration is we're going to be best in channel across because that makes you really, really relevant.
- [17:40]
A
Yeah.
- [17:40]
B
Right. And really efficient. And then we're going to layer the AI on top of that and we'll meet you wherever you need to be. So if you're just doing digital at home, great. We're going to, we'll layer our AI capabilities on top of that. If you're plugging CTV into it, it perfect. The AI model should be able to replicate those media plans and the ways in which you optimize to create that consistency. So, yeah, I think it's going to be an interesting time in the next four to five years.
- [18:03]
A
Yeah. Any things you're hearing about AI's impact on web traffic, the open web and stuff like that, is that something you've got your ear to the ground on?
- [18:14]
B
I mean, having spent a ton of time there, I was asked that question in another interview. In terms of you're doing other interviews.
- [18:22]
A
Your two timing podcast, I mean, I don't feel good.
- [18:25]
B
They're not as important as yours.
- [18:27]
A
I need to own the guest. Own the guest.
- [18:31]
B
As I look at my competitors. Outside your door. No, but when you look at the open metab, I do think having AI Being able to kind of leverage and look through the landscape for you for relevancy and relevancy. How appropriate it is. And. And the probability of this is what I'm looking for. I think it actually will offer some opportunity for some of the smaller publishers to actually kind of surface to the ground and have actually better chances at showcasing their value. I do think there's a lot of chatter in the open web's. Dad. I think there's a little bit of truth in the decline of the open web, obviously, because everyone sees it in their numbers. But will it be fully destroyed? I don't think so.
- [19:22]
A
Numbers are going down, but a lot of what's going down is the lowest quality traffic. Yeah.
- [19:28]
B
And so I think by that, the question then gets turned right is, can AI help us clean up the ecosystem of the bad actors? Because we've seen it. Bad actors are ridiculous.
- [19:40]
A
Yeah, bad actors, people who are creating hundreds of pages of content so they could spam, SEO. I mean, you. We're old enough to remember the, like, what time is the super bowl content? Right.
- [19:51]
B
What time is the Super Bowl? Wow. Yes.
- [19:53]
A
Those were good times when you'd publish an article like that and get thousands of views just because Google had no concept of when the super bowl was.
- [20:00]
B
No. And it was. I mean, and that's the funny thing is that I think there are going to be ways. Right. In terms of the word of caution, which is it always seems like the bad actors are always one step ahead of the people who actually want to actually be beneficial to the. To the industry.
- [20:18]
A
Sure.
- [20:18]
B
And so I do think we'll still find ourselves in a weird place. And I think we see it today. Like the, The. I mean, some of the. Some of the spoofs are out there are incredibly impressive. Like, I'm. I'm afraid to actually answer any of the, like, either phone calls or inquiries.
- [20:34]
A
Oh, yeah, like phishing stuff.
- [20:35]
B
Yeah. Because it's like, oh, I'm working with your doctor's office. I'm like, are you?
- [20:39]
A
I told. I gave. Everyone in my family gave a cook that supposedly we all remember.
- [20:45]
B
Why is that so brand appropriate for you?
- [20:47]
A
If I ever call you and ask you to do anything really weird, like send a ransom amount in cash, there's a code word, and I will use that code word if it's really me.
- [20:57]
B
Hi, Mom. Right after my podcast, I flew to the Himalayas and I ran out of money.
- [21:03]
A
It's like, mom, something about chrysanthemum is. You know, it's a smart. Anyway, it's a Best practice. Let's just say that. So we usually end these interviews with a lightning round. I'm just going to do a very abbreviated.
- [21:16]
B
Okay, let's see. This is going to be interesting. I'm holding my breath.
- [21:19]
A
Just a couple quick questions. What is your number one competitive advantage?
- [21:25]
B
I think it's our technology, to be honest with you. Yeah. I think we have the best technology out there in terms of the channels that we play in.
- [21:32]
A
What is your number one challenge on the market?
- [21:34]
B
I think our number one challenge, given kind of this shift to param1, is just people getting familiar with the brand you rebrand. You know how that works. People have been brand loyal. They think they know what you do, but you've evolved. And so trying to actually make sure you're getting the right message out there in terms of what challenges are you looking to solve for sure. And how you can help me and that you're actually good at what you do.
- [21:59]
A
And last question. If Perion was an animal, what animal would it be?
- [22:03]
B
Ooh, the honey badger.
- [22:08]
A
Honey badger? Why is that?
- [22:11]
B
Paryon's been through a lot in its history, and it just. The thing never gives up, and it's absolutely ferocious.
- [22:17]
A
Honey badger don't care.
- [22:18]
B
Honey badger don't care.
- [22:19]
A
All right, so, Steven Yap, the CRO of Perion. Thank you for joining us.
- [22:23]
B
No, I have one quick question for you.
- [22:25]
A
Oh, you have a question for me? We're not done yet. No, no.
- [22:28]
B
When am I going to get. You promised me when we saw each other at the live ramp at the Ramp up event. You're going to give me signed copy that said help. Thank you for helping to ruin the web.
- [22:39]
A
I only vaguely recall that. Have you bought a copy?
- [22:41]
B
I have, yeah.
- [22:42]
A
Okay, well, you could have brought it. I would have signed it.
- [22:44]
B
Okay. I know. I actually thought about that on my way over there.
- [22:46]
A
So, you know, I don't know. Teller, are you going to be at marketecture Live?
- [22:50]
B
Yeah, actually, I think I'm actually in your copy.
- [22:52]
A
I'm going to be happy to sign.
- [22:53]
B
I'll be one of a thousand people standing in line waiting for the signature.
- [22:56]
A
And this is a public service announcement. I am not above signing your copy in any circumstance, at any conference, at any dinner, in the men's room. Just give me a pen and I'm ready to go.
- [23:06]
B
It's not the cleanliness way to do it, but okay. Thanks for having me, man. It's good seeing you.
- [23:10]
A
Absolutely. It's a great conversation. Thanks for being here.
- [23:12]
B
Yeah, cheers.