← Marketecture: Get Smart. Fast.
Marketecture: Get Smart. Fast.
SSPs Beyond “Pipes”: OpenX’s Matt Sattel on the Future of Curation, SPO, and AI
00:25:41
Loading summary
Transcript95 lines
- [00:10]
A
Welcome to the Market Sector podcast where you can get smart fast with in depth interviews of leading technology executives. I'm Ari Paparo. I'm joined today during Advertising Week in person with Matt Satel, who is the CRO of OpenX. Matt, thank you so much for being here.
- [00:24]
B
Yeah, thanks for having me and thanks for hosting me here. It's great.
- [00:28]
A
We should do these in person things more often, but it requires me to leave the house.
- [00:31]
B
I was gonna say you work out of your house. Are you inviting us over?
- [00:33]
A
This is actually the first podcast I've ever recorded with pants on.
- [00:36]
B
Very cool. Good to know. I'm gonna watch all of your videos.
- [00:39]
A
Differently now, more carefully. Yeah, exactly.
- [00:42]
B
I think that's why you encourage everyone to listen and not to not watch the YouTube.
- [00:46]
A
Yeah, exactly. We're gonna cross post on OnlyFans and make a little extra money.
- [00:52]
B
Innovator. We're here to talk about innovation, right?
- [00:54]
A
We are, we are. You have a big week ahead at Advertising Week. Yeah.
- [00:59]
B
Feels like everybody is here.
- [01:01]
A
Everyone seems to be here. Yeah.
- [01:03]
B
A lot of good nighttime events yours tonight. Excited for that one?
- [01:06]
A
Oh, yeah. We're recording this the day of our party. So the Antech God party and then a lot of other good stuff. But let's talk about OpenX, because OpenX is the OG, the ox. Been around for a long time. It's evolving, it's changing. So what's the latest and greatest? Why should listeners care? What's exciting about OpenX nowadays?
- [01:33]
B
I don't know if we've ever really talked about this, but I came more from the buy side. I worked at a company called miq, previously sold to buyers for a while. I think as OpenX was looking to evolve their strategy, get deeper with buyers, they took a different approach than I think a lot of people did. They didn't hire from another ssp, they didn't hire from a publisher. They hired somebody who had pretty limited SSP experience and more of the buy side experience. And I think if you pay attention to any of our product launches or the strategy, I think it's pretty indicative of that strategy because a lot of what we're building towards is buyer tools. That's why I think we were very. I'm sure we'll get into curation. Why we are very early in creation with actual product and actual technology behind it. But I think for us, I think it's around staying true to what an SSP is always meant to be. Not straying from that, but also evolving. Because. Because I do think that there is a group of SSPs who are not innovating and they do look like a commodity. I think for us it is very centered around driving innovation while also staying true to what an SSP is.
- [02:45]
A
All right, well, let's talk about what is an SSP nowadays? If someone's cynical, they could say it's a pipe. It's just traffic volume. Why is OpenX bigger or better than that?
- [03:00]
B
You know what's funny is I don't necessarily think being a pipe is a bad thing all the time.
- [03:04]
A
People need pipes.
- [03:05]
B
As long as you try not having pipes. My dad is an engineer, so maybe it's my civil engineer. Growing up with a civil engineer that makes me think that way. But for me, I don't think it's always bad to be a pipe. I think if you innovate on top of pipes, you have the strongest story. So I think if you look at a lot of the press that's happened over the last couple of years, it kind of started with the ANA report, man. Was that three cans ago that talked about mfa?
- [03:28]
A
Mfa, yep. Around this time last year, I think.
- [03:30]
B
Yeah, I think it was two years ago actually. I remember being on a panel and I hadn't even had a chance to review the study yet and someone asked me a question on it and I'm like, I was on a flight yesterday, I don't know yet. But I think that back then we were looking at doing our own in depth analysis on mfa and we realized that the agencies and the buyers that had close relationships with their SSP partners were doing the earliest version of curation, which was site list creating PMPs, eliminating things that they didn't see as quality. Where the ones that didn't have the MFA problem, I look back at Omnicom, they were at that point running almost exclusively off of Allowless. And they were not emailing us frantically about MFA because they had very strong relationships with their SSP partners. So that's why I say being a pipe isn't always bad, long as you're a smart pipe. And I think historically SSPs were called the dumb pipes. And I think we really are focused on being a smart pipe where we're working closely with our DSP partners to send exactly what is going to be bid on. It is about sending the right requests, sending high quality supply and not getting as involved in, I think some of the bad practices that are happening. When it comes to being commercial driven, I think SPO has become a commercial term instead of an advancement in how people buy. I'm a Firm believer that SPO should go back to being supply path optimization and not a commercial term that you negotiate.
- [05:10]
A
Yeah, this is a pretty controversial issue. Can you explain what you mean by this? Because I know what you mean, but spell it out for folks. What does it mean when SPO is a commercial term?
- [05:18]
B
Yeah, I think good SPO is data driven. So it's looking at the SSPs that are hitting the metrics that you're looking to hit as an agency or a buyer. Often what that looks like at the highest level it's going to be quality of supply, directness, eliminating things like mfa, as few hops as possible. I think global, how global you are is another part of an evaluation process. I think for spo. I think agencies are looking to consolidate partners in general. So I think having different partners in different countries or regions is, you know, has become challenging to navigate, especially when you tie the commercials into it. But I think where we look at it is that data should be the first step of an SPO exercise.
- [06:09]
A
Right?
- [06:09]
B
It's looking at the quality of the supply. It's looking at are they going to hit the goals that we are looking to drive? Are they willing to co innovate, Are they willing to actually work together on building tools that are going to drive the market forward? Are they doing what's right and paying publishers on time? There's a lot of those types of areas that I think should be the main qualifiers for spo. I think a lot of times what it ends up becoming is we're consolidating to you four or five SSPs. So here's the commercial terms that we expect.
- [06:37]
A
We're talking about discounts or paybacks.
- [06:40]
B
Exactly. Yeah. The post auction discounts fee caps were popular back in the day. But some version of a commercial structure which I get it, when you consolidate there should be some commercial advantages to that. But I think it has to be rooted in the data first and I think consolidating to what has historically been called the biggest SSPs, there's a reason why they're bigger in terms of requests and it's not what you're looking to buy. And we can talk about that later. I'm sure there's plenty of stats that we can pull in.
- [07:07]
A
Well, what do you mean? I think doesn't everyone have the same pre bid adapter on the same publishers?
- [07:13]
B
Yeah, but I'll say at the same time, if you look at we send about 500 billion requests daily. A lot of our public competitors are around 850 billion to 1 trillion a day.
- [07:26]
A
That's a lot.
- [07:26]
B
If we all have like if you look at a Jounce report or if you look at most of the directness reports, everybody in the top, let's call it 5 to 10 in terms of highest ranked SSPs are going to be probably between 87 and 94% coverage. Don't quote me on that, but somewhere in that general range.
- [07:47]
A
Yeah.
- [07:47]
B
So how is it possible that some people are sending a trillion requests while we are sending 500 million?
- [07:55]
A
There's reselling and there's open bidding.
- [07:57]
B
So there's two, I think there's two big culprits in that. It's just the supply that you don't want to buy. So it's mfa, it's seller hopping, it's in direct supply. So multiple paths to the same publisher. But I think one of the biggest one is request duplication. It's sending out 10, 15 requests for the same. And ultimately what that's doing is manipulating the DSP algorithms so hoping that sometimes you bid higher.
- [08:24]
A
I thought that was all stamped out, that bid duplication stuff. Because it's so easy to find.
- [08:29]
B
It's so easy to find. But I think that's why we go back to that conversation of what SPO actually is. I think it's first it's evaluating your SSP partners on those qualifiers but then also it's continuing to evaluate them on an ongoing basis. Yeah, but I also think, you know, a lot of that is it, it's allowed. And I think that lately, I would say over the last six months to a year, I've noticed a big difference in how independent agencies and holding companies and in some situations buyers directly are what they're evaluating on. I think the eyes have been opened a lot to quality. And you know, I think OpenX has always really leaned into that quality narrative. We've never claimed to be the biggest. In fact we see one of our superpowers as that we are not the biggest in terms of requests. Right. And we've remained pretty true to that. That quality comes first. And I think in an AI driven programmatic world it becomes even more important because garbage in, garbage out. If you don't have the highest quality supply, if you're not passing the right signals to DSPs and we start looking at like what is an SSP without a DSP and what a DSP is without an ssp. I think it looks a lot like people not sending the right things, not setting the other up to succeed because you're fighting the two sided marketplace.
- [09:56]
A
Do you have a buy side? I'm sorry, do you have a buy side?
- [09:59]
B
Like a bidder?
- [10:00]
A
Yeah, like a bidder. Because there seems like it's getting pretty popular among SSPs to have a buy side.
- [10:04]
B
It is getting pretty popular. We've not focused on building kind of that direct path.
- [10:08]
A
Yeah.
- [10:10]
B
For us, innovation has looked a lot more like keeping that clean supply. It's some of the AI driven products that I'm sure we'll talk about. In general, it's working closer with our DSP partners. I think we at OpenX are working closer with their DSP partners than ever to really make sure that we are the best SSP partner to them.
- [10:28]
A
Got it. All right, so we talked about the pipes. Let's talk about what goes through the pipes. Curation is the hot topic, or has been for about a year or so. Basically. SSPs enhancing the stream with all kinds. So what's OpenX up to there and how does it relate to industry trends?
- [10:46]
B
Yeah, so oddly enough, curation is not remotely new to OpenX. We built Identity Graph and launched it in 2018.
- [10:56]
A
I remember that. Yeah.
- [10:57]
B
So we had a product called, originally called Open Audience that has been rebranded and reshaped into OpenX select and that is ultimately our curation platform. So we relaunched that at the beginning of this year and it's been really quite interesting to see that we've actually it's 2x the amount of adoption since the relaunch in terms of spend going through OpenX Select. It originally was built for our curation partners. So data providers and strategic partners, they were getting in, they were ultimately building and creating their own deals and passing it to the DSP through some sort of a deal id. And the revamp of that has been giving agencies directly. So now we have quite a few independent agencies and all of the major holding companies touching OpenXelect and so on.
- [11:44]
A
Wait, what are they doing?
- [11:46]
B
Why is their keyboard.
- [11:47]
A
What are they doing? They're creating deals.
- [11:49]
B
They're creating deals. They're creating their own deals within the OpenXelect platform and still passing it to the DSP of their choice.
- [11:58]
A
My brain is starting to hurt. The DSP allows you to create your own campaign targeting, which presumably has the same data that the deal does. Why do they want to create. Why would they want to create. Create a deal that's effectively replicating functionality that they would get on the buy side.
- [12:17]
B
There's a couple different reasons. There's the more meaty reason, which I think is transparency and control. I think working directly with the SSP and curating within our platform does give you about as granular of controls as what you're buying as possible because you're.
- [12:33]
A
Closer to the supply.
- [12:34]
B
Correct. Us owning our own identity graph means that the match fidelity on on audiences are just much higher. But I also think transparency into fees, we are fully transparent around what you're buying and the price that it costs to do that. I don't think that's an industry norm as much anymore. So our platform really is rooted in control and transparency.
- [12:58]
A
So we're talking about cost of, let's say the data. If I want to overlay, I don't know, an Equifax segment on top of some inventory from a white list of publishers. Include list of publishers. The transparency is how much of each of those pieces is going to cost.
- [13:14]
B
Yeah, yeah. It's kind of the. We talk a lot about the supply chain and I've seen a lot of LinkedIn posts as of late around kind of the lack of transparency. I think I always find that a little bit interesting because of how granular we're willing to provide that transparency. We have a product called Bids that's built on Snowflake that gives about as granular of reporting as you can possibly get. I think we're really excited about the way that the industry is going because I do think transparency and quality have always been probably the biggest differentiator an SSP can provide. The thing that we've been evaluated the least on.
- [13:54]
A
Yeah. So there was a press release recently that said basically OpenX is the first platform to offer automated discounts to agencies. Firm curation. So how does that relate to what you're talking about?
- [14:06]
B
Yeah, so we have about, I believe it's 85 data providers that are directly matched into our identity graph. And you can actually access and buy directly through OpenX through a PMP deal or any sort of deal, I should say. And what we've done is we believe that the best way to continue to grow our partnership with our publishers is to get them the most amount of incremental demand and just revenue. Right. That's what publishers are looking for and buyers are looking for tools to drive performance. How do you combine those two things? We talked a little bit about post auction discounts. It's been a pretty popular commercial strategy for agencies to utilize with their SSP partners. It's in essence getting kind of in real time a discount back. So the way that we do it is what we would report back, what we'd pay the publisher is more than what we would collect from the dsp. So that comes out of our margins. So that's how we run post auction discounts. That's not how everybody does, but. But it comes out of our margins. What we've launched is for the ability for data providers to do the same thing in real time. So they're providing a discount on their data fees that goes back into working media dollars. So again, for us it's around how do we get all of those dollars? A lot of people are talking about how do you get more money into working media? The cost of everything. We're doing real work to actually drive results there. So you might not see us talk about it all the time. I'm not posting LinkedIn every day. That press release was great. It got picked up a lot. But I think we're doing it because we believe that that's the right thing for both the buyers and the publishers. I think for us, we're going to continue to stay true to the fact that we believe that's the role of an ssp.
- [15:55]
A
The data cost question is a complaint of the buy side quite a bit. There's a feeling on the buy side that they can't really project or reign in data costs or really understand where they're going to show up on the bill at the end of the month.
- [16:11]
B
Yes, 100%. It's interesting because I think a lot of the way that you're buying data now looks a lot more like a percentage of media versus a cpm. When I was more in the weeds at MIQ and I came from agencies in the past as well, you were used to having a good amount of transparency into what you're paying for a data provider because it was a cpm. And I think that there has been. I think most platforms are charging a percentage of media now, which does make it really challenging to understand what you're going to be paying. And there are particular verticals that are. Pharma brands have a much harder time. The cost of data is very high, very expensive. So for us, again, doing what we believe is the role of an SSP is to get the most amount of demand and incremental demand, specifically to publishers. And this is a great way of having strategic data providers that we work with buyers and publishers to all benefit through the OpenX select platform.
- [17:16]
A
Got it. Switching topics a little bit, what's the role, do you think, of an SSP or OpenX in particular, around performance and outcomes?
- [17:25]
B
Yeah, I think that's where outside of the internal efficiencies that we've been looking at with AI. I think AI is the way that we've seen our impact greatest on performance and driving performance for brands. We launched this year a product called results by OpenX and it's looking at different attributes within our bid stream and utilizing AI to predict the likelihood of an outcome. So it can be something as simple as a click through rate or higher levels of viewability. You can match your own first party data through our identity graph and utilize that to create a lookalike model to then drive performance that way. Or you can start getting into the real metrics. So a cost per completed or a cpa again that is an AI driven product that is actually utilizing a combination of our midstream data, our curation abilities and our internal teams that are doing a lot of the optimizing to actually drive performance. So we have a large tech brand that utilized this product somewhat recently. They were a really early adopter and I believe the stats were it drove about a 23% efficiency in click through rate and 11% percent efficiency in a CPA.
- [18:39]
A
So the buyer, either agency, advertiser, DSP is sending you back signals about what's working for their brand and then you're using that to predict which inventory will be performant and creating a deal ID.
- [18:54]
B
Different combination of our bid stream data and looking at that and then predicting the likelihood of a particular impression or request to actually drive performance and bidding specifically on that. So for us it's in my opinion that's like curation 2.0. You can talk about curation from a site list. Every article I read around curation I feel like people are talking about a different version of curation. For me this is an example of innovation in curation and it's utilizing AI, which I think is a really interesting use case. It's going beyond the noise of AI and it's actually a tangible product that people can utilize and actually see results for.
- [19:35]
A
This is a subtle question, but are people using this as a supplement to their other deals? Are they adding a deal that's like the performance deal or are they adding performance to all their deals?
- [19:49]
B
I would say it depends on the advertiser that tech brand started with. Everything was performance driven because this particular strategy, it was a. They have a lot of different lines of business. This particular lines of business was centered around performance. They had quite a few performance line items. I don't know exactly how the deal was set up. I guess it was a bunch of deals within one line item. But however it was set up, it was really Optimizing in and out of the partners that were actually driving the best performance. We were the partner that drove the most performance. By the end of the campaign, we were the actual only partner on the plan. That's an example of they tried a bunch of different tactics and ultimately the one that drove the highest level of performance, which I think is surprising that it's an ssp, but an SSP who are doing the type of work we're doing with AI. And for us we're really thinking about it from an approachable lens. So internally we call it a lot approachable AI. So if we're building this, is this something that's going to make the buyer's life easier? Is it something that they're going to be able to understand how it works and are they going to be able to go to their end client and explain it? Because to me that's, that's how right now AI is going to be successful.
- [20:59]
A
You brought up AI, so I feel like I have to round the bases on this question. Are you seeing any impact of AI on web traffic volume, publishers?
- [21:08]
B
Yeah, we look at that very frequently. I think what we have been noticing and talking about is there's particular categories within so shopping, for example, we see the biggest impact kind of in the shopping category. Then we're doing a lot of work to also look at that from an audience lens. So again, we have the 85 plus different data providers that we typically buy at a log level. So we can do some interesting analysis. We've been looking at is there an impact on the audience type? You and I might be utilizing AI more than my mom, for example. So we're really trying to innovate in that way to understand is there flooring work that can be done more on the audience level with publishers as opposed to the domain level.
- [21:54]
A
Interesting. Well, we usually finish these interviews with the lightning round, so I'm going to do sort of an abbreviated version. Just a couple quick questions. What would you say is your one biggest competitive advantage?
- [22:06]
B
Our one biggest competitive advantage? I would actually say it's our agility. I think we spend a lot of time working with our buyers and our publishers to understand what they're actually looking for. We have an amazing client advisory board and I think we are at the size our company is one independent and two at the size where we can pivot our strategy based off of the needs of our clients and what they're telling us. So, you know, it doesn't. We're not as clunky as others. I think we can move with speed and agility.
- [22:35]
A
Clunky. Those are fighting words.
- [22:38]
B
That's the second time you said it. I'm going to get in trouble.
- [22:41]
A
You are. And then what's your biggest challenge in the marketplace?
- [22:47]
B
I think the 12 sided marketplace conversation. I think a lot of SSPs are making moves, a lot of DSPs are making moves and I think there's a group in the middle that believe in a two sided marketplace and believe that Programmatic was built this way for a reason. And we might have some things wrong, but I would rather work on what is wrong. The request duplication around fees, around these commercial structures, around curation partners taking additional fees, curation partners outside of the SSPs taking additional fees and, and the finger pointing I think is ultimately the biggest challenge right now where it's like if we could actually just really start putting some standards in place and understand why Programmatica was built this way, I think we'd be more successful.
- [23:36]
A
All right, fair answer. And then the final question. If OpenX was an animal, what animal would it be? Oh man.
- [23:46]
B
There'S so many different routes I could go with this. I think one of the things that.
- [23:49]
A
Just don't say lion.
- [23:49]
B
I don't want to say tigers.
- [23:52]
A
Bear.
- [23:52]
B
No. I think what's interesting is we hear that we've kind of been this like sleeper for a while where you know, we were an OG. We are an OG, right? We're one of the original SSPs. We've been around for a long time and I think we were quiet for a while.
- [24:08]
A
Yeah, we're.
- [24:09]
B
And over the last four years I think a lot of people like probably the most common thing I hear at events like this is like where did OpenX come from? And like you guys have been like, you've always been around but you haven't been quite as. We haven't noticed you as much as we have over the last couple of years.
- [24:26]
A
I hate to drop. I'm still not hearing an animal.
- [24:28]
B
I know, I'm getting there. This is my delay tactic. So I think the whole reason that I'm saying that is I think there's this kind of sleeping giant mentality that I'm trying to think of. So I'm trying to think of what animal I would put that. Yeah, hibernate. Maybe we hibernated for a couple years.
- [24:43]
A
Let's do it.
- [24:45]
B
But yeah, I think ultimately how I think about OpenX is like it's been a fun ride and it's been a big change.
- [24:53]
A
I should disclose it's always good to have full disclosure. I own eight shares of OpenX.
- [24:59]
B
You do?
- [25:00]
A
Eight. Exactly eight. I was an advisor to a company. It got acquired by OpenX in stock and my tiny, tiny little bit turned into eight shares. Every year or so I get some notice from your investment.
- [25:11]
B
So you're rooting for us?
- [25:13]
A
Yeah. One day they may be worth a dollar or two. Yeah, maybe that would even be if you had a great exit. Matt Satel, thank you so much for being here. CRO of. I hope you have a great advertising week.
- [25:27]
B
Yeah, I'll see you tonight.
- [25:28]
A
See ya.