Description
What do gaming and programmatic advertising have in common? Stephen Johnston is joined by publisher development and yield management expert, Matt Kaye. The two reflect on their common interest and expertise in video games and programmatic advertising to uncover insights publishers can apply to monetization in 2023. Can what we learned in years past from premium subscriptions, paywalls and freemium upsell models be applied to the modern web? The issues explored in this episode for the benefit of creating sustainable revenue for publishers in the post-cookie era include supply and demand, creating scarcity, finding the ideal unit-to-volume ratio, and fill rates.
Transcript
Tony Winders: Hello, and welcome to another edition of On the SPOT. My name is Tony Winders. As always, we’re joined by programmatic technology expert, Stephen Johnston, Founder and CTO of PubWise. We’re also lucky to have a special guest today: publisher development and yield management expert, Matt Kaye is with us. I’m going to be putting both of them on the spot today about monetization, optimization and much more related to programmatic advertising and pub dev. Gentlemen, welcome to On the SPOT.
Matt Kaye: Thanks for having me.
Stephen Johnston: Thanks.
Tony Winders: Matt, you and I met almost 20 years ago, when we were both at ValueClick. I was leading marketing; you were on the pub dev team. It’s been a long time since we caught up. What have you been up to? What are you doing now?
Matt Kaye: Hard to believe it’s been 20 years, especially the last two or three being fast forwarded. I spent about a combined seven years there between FastClick and commerce and where we had met, which I would just kind of broadly qualify as the display advertising and retargeting world. From there, I actually moved to San Francisco from LA, where we went and got into the mobile app and gaming space. I did a four year stint at a company called Tapjoy, which was just recently acquired by ironSource. (It was a) really cool experience kind of comparing and contrasting mobile, web and web versus the app space and seeing how they compare for a while. And then, after Tapjoy, I’ve spent another 10 or so years in the native and content recommendation/programmatic space across a couple of really interesting companies. And currently, I’m just kind of consulting on a couple of non-ad tech ventures.
Tony Winders: I value the experience that you’ve had and dealing with publishers and understanding what they care about. Stephen and I talk a lot about monetization and supply path and demand path optimization and the technology that makes that work. The pressing question that I have for you, because I perceive that you have the ear of publishers and really are thinking about them, is what’s on publishers minds as we head into 2023? What are the pressing issues and the things that, as an industry, we should be concerned with and helping them do their work better?
Matt Kaye: It’s pretty simple in my mind. The number one thing publishers care about is generally revenue, as kind of cliche and simple as that sounds, it’s literally the lifeblood of the company. Secondarily, or tied with that, is managing their brand and growing their audience and doing stuff that’s obviously critical like that, as well. But at the end of the day, they’re always trying to grow revenue. I would say, trying to do so in a way that doesn’t ruin the user experience, which again, would kind of tarnish the brand and do other things that are undesirable. So, that’s kind of a good segue into some of the trends that I’m seeing are helping publishers earn more money these days, sort of what’s old is new again, in a sense. Over the last twenty years or over the last five or ten years of that, we’ve seen an explosion in programmatic advertising. As such, an explosion in supply, just the raw number of impressions that are out there won’t even go into the whole conversation about how many of those impressions are legitimate. But needless to say, there’s been sort of an abundance of impressions over the last ten years. And I think we’re seeing some elastic action here where we go back to creating a little bit more value through scarcity, not artificially, but just being a little bit more considerate about ad placements and where they are and how often they appear. That’s just one thing I see as being popular – creating more value for advertisers by frankly, just spamming less ads and being more thoughtful about the placements and really doubling down on those placements that are working and getting rid of the placements that aren’t working to create more value across the whole programmatic landscape. Secondarily (popular) is sort of going direct as much as possible, whether that’s just direct partner to partner with an advertiser and a publisher, using the programmatic pipelines to establish direct deals, or using private marketplaces and things like that. That, to me, seems to be the end game where you really can accomplish both of those goals. One, you’re getting more premium rates, you’re driving more revenue, but you’re also protecting your brand and keeping things clean with a much more predictable pipeline of advertiser types coming through.
Tony Winders: Stephen, are you seeing that from the publishers that PubWise works with or what does that bring up for you?
Stephen Johnston: Absolutely. I think maybe Matt was going to mention it, but there’s absolutely some focus on what really the post cookie thing means. In my opinion, that identifier chaos, if you will, of what is going to replace the cookie really is the transport mechanism of the identifier. What is going to replace it in the market actually contributes to the three things that Matt was just talking about heavily. Primarily, if you have a site that just lives off that high fructose corn syrup of third party cookies, then it’s going to be very challenging to continue to be worthwhile in the future. I think lopping off that segment traffic flow, be it arbitrage or something a little bit more valuable, but still just getting as much traffic through and siphoning off whatever you can through targeting is going to create part of that scarcity. That whole side of the market, along with header bidding that we’re doing, has really contributed to supply and the bid volumes going massively higher. Before, we had that waterfall and it controlled some of this, like, the cream got taken off, and then what was left and everybody fit in. So if you had one hundred at the top and you had two at the bottom, now you have 100-100-100-100 going out. Direct is just getting back to the place of, can we build a relationship so that we can control the ads that are coming in from the publisher side and get a better sense of the traffic that’s coming in from the buyer side? Because we have a real connection, now you can service that through more of the programmatic technologies like PNP, or we’ve created lower friction ways to do that, recently. Those options solve, for a lot of people in the market, complaints or challenges or concerns about traffic in demand.
Matt Kaye: That totally jives with what I was feeling. One other thing that I’d like to kind of throw in there that I sort of see as something that’s happening more often on premium publishers, and just across the board in a couple of different ways, is sort of building publisher or media experiences with different users in mind. You have, obviously, users coming in from mobile or social or coming from desktop or coming from search or various places. They might be receiving a different actual experience in terms of how they’re consuming the content, but most certainly in terms of what type of ad stack they’re getting based on where that user may or may not be coming from. And then one step even above that is this trend that we’re seeing in terms of subscriptions and paywalls and stuff like that, that I think is going to keep happening. I think the wall coming down at Netflix was the ultimate indicator that even the most premium media properties are going to have multiple paths for users to consume their content based on how they want to do it. I think that’s something that websites are going to be doing a lot more frankly, and not just the big guys doing this. So I just see that as another path to revenue and also creating scarcity legitimately, by gathering revenue from those that would rather just have the premium experience and then taking some of those impressions off the marketplace and making the other ones more valuable.
Stephen Johnston: There’s always that sort of unit price dynamic, and Netflix is the poster child of going the other way, where they had this premium product, and they’ve perceived that they’ve maybe capped out that channel. So they’re saying, “Okay, is there another place where we’re going to get a different amount than we get out of the premium? But can we open the scale back up?” Publishers, to some degree, are traditionally a web publisher or gaming publisher. Gaming publishers probably are closer to what you’re describing where they’ve always had this mindset of ‘there’s a way to pay for my app and get rid of the ads, or you can sit here and wait a little bit longer, or you can view this ad and get in.’ Publishers have really been focused on the web and on that ad experience. I think that making the platform capable of creating that subscription experience was lacking on the website. So, we’ve seen some growth tooling around with that, but I think the value proposition in gaming was already set in the customer’s mind of there being a premium and a freemium kind of thing. And on the web, I think people still chafe against that a little bit. Because if you put the negative incentives in the way on the web, like you have in gaming, people would revolt. So, that dynamic that was created that allows the upsell to happen in games, particularly mobile games, is a tough one to retrofit into the web. And so the web is having to say, “How do we create true upsell value?” whereas gaming got away with saying, “Can we create an acceptably negative experience that then creates the upsell?” It is interesting to see those dynamics a little.
Matt Kaye: One thing that it makes me think of is that game and app developers are a lot better at and do a lot more regularly than web companies and web media companies is understanding what the actual lifetime value is of a user. So it’s the understanding that if someone has your app, they’re going to have it for X number of years, they’re going to spend X number of dollars on average or consume X number of views or X number of clicks or whatever your metric is. Part of that was probably just tooling and identifier issues that we kind of mentioned before; it was just prior to the last year or two, it actually was very easy to identify users across apps and all over the place and really understand what the value was of a user over a couple of years. You could make a really informed decision before you decided on doing something so drastic, like putting up a paywall, or making these different experiences or things like that.
Stephen Johnston: So what percentage of game publishers do you think were really doing that sophisticated analysis of watching a player evolve into a different kind of buyer?
Matt Kaye: I think almost all of them. I was there for the sort of revolution of freemium, I would say, in 2010-2011-2012, when I was sort of going bananas. When the iPhone came out, everything was sort of premium and paid for the first couple of years, there wasn’t this notion of getting any content that was worth anything outside of a flashlight app for free. But then quickly after that, things evolved, again, across these lines. And, I think part of that was, when I was at Tapjoy, we were one of the first proprietors of the cost per install model for apps, generally. And we were doing it at scale really ahead of anyone else for a good solid year or two. We had like 10X revenue and had true product market fit, which was one of the rare times in my career where I really had that much wind in the sails and experienced that. I would say at that point, it was a rush for almost every game, maybe not every game, but 70-80% of games, to have multiple options especially if you think about the international aspects that were coming online back then and still coming online to some point now with Android phones and stuff like that. Premium option just wasn’t a thing for a lot of these people in India and other places.
Stephen Johnston: So what are they primarily driving with that persona or identity analytic? It’s not the models because those are kind of baked into the app and they’re there. Is it the price points and configuration that drives that? They’re looking at how much value you get in game for your dollar out of game, and then that aligns with the type of person and they’re tweaking price? Because they’re not really adding on a whole other model based on this?
Matt Kaye: No, I think it’s simpler than that; it’s kind of trying to give the user an experience that they want. If you’re someone that has the money and really enjoys the content, and the ads really annoy you, make that experience available. And then, if you’re another person that doesn’t want to spend that kind of money and is okay with the ad experience, try to tailor that ad experience to them as much as possible, given as many signals as you have about that user.
Stephen Johnston: And so then that ad experience is really different based on that user? You’re taking in information to say, “This person responds when ads are coming in at this rate, or are this long” versus, essentially, “How long is the X time and that kind of thing becomes a tunable?”
Matt Kaye: The gaming space is honestly fascinating, just because there are so many more data points, and so many more levers you can think about, in terms of showing the classic right ad to the right person at the right time. You’re not only trying to do that, but you can also capture emotional state to some extent. You know sequentially in the game that they’re really frustrated and just lost a certain level three times in a row, or you know if they just passed a certain level, they’re feeling really great about it, and they just added three friends on their social list. That kind of psychographic stuff, not that any of it is super precise, is still really interesting to tip it off as a data point that isn’t available in other places. You don’t really know the emotional state or intention of someone that’s reading an article online. And famously, that’s why Google has done as well as they’ve done over the years, since they capture all of the intent because you’re typing in exactly what you want and you’re at the bottom of the funnel, and they’re about to capture it.
Stephen Johnston: It’s interesting. So to some degree, the web has focused so much on, “How do we bucket them into product interest?” And it’s done a horrible job of where they are in that interest of the product. We’ll use a cookie; we know they’re interested, and we’ll just slap them with ads all over the place. And then the state that we’re in will just reveal itself because we’re going to be there all the time. I feel like that’s often the strategy, maybe not the only one. Do you think there’s an opportunity to get more of that with some of this identifier change or a re-prevalence of contextual? Could someone build something that gets a little bit more into where you are and what you’re feeling about it?
Matt Kaye: I think we’re on the way there. I think that a lot of the stuff that we’re talking about already in terms of providing different experiences for people to enjoy things in different ways gets us closer to that. I think again, just being more mindful about the ad experience gets us closer to that. Ideally, there’s ads that people enjoy, which is sort of the famous magazine model. Half of the magazine is ads, but they’re just so contextual and they’re so good that people are okay with it. I think the closer we get to that, and again, coming back from the explosion of every advertiser trying every impression everywhere all at once to just getting a little bit more reasonable, it gets us closer to that.
Stephen Johnston: If we lop off X percent because cookies go away, and we lop off Y percent because easy identifiers through mobile goes away, there’s still plenty of money flowing into the system out there. It starts to make sense that the unit price has to go up. Potentially, you may have a smaller number of publishers capturing more money, and now they’re able to invest in some of this extra work that it takes to get at that information.
Matt Kaye: One point that just came back to me is for publishers, it’s important to follow the revenue and the budgets more than following the tech, because a lot of times the tech is super far ahead of the budgets and the revenue. So even though header bidding is way better than a lot of other stuff that came before it, and even though native ads and social display ads are way sexier and way cooler than display ads and TV commercials, guess what? There’s way more money in TV commercials and display ads and these super old formats still. So I think that’s an important point, too, for publishers just generally, is to not get the shiny new thing syndrome and to not just say, “Hey, we’re going to do all these crazy types of formats and neglect display ads” because display ads aren’t sexy, but guess what? There’s probably a lot more agency dollars in those display ad campaigns than there are in those native campaigns.
Stephen Johnston: And that’s getting back to that unit to volume kind of thing. So much of the industry asks questions, and this is one of the things that we’ve been challenged to talk to customers about and do some education. People ask all the time, “What CPM are you going to get?” It’s irrelevant if you don’t ask what the fill rate is that goes with it. The real answer to that question is you might get X at the top 20%, Y at the middle 60%, and Z at the bottom 20%. That really is the answer to the question. Otherwise, we’ll just say $50 CPMs because they exist somewhere out there. It’s all about fill rate at the CPM; we’ve always tried to educate people about not just rate but volume. And then for us, we bring people into this conversation about sustainability, not in a climate sense, but in a sense of sustaining revenue. Can you make sure that you’re getting that for the next six months, a year, five years? Once you talk about saying, “Look, there’s the old steady stuff – display advertising,” it fits all of those things: high volume, decent rates. It will last over time, and then you pepper in there these more adventurous or more up and coming formats and things where you can capture some things. You’re talking a lot about some of the things we are, where this landscape is going to get more complex for some period of time because there’s changes, whether it be regulatory or by choice in the industry and cookies. Both the publisher’s path to revenue and maximizing (revenue) has become more complex but the opportunity is for them to capture a more robust set of revenue.
Matt Kaye: I think it’s an opportunity for people to work with platforms, like PubWise, and looking at stuff, like PubMatic is doing with their identity product, where they’re just bundling it all up makes it easy for people. They don’t have to use a bunch of their own engineers and do a bunch of decision making and business development around getting some identity data out there. I think that’s another interesting trend, too, that’s going to continue is just fewer partners. If those partners can do more things for people, that’s going to be looked on really favorably because I think that people have some partnership fatigue to some level.
Stephen Johnston: It is a thing, and it’s partly driven by people coming in. They probably promised a whole bunch of stuff and they delivered, unfortunately, outside of what they said. And that just drives that fatigue. You get people who are like, “Well, I just feel better doing what I’m doing instead of trying to do more.”
Matt Kaye: Yeah. And it makes it really hard to put together that lifetime value equation for yourself because you don’t know who’s exactly contributing what. So many people are sharing impressions, and the video advertiser guys are talking about RPMs. Everyone’s talking about their impressions and viewable impressions with different goalposts. It’s just very muddy, and that makes the decision making hard, but again, things like identity and all this other stuff make it a lot simpler, and I think it’s getting there.
Stephen Johnston: We try to draw everything back to net revenue. We talk about a lot of these metrics and even some of the things that you’re talking about, like the customer cares about revenue. And ultimately, that is the measure, but sometimes it feels a little bit like focusing on home runs in baseball, when there are things you have to do to get on base. Those things add up, and what we’re talking about in terms of whether it’s diversifying revenue or playing some smart technology to make it easier, those are the base hits along the way. But I think they’re a stand-in for revenue. That has to be the beacon, but you have to be out there doing those base hits and the things that are just smart along the way.
Matt Kaye: For sure. I mean, you have to eat your vegetables; you have to be making sure your ads.txt is good, your sellers.JSON is good, that you’re indexing well, and your SEO is good. The other thing, too, that I see happen sometimes is people connect to too many platforms, given the bandwidth they have in terms of operators they have on board to do actual trading and do some human level decision making. I think that’s maybe another interesting thing, or I don’t know if it’s a dirty secret, or just something that doesn’t get talked about a lot – yield management. Or you can call it curation, which is something that happens everywhere and is not necessarily talked about a lot because it’s not as sexy as saying, “Your machine learning did it,” or “Your AI did it,” or “Your algorithm did it,” or whatever did it. But, as smart as Google is, and as many data centers as they have, I guarantee you, those top 10 results have a certain level of human curation going on for those Google results, on top of everything that the algorithm and machine learning and whatever are doing. At highly functional ad platforms and exchanges, the same thing is happening. There is, in a lot of cases, yield management teams or traders or other people that are trying to find efficiencies within the platform, to come up with win-wins for these publishers and advertisers to help them get to more of these direct connections and add value themselves to the platform and justify their existence.
Stephen Johnston: That’s one of the reasons why we built our machine learning technology, not as a black box of throwing things in and seeing if something happens. We use a lot of human informed starting points for those decisions. We are completely reliant on where’s the starting point for timeouts? Who are the partners that you want to select between? Because the human should be making those good, really intuitive decisions about what they understand in the market and what seems to be working. The machine should be doing what it’s really good at, which is just verifying that intuition and verifying that net revenue production is greater than the other. Whether, from our perspective, companies should work with PubMatic, Rubicon, AppNexus, is up to them. What we’re using machine learning for, is to determine which segment of the audience, what time of the month, what time of the day, what part of the world and what devices, those three that they’ve made good decisions about, work and produce the most revenue. We look at, not doing the human things better, but where is the human essentially unable to do these things? And so, we’re producing tens of thousands of configurations for a site because the human team can’t do that, but we’re doing it based on what they told us was their good sense of what would work. And then we, as a company, are coming in and saying, “Hey, you’ve got a lot of traffic from X, Y and Z. You’re not working with this company yet, but they would help because we know that they monetize that well. Let’s put them into the mix and see how the algorithm tunes the mop and aligns the traffic.” I joke that if I could remove every process where someone says, “We’re going to wait a week, and then see what number is higher, and then just keep doing that more,” that would be my company’s mission; remove all the processes where all you do is get to the end of the week and see which number is higher than the other. But then, let the humans decide who’s in that competition.
Matt Kaye: Yeah, that sounds like you guys are going about it in a really smart way.
Stephen Johnston: Thanks. I appreciate that.
Tony Winders: Removing toil and waste, right, Stephen? That’s our mantra for the week and perhaps moving forward. Well, guys, I could chat with you for another 20 minutes and ask 10 more questions, but I think this is a great place to leave it. This has been insightful; I most appreciated the conversation about gaming and what web publishers can learn from some of the things that have happened historically in that realm. That was really interesting, and I just appreciate all of your insights, so we’ll have to leave it there for today. Stephen and Matt, thank you for all of your insight, and thanks to our audience for tuning in. This has been On the SPOT with Stephen Johnston, Founder and CTO of PubWise and our special guest, Matt Kaye. For more insight on programmatic ad tech and publisher monetization, please listen to our other podcast episodes, and visit our blog at pubwise.io. See you next time.
